Take the hearing test challenge!

[windowman]

Listen up sports fans! Here's a listening challenge for you that I think will be fun and hopefully benefitial for all of us. There's a lot of controversy (well at least I'm being contraversial http://www.audioforums.com/forums/smile.gif)about the notion as to whether or not there's any real benefit of recording at anything higher than 16/44.1. Actually I'm convinced that no one can even tell the difference between a well recorded 16/44.1 wave and a 320k MP3 copy of it (at one third the amount of information), but I've taken that a few steps further by doing a total of 6 conversions/resamplings of a percussion wave file that does a pretty good job of covering the frequency spectrum (even has triangles in it.)

If you can't tell the difference between the original 16/44.1 stereo wave file and it's copied/resampled/converted brother then I think it's safe to say that the world can well do without 24/96 and you can save your money and not fall for these sound company gimmics. (Just my opinion, you don't have to share it.) Please come back here and post your decisions as to which file is the original and which is the bastardized copy.

And for goodness sake please be honest about it and only use your ears and not test equipment. You've got untill next tuesday (New Years Day) after which I'll take the page down.

Thanks! Bill
http://www.geocities.com/billseper/index.html

PS, This is just a free GeoCities page so beware that they can't handle much traffic over there. If a lot of peoiple are doing downloads the page will automatically be shut down for a while (by them, not me) so if you get an error page try back later.

[This message has been edited by windowman (edited 12-29-2001).]

[huxtable jones]

I've mirrored the files:
http://hux.artistcollaboration.com/swampie.wav
http://hux.artistcollaboration.com/swoopie.wav

hope ya don't mind..if ya do lemme know, I'll snatch em back off there.

[huxtable jones]

well, here's me being perfectly honest...I'm somewhat disappointed to report that I can't tell any difference. Dunno if it's my ears, my equipment or what, but they sound the same...when all I've heard tells me I should hear a difference quite easily..I was really hoping to hear a difference...dang...might you have pulled a dirty trick and they're both the same file? http://www.audioforums.com/forums/smile.gif

[Dark Phader]

If you can't tell the difference between the original 16/44.1 stereo wave file and it's copied/resampled/converted brother then I think it's safe to say that the world can well do without 24/96

Rather improper. It's the logical equivalent of proposing: "If you can't tell the difference between the red cow and the green cow then I think it's safe to say that the world can well do without the blue cow."
Sorry to say but it is quite ridiculous to suggest that the scope of your survey, which includes no 24/96 data, has any bearing at all on the value of 24/96.

[huxtable jones]

I think Phader there has a valid point, but I notice you didn't say which was which.
Can you tell? Just curious.

[Sonic Valley]

My guess is that swoopy is the one that's been thu the ringer a few times. This come froms from a purely visual comparison as a 10 second drum loop makes it very hard to judge.It's also not really a good test due to the actual MP3 encoding. A better test would be a MIX with a wider spectrum that was recorded at 24bit then properly dithered to 16bit and the same mix that was done at 16bit all the way thru and see which one was the 24 bit dithered.



[This message has been edited by Sonic Valley (edited 12-29-2001).]

[huxtable jones]

Sonic I think yer probly one of the only folks around here that has the stuff to do that, and do it right http://www.audioforums.com/forums/smile.gif

After I saw it in a spectrograph I came to the same conclusion you did...but only after lookin at it...my ears tell me there's no difference...can anybody else not hear a difference? Come on, I'm beginning to seriously doubt my ears here... http://www.audioforums.com/forums/smile.gif

[windowman]

Well, thank God I'm not the only one with neither a date nor a gig tonight. This has got the makings of a baaadddd weekend. If the Rams loose tomorrow I might well be suicidal but nightfall. http://www.audioforums.com/forums/wink.gif

Hux, no I don't mind. It's your storage space man! And I really didn't pull a fast one. On Tuesday I'll post the answer as to which is the original and I'll give you exact steps as to how I changed it (I pretty much said it all on the web page anyway) so you can go through the procedure yourself.

Dark One;

Sorry you feel that way but for the life of me I can't see how anyone should expect to hear any difference in higher order sound files if they can't hear the difference in lower ones. I'd say your analogy is backwards myself. I'd say the difference between a 16/44.1 and a 24/96 wave is like the difference between the ride of an Infinity and a Rolls Royce. You can pay a whole lot more for a car but the ride just don't get any better than an Infinity. Whereas the difference between a 16/44.1 wave and a copy of it that's had 2/3's of it's guts pulled out not once but twice and then resampled several times should be like the difference between the ride of an Infinity and a much cheaper Taurus. But have you driven a Taurus lately? http://www.audioforums.com/forums/wink.gif

Whatever turns you on man. I ain't trying to beat you down my own path. But when these 24 bit processor workstations first came out, an acquaintance (don't really know em well enough to call em a friend yet) that works at one of the big St. Louis music stores and myself both thought there was no difference in the sound at all. And he sells this stuff for a living. The first thing I heard out of his mouth about it was how he thought the big makers of audio workstations weren't selling as much recording gear as they once had (seems everyone already had an Alesis adat, etc) so they devised a way sell more gear by telling people this 24/96 stuff would make em sound better. Nonsense I say! I have yet to hear anything done on 24 bit gear that sounds any better whatsoever than what myself and countless others have done on 16/44.1 I mean not once.

What kills me is that I know a lot of young guys love their music and are begging and borrowing every penny they can to get gear that they've been told will make them sound better and it just ain't gonna happen. I mean these guys are never gonna make it in the music business anyway and here they are out getting themselves in deep debt just to make some already rich guys even richer. I'm almost 43 and I've been suckered my share over the years too and I can't help but speak up on this one.

Save your money kids; and you won't have to marry some mean, ugly, rich woman so you'll have a roof over your head. Do you really want to live in your mom and dad's basement till you 30?

Be a great musician first and foremost because if you're truly great you'll always have an audience for your CD's even if they were recorded on $50 cassette deck. Heard any of those old Django Reinhardt records lately?

[This message has been edited by windowman (edited 12-29-2001).]

[windowman]

Hey Moderator Man! I sent you an email at the address listed in you profile earlier today. Also, see the post I left (partially for you) in the other thread which explains things. I need someone to keep me honest or Hux is never gonna believe me. http://www.audioforums.com/forums/biggrin.gif

[huxtable jones]

heh..naw, I believe ya - but only after I saw the file on the spectro-grapho-meter...there are differences...very small, but there...I can't hear em, though, even though the machine says "here they are"....but after lookin at the two graphs I don't feel so bad about not hearin it.. http://www.audioforums.com/forums/smile.gif

[Steve H]

I know exactly what you are driving at. The perceived quality is hard to justify. I work in film post and have for a couple of decades now. I've seen the evolution of mixing from/to 35mm mag to "all digital post". What is the effect, ..... it's had very little impact on the final film soundtrack quality. Analog 35mm Mag versions are still being recorded simultaneously to the digital versions of the soundtracks. The quality of the mag recording has to be a match to the digital or
the process stops! In fact, certain "digital" release formats are actually derived from analog recordings (DTS mainly).

It makes the sound post different, ..... it's quicker and cheaper for the director to change his/her mind, and the sound track. It's all about the money and this results in nobody wanting to commit until the last moment. Movies are frequently recut dozens of times before the final version. Each recut means hundreds of changes to the soundtrack. Each change results in some "re-recording." With changes comming like machine gun bullets, the only way that the soundtrack can avoid sounding like "mush" is if everyone uses the highest bit/sample rates possible. That way we can keep the quality up.

If you want a real test of the effects of 44.1k/16b quality, just re-record that same loop 50 times using 44.1k/16b and compare the same result at 96k/24b. The CD quality sample will sound like "crap" while the 96/24 sounds reasonably un-changed.

The real benifit of higher rates is that processing artifacts do not accumulate at levels which become audible very quickly. With magnetic recording, the artifacts were horrible. This was called a "generation loss." Now, with digital recording, you are not bound by the same limits. However, digital recording is not perfect either, Hence this thread.

The moral: For better quality, record at higher rates. Even if you can't hear the difference.

Steve H.

[Jazz E Bob]

Hmmm...

I listened for top end distortion and softening of attack.

I would put my money on "swoopie" as being the MP3. Don't really know why... Using the force here really - I stretched out with my feelings... :-)

PS - My Ensoniq EPS16+ still sounds brilliant in a mix and it's sampling rate is 32kHZ at 16 bit!!! It seems to have a "glossy" sound to it...

Good fun!

[Sonic Valley]

Windowman...I did get your email but got caught throwing a party I tried to cancel and couldn't...lol. After everyone left I had about 8 glasses of wine yet trying to do the comparison...lol

I'd rather not be a moderator as such for this and I don't think it's needed to validate the honesty of how you divised the test. We all trust ya.

I'm in complete agreeance with Steve H. It is very hard to hear the difference between 16 and 24. All our ears hear different ranges and the true benefit of 24 bit is to preserve the resolution while working on a file. Remember whenever you change eq, compress or change levels in a hard disk system, You are re-calculating in the computer. The actual word length is changing and the original "analysis" made by the A to D converters is being changed. The sound is being altered, and generally this means that some depth and resolution in the tone is lost. This translates into a harder, shallower, less detailed sound. Smoothness, width, depth, and openness are some of the first qualities to go whenever these new calculations are made. Keeping it 24 bit will at least give you some "resolution headroom"

I'm anxious to hear the results.

[windowman]

Oh, I agree whole heartedly about staying at 24 bit if you're gonna make a ton of changes/re-saving. I think however that it's clear, to me anyway :-), that if you only make 4 or 5 changes to any given wave file recorded at 16/44.1 there will be no noticable degradation of the signal. I would think most people don't make more than 6 changes to a wave after it's been recorded...dunno for sure though, I only have my own experience to go by on that.

Hey, I noticed Bob & Trock mentioned something over in the other thread about opening already recorded tracks with another program that made them sound better. While I would dispute that anyone could hear any difference in a wave recorded on CE, Cakewalk, Pro T, etc, I have noticed that when I use the "import file" feature in Cakewalk to import a wave I've made in CE, that it'll often get a little distorted. However, if I just copy the wave to the clipboard and then "paste" it into CW, it'll sound fine. Anybody have a clue as to what's going on with that?

[Sonic Valley]

Thing is you don't have to make a TON of changes. One EQ applied to a track means the whole file is re-calculated.

Going to check out the other thread...

[Dark Phader]

Originally posted by windowman:
Dark One;

Sorry you feel that way but for the life of me I can't see how anyone should expect to hear any difference in higher order sound files if they can't hear the difference in lower ones. I'd say your analogy is backwards myself.

It's not a matter of "feeling" it's a matter of scientific method and you present absolutely nothing that would enable anyone to infer anything about the qualities or lack thereof regarding 24/96.
Maybe the cow analogy was too far out for you and you need something a little closer to help you see the absurdity of your proposition.
Think about the following:

If one constructed a very similar experiment but used instead an 8 bit 32kHz file as the original source would the inability to detect a difference between the source file and one that went through some highly accurate encode/decode routines properly allow one to therefore categorically state, with reasonable confidence, that 16 bit 44.1kHz processing is overkill and that no real benefit can be achieved by recording at anything higher than 8 bit 32kHz?

[windowman]

Thing is you don't have to make a TON of changes. One EQ applied to a track means the whole file is re-calculated.

But I just made 6 changes and no one noticed…. Of course the track has been degraded but if it's not noticeable…so what? It's not so much like, "If a tree falls in a forest and nobody's there to hear it, did it really fall?" It's more like, "If a tree falls in the forest and nobody's there to hear it, would anybody care?" See what I'm saying?

It's not a matter of "feeling" it's a matter of scientific

It's the sensatory world we're talking about, do you HEAR a difference? If you don't then all the calculations in the world don't mean a thing.

If one constructed a very similar experiment but used instead an 8 bit 32kHz file as the original source would the inability to detect a difference between the source file and one that went through some highly accurate encode/decode routines properly allow one to therefore categorically state, with reasonable confidence, that 16 bit 44.1kHz processing is overkill and that no real benefit can be achieved by recording at anything higher than 8 bit 32kHz?

But you and anyone else WOULD notice a difference in taking an 8/32k file and making it smaller and then changing it back again. You would very easily notice. The lower you go in file size the easier it is to detect signal degradation. Take a Take a 256k MP3 file and cut it in half and there'd be very little difference. But take a 128k MP3 and cut it in half and the difference is much more noticeable. Take a 64k MP3 and cut that in half and the difference is astounding! The lower you go the more astounding the change.

I have yet to meet anyone that can hear the change even when you cut an 320 MP3 down to 160k. I have yet to meet anyone that can hear the difference between a 16/44.1 wave and a 320k MP3. By the time we get to a little 320K MP3 we're already at a place where no further data bits will make any difference in the sound no matter how much higher you go. The only reason to record as high as 16/44.1 is to help keep the signal strong after making conversions. I think I've already proven that you can make up to 6 DRASTIC conversions and still no one can detect it. (Or do you consider taking out 2/3's of the file twice "not" to be a drastic change?) Sorry, but that's just bad science to think that you'll hear a difference in sound if you record any higher. If you can then by all means let's hear it. I'm game.

I'm not trying to make any enemies here. If you've got it in your head that for some reason you'll be better off recording at 24 bit and you've got the bucks to monkey with then more power to ya. But I just want the younger kids and run-o-the-mill home recording enthusiasts to know that it's totally unnecessary and they should put their money to better use. College comes to mind…. http://www.audioforums.com/forums/smile.gif

~Bill~

PS, 35 Rams—14 Indianapolis, 7 minutes into the third quarter. Wahoo! The weekend isn't lost after all!

[RichMoore]

Hi Bill,

Your test is very interesting and while I'm totally new to this board, I want to throw in my $2. In grad school I took a course for fun called "Psycho-acoustics and perception processes" (something like that).

Mainly we learned about how scientists can say things like "Sweetner A is 1000 times sweeter than Sweetner B", with some certainty, even though A and B are totally different substances.

Turns out there's a whole science behind designing perception tests like this (which is essentially what you've tried to do here to prove a point)

One of the fundamental concepts is the "Just Noticeable Difference" or JND. It goes like this: You present the human test subjects with a series of choices like "which sample sounds better, A or B" and they have to answer A, B, or C-can't tell.

Now, you the tester, are controlling how much difference there actually is between the samples, and you run the gamut between "very close" to "obviously different"-- randomly of course. When you look at your subjects answers, at some point along your input difference curve, you will see them go from 90% wrong to 90% right.

That point around when they start answering correctly is your Just Noticeable Difference -- for a particular noise/sweetness/whatever level.

You have to run tests at a range of levels & you end up with a curve. In this case, noise on the X axis and perceived noise on the Y axis -- this is almost never a linear relationship, BTW, regardless of taste, sight, smell, touch or hearing.

So how does this relate to your test? What you've proven is that whatever noise difference you introduced into that pair of samples was beyond the hearing of most people. That's an OK conclusion.

But to extend this and conclude that the difference between a 16bit/44.1kHz sample and a 24bit/96kHz sample is undetectable does not hold any water at all.

Personally, I tend to agree with your hunch that most people's hearing will not be able to detect much difference, but you need a better test to prove it.

If you're interested in the perception measuring stuff, look for references to a guy named Fechner on the web -- he was a pioneer in this area.

Thanks for starting a lively discussion!

Rich

[windowman]

Say Rich, have you ever heard of a totally bizarre disorder called "synesthesia" where apparently it's a mixing of the senses in general where people see smells or taste sounds and so on. Really strange!
http://www.ncu.edu.tw/~daysa/synesthesia.htm#title%20explanation

[Dark Phader]

Originally posted by huxtable jones:
I think Phader there has a valid point, but I notice you didn't say which was which.
Can you tell? Just curious.

That's an interesting question. Telling which one is which is an entirely different matter than being able to tell that one is different from than other, which was the original propostion forwarded.

Consider an analogy of a digital photograph that didn't come out just right, maybe a little too dark and not enough contrast. So you take it through a few steps of processing in Photoshop and now it looks just great.
Now in this case the heavily processed file (photo) is the one that looks the best. The original is lacking in comparison.

The point is that in the photo example it is clearly easy to decide which photo looks better. But with no prior knowledge it is quite impossible to tell which was the original. I could have, for example, started with a perfectly fine photograph and darkened it. Which one is more pleasing would still be clear but you would be just as in the dark as to which one was the original.

It might be argued that MP3 being a lossy compression of a destructive nature would, if anything, make something sound worse. But in reality we use such destructive processing all the time to make things sound different in the direction of what we call better! Turning up a tube guitar amp to get the sweet saturation when the output tubes cook is such an example. Hitting the tape hard to get tape saturation is another such example. In both examples the audio data has been distorted and damaged compared to its original form, some frequencies have been lost, and the dynamic range of the signal has been compressed. Yet we approve of the sound and call it better.

In the case of the two soundfiles, as in the case of the two photographs, the only way to begin to know which was the original is to have a third file which is identified as the original that we can compare them to.

I can hear that the files are different but I do not know which is the original.

[This message has been edited by Dark Phader (edited 12-30-2001).]

[Dark Phader]

Originally posted by windowman:
It's the sensatory world we're talking about, do you HEAR a difference? If you don't then all the calculations in the world don't mean a thing.

Yes, I HEAR a difference.
I don't, however, understand what calculations you're talking about, nor is sensatory even a word.

Originally posted by windowman:
But you and anyone else WOULD notice a difference in taking an 8/32k file and making it smaller and then changing it back again. You would very easily notice.

Really? At http://www.64mustang.com/832_wav.html I took one of your files (not that I think they're a great example, but you seemed to like them), converted it to an 8 bit 32kHz wav then ran it through 6 conversions (3 round trips) to MP3 and back. The final 8 bit files were converted back to 16 bit (still 32k) as some high end audio programs no londer deal with 8 bit files.
How easily do you notice? Which one sounds more degraded to you?

Of course, it would be as silly to infer that one can make judgements about 16 bit 44.1 kHz based on this "test" just as it would be to infer that one can make judgements about 24 bit 96kHz based on your provided files.

Originally posted by windowman:
I think I've already proven that you can make up to 6 DRASTIC conversions and still no one can detect it. (Or do you consider taking out 2/3's of the file twice "not" to be a drastic change?) Sorry, but that's just bad science to think that you'll hear a difference in sound if you record any higher.

Actually you've proven very little, it would take a much more scientific and closely monitored experiment that produces statistically valid results that are then corraborated by others in similar tests. You've basically only made claims, with at least one of them, the one having to do with 24/96 rather silly, totally unfounded, and having nothing at all to do with your MP3 conversion experiment. The most you've proven is that some people can not hear the difference between one particular file and the end result of it being run through an MP3 algorithm several times.
If you're interested in finding out what you can and can't hear I suggest a trip over to http://www.pcabx.com/.

Originally posted by windowman:
If you've got it in your head that for some reason you'll be better off recording at 24 bit and you've got the bucks to monkey with then more power to ya. But I just want the younger kids and run-o-the-mill home recording enthusiasts to know that it's totally unnecessary and they should put their money to better use.

I do have that in my head. Just the headroom alone is reason enough.
Maybe your little experiment would be better received if you didn't try to draw conclusions that are totally outside of its scope.
You might also re-think what is really happening after the first MP3 conversion.

Originally posted by windowman:
College comes to mind…. :)

I do hope you're saving up.

[windowman]

I can hear that the files are different but I do not know which is the original.

But everything I did was to damage the copy of the original sound file, not make it sound better. Still, you do bring up another topic that I've been kind of thinking a lot about.

How bout we take a new approach to things? I know some of you guys must be close to my age and if so, probably made the transition from open-reel (that's reel to reel for you youngins http://www.audioforums.com/forums/smile.gif) to digital too. Have you noticed your recordings sounding any better? I can't say that mine do. In fact, I can't say that anybodies does. Sonically speaking, the best recordings I know of were made in the late 50's early 60's. Take a good hard listen to Brubeck's "Time Out" record sometime. It's absolutely stellar and I can't think of a thing I'd do to make it sound better. Ever hear of The Glass Harp? They had three of the best-recorded pop/rock records I ever heard and were made in the early 70's.

Are records made today any better sounding? I don't think so myself. I blame a lot of it on poor production though. One of the things I don't like about a high-end digital machines is that there's a lot of temptation to bring out as many highs and lows as you can. It's like there's this thing today where people dial so much treble into their waves that they just don't sound right to me. And I don't even wanna talk about the lows on that hip-hop stuff. Yuck! Just because you've got this wide frequency spectrum to deal with doesn't mean you have to use every ounce of it. And kind of closing in on what Darkman was talking about (using his photo analogy) since all MP3 encoders have a little different way of spreading their codec, sometimes you can take a poorly produced wave file and make it sound better but making it into a MP3. Some music sounds much better with a lot of the highs rolled off. Analog tape has a way of doing just that and introducing some lower mids that will generally sound great.

Do you really need a frequency response that goes 20 to 20k or better anyway? I really question that and definitely think a lot of stuff sounds better without the extreme high end that digital systems tend to produce. Now I understand about the upper harmonics and whatnot but I still question it. I'm going to use what I think is a great example of the less is more school. Anybody remember when Fostex first came out with their little 8 and 16 track machines that only used ¼ and ½ inch tape respectively? The whole industry thought it was a joke…at first. Next thing you know hit after hit made its way onto the charts that had been made on those 16 tracks. In fact it got real common for people to sync two of them together to get 32. They only had a frequency response of 40 to 18k and with built-in Dolby C a S/N of just under 80db at 0 VU. Not a whole heck of a lot better than a Nak cassette deck! But specks bad or not, they sounded fine.

Did anyone get Steve Vai's "Flexible" album? It was recorded in 83 in his home studio on one of the Fostex A-8 machines (8-track.) This is when he was still playing for Zappa and not many people knew who he was. It absolutely floored me when I first heard it. He was just a kid but his production was wonderful. The drums are tight and punchy without the cymbals being too bright or overly compressed. Here's a MP3 of a tune called "The Attitude Song" from it. Normally I wouldn't post an MP3 of someone else's work like this but it's an old tune now that never had many sales in it anyway and it's just for demonstration purposes so I don't think he'd mind.
http://www.geocities.com/billseper/vai.mp3

Now you might come back and say, "I could have made the drums/cymbals brighter or the bass even growlier" but I'd have to ask, "Why would you want to?"

[Dark Phader]

Originally posted by windowman:
But everything I did was to damage the copy of the original sound file, not make it sound better.

So you say. But in your little experiment you have not asked which one sounds like the original nor which one sounds better, only if one can hear the difference.

And what several MP3 conversions of a 16/44.1 file can have to say about 24/96 recording/editing is absolutely nothing.
Drawing absurd conclusions from close to zero posted results of a limited, unscientific experiment is clearly no way to help the "the younger kids and run-o-the-mill home recording enthusiasts " that you seem so concerned about.

[huxtable jones]

>>I do hope you're saving up.

That was necessary...where's hennyco?

Here's what I am getting from this thread:

I can't tell a difference between a 44.1/16 bit wav and a 320 kbps mp3...mebbe the sample was a bit limited, mebbe with some cymbals and a bassline I could have, dunno. I'm doing my own comparison on my own material to satisy my own curiosity.

I felt sincerely bad about it till I loaded the two samples up in a spectrograph...after seeing how tiny the differences are, I don't feel so bad about it.

Dark Phader, I think he's asking if one sounds "worse" than the other - the common belief is that mp3 encoding (at any bitrate) degrades sonic quality...I gather you did not hear one as being degraded, since you didn't specify - only that there are differences.

Now, can I take this and assume that I wouldn't be able to hear the difference between 16 and 24 bit? Weeeeeeeell, much as I'd like to say yes, I can't. Like Mr. Phader is pointing out, that ain't quite what we would establish with this test...I would like to hear the comparison as Sonic described, but having said that, I must also say that I've heard some of his work - if he says he can hear a difference between 16 and 24 I believe him. I can definitely see the advantage if yer into editing/processing the sound a whole lot..dry tracks? Not fer me, I don't have the talent or money for 24 bit or 96k. Others might. I'm happy at 16/44.1.

[windowman]

Yes, I HEAR a difference.

No way—no how.

I don't, however, understand what calculations you're talking about, nor is sensatory even a word.

Yes, that should read sensory. Aside from being rude what's your point? And it was you who mentioned calculations.

I took one of your files (not that I think they're a great example, but you seemed to like them)

Actually I'm not too big on Turkish percussion. I know of no better instrument for testing high-end than a triangle or lower mids than a batas though, and an egg shaker seemed to be well enough the for upper mids.

as some high end audio programs no londer deal with 8 bit files.

A…gee…kiddo, is "londer" in your dictionary? Can't seem to find it in mine.

I took one of your files, converted it to an 8 bit 32kHz wav then ran it through 6 conversions (3 round trips) to MP3 and back.

Actually when you went to convert the 8/32 wave to an MP3, what you got was a pop-up box from CE telling you that it "can not handle the current sample type" and asking if you wanted to convert to 44.1/16 before you ever made your first MP3 conversion.

The final 8 bit files were converted back to 16 bit (still 32k) as some high end audio programs no londer deal with 8 bit files.

"final" 8 bits my ass. You were in 16 bit long before that partner as I've already shown.

How easily do you notice? Which one sounds more degraded to you?
Of course, it would be as silly to infer that one can make judgements about 16 bit 44.1 kHz based on this "test" just as it would be to infer that one can make judgements about 24 bit 96kHz based on your provided files.

Actually I did notice. 101010 was the original. It was darn close though. It ought to be. They were not only buried in hiss so much that you could barely hear them but you never once made the copy of the original any less then 409,600 bytes in size compared to the 647,168-byte original 8/32 wave that you started with. How is that at all anything like the less than ¼ size reductions that I was doing?

I took a 1,769,472-byte 16/44.1 wave and after changing it to 48 res, made 320k mp3 file that measured 409,600 bytes. Do the math. In order for you to duplicate the amount of reduction I did (divide 1,769,472 by 409,600 and you get 4.32) you would have had to reduce your own 647,168-byte file to (divided by 4.32) 149,807. That would mean converting it into something less than a 128k MP3. (If you convert it to a 128k MP3 it you'll create about a 157,000-byte file which is still a bit larger than the 149K you need to duplicate what I did but on a smaller file size scale.

But you never even bothered to tell us what the size of the MP3 you created was. Don't bother, I already duplicated your tests and you converted at 320k just like I figured. Mine sounded just like yours. if you would have made a 128k MP3 for your conversions like you should have (at not quite ¼ the file size like I did) those triangle hits would sound an awful lot more like underwater, alien duck warbles. Even buried under a ton of hiss it's quite noticeable.

So making MP3 files that are only 1.58 times smaller than the wave you started with and burying it in hiss is like my test how?

The most you've proven is that some people can not hear the difference between one particular file and the end result of it being run through an MP3 algorithm several times.

Actually I proved at the very least that you could take a 16/44.1 wave and cut it's file size by less than ¾'s and still have an identical sounding file by spreading the remaining snapshots into the areas where human's best perceive musical sounds—that my boy is what an MP3 codec primarily does.

You can take that 16/44.1 wave and add ¼ more info there if you want and strategically place that info anywhere you want to in the frequency spectrum (wave files are spread evenly by the way which is why all wave recorders sound the same as Jazz E Bob was trying to say earlier) and it won't amount to a hill of beans. That was studied many times over years ago when the 16/44.1 rate was settled on for CD's. They settled on that rate primarily because no one could hear any difference when more bits were added.

[windowman]

Listen guys; it seems that this thread has gotten heated for some reason and that was not my intention. I'd like to think that forums such as these not only provide us with something to do on those weekends that some of us can't seem to come up with a date http://www.audioforums.com/forums/wink.gif but that we're here to help each other out. I'm not sure how anyone can settle this matter without buying an expensive soundcard that can do 24/96 and trying it themselves but then they'd be out some bucks for something they may find they didn't need after all. What seems like common sense to me apparently doesn't to Darkman and maybe a few others. So be it.

I had thought about having people like myself present their best "sounding" recording at 16/44.1 and have others do the same at 24/? to compare them but something tells me that there are some people who would never agree on their findings with that scenario either (as in, who used the better mic in the best sounding room and so on.)

As far as I'm concerned, if audio is ever going to sound better it'll come from making better speakers, or amps, or...heck I don't know. But pumping even more data bits into an already well-filled computer algorithm doesn't do anything for me. I'm well satisfied. To each his own.

[Dark Phader]

Originally posted by huxtable jones:
>>I do hope you're saving up.

That was necessary...where's hennyco?

I really don't know what was necessary. I didn't start the college bit.

Originally posted by huxtable jones:
I can't tell a difference between a 44.1/16 bit wav and a 320 kbps mp3.

Most of us can't, most of the time. However, and not to make you feel bad, you're the only one in the whole thread to admit to it in this case. Do you think anyone should proclaim scientific proof under such circumstances?

Originally posted by huxtable jones:
I think he's asking if one sounds "worse" than the other

Why would you think that? He never asks this question.

Originally posted by huxtable jones:
I gather you did not hear one as being degraded, since you didn't specify - only that there are differences.

And which photograph in my previous post would you pick as being degraded? Do I have to post some photo examples to assist?

If Bill wants to claim that it may be difficult for some to hear a difference between some 16/44.1 files and the same files after three passes through a good MP3 encoder @ 320kbps he may find me in agreement. Heck, many think that a 128k MP3 is CD quality. But to draw erroneous conclusions from a psuedo experiment, with only one result posted result in the "affirmative" (not able to tell the difference) is sheer folly.

[Dark Phader]

Originally posted by windowman:
No way—no how.

Why? Because you can't? Or just because you say so?

Originally posted by windowman:
Yes, that should read sensory. Aside from being rude what's your point? And it was you who mentioned calculations.

Rude? You're making up words so I'm being rude? And where exactly was it that I mentioned these calculations you speak of? Since I didn't does this fall into your rude category?

Originally posted by windowman:
A…gee…kiddo, is "londer" in your dictionary? Can't seem to find it in mine.


It's clearly a typo, gramps.

Originally posted by windowman:
Actually when you went to convert the 8/32 wave to an MP3, what you got was a pop-up box from CE telling you that it "can not handle the current sample type" and asking if you wanted to convert to 44.1/16 before you ever made your first MP3 conversion.

Actually CE was only used for the initial sample rate converison to 32k and not at all thereafter. So only you got the little pop-up box. I never saw the little creature.

Originally posted by windowman:
"final" 8 bits my ass. You were in 16 bit long before that partner as I've already shown.

Refresh my memory. How is it exactly that you showed this?

Originally posted by windowman:
Actually I did notice. 101010 was the original.

You think so? How was your score after running the two through the PCABX test?
I wonder what you think about this file: http://www.64mustang.com/files/021210021210.wav .

Originally posted by windowman:
I took a 1,769,472-byte 16/44.1 wave and after changing it to 48 res, made 320k mp3 file that measured 409,600 bytes. Do the math. In order for you to duplicate the amount of reduction I did (divide 1,769,472 by 409,600 and you get 4.32) you would have had to reduce your own 647,168-byte file to (divided by 4.32) 149,807. That would mean converting it into something less than a 128k MP3. (If you convert it to a 128k MP3 it you'll create about a 157,000-byte file which is still a bit larger than the 149K you need to duplicate what I did but on a smaller file size scale.

But you never even bothered to tell us what the size of the MP3 you created was. Don't bother, I already duplicated your tests and you converted at 320k just like I figured. Mine sounded just like yours. if you would have made a 128k MP3 for your conversions like you should have (at not quite ¼ the file size like I did) those triangle hits would sound an awful lot more like underwater, alien duck warbles. Even buried under a ton of hiss it's quite noticeable.


That's right I didn't mentioned the MP3 bitrate. I kept it the same as yours. Thought that was fair. You never mentioned the 48k step. So what? I can hear a difference in your two files and I can hear a difference in mine. And in neither case does it have anything to do with 24/96. And why I should have made a 128k MP3 instead of a 320k MP3 is beyond me, especially when your whole experiment is regarding the effectual quality of 320kbps MP3 conversion and nothing else. Which, if you had a reasonable number of replies, you may have been able to correlate some data followed by a posting of some possibly meaningful results.
You say you've duplicated my tests but I highly doubt that as it's apparent from your own post that you're not using the same software or methodology, and I doubt you could come up, on your own, with matching files.

Originally posted by windowman:
So making MP3 files that are only 1.58 times smaller than the wave you started with and burying it in hiss is like my test how?

Although it starts with a lower quality file it (BTW the hiss was not an attempt to bury anything it was the dither used to bring the 16 file to 8 bits. Yes it's noisy but the alternative of no dither is worse.) uses the same MP3 bitrate and is built from the same source. And the results so far are about the same. And it has as much to do with 24/96 as your test.

Originally posted by windowman:
Actually I proved at the very least that you could take a 16/44.1 wave and cut it's file size by less than ¾'s and still have an identical sounding file by spreading the remaining snapshots into the areas where human's best perceive musical sounds—that my boy is what an MP3 codec primarily does.

You proved this by tabulating the result of one person who claimed to not hear a difference and ignoring any claims to the contrary? That's some powerful science you got there.

And what is it again that you proved about 24/96, which was my original bone of contention? And how was it exactly that you proved this?

[This message has been edited by Dark Phader (edited 12-30-2001).]

[Dark Phader]

Originally posted by windowman:
[B]Listen guys; it seems that this thread has gotten heated for some reason and that was not my intention.

That's really silly. There is nothing heated in this discussion.
You're making a bunch of unwarranted claims and I'm simply exposing them for what they are.
This really is nothing personal.

[windowman]

Listen kid, I've got nothing more to say to you. Get it? Go pick on your little sister and come back when you've got at least something halfway reasonable to say. You're only hurting yourself. Nobody ever won a debate with childish lies.

[SYG]

Sounds like Swampie to me, because the bell or whatever it is sounds richer and truer.
SYG

[huxtable jones]

>>However, and not to make you feel bad, you're the only one in the whole thread to admit to it in this case. Do you think anyone should proclaim scientific proof under such circumstances?


fair enough. thanks for the audio testing link, great stuff. In fairness to windowman, the tone I get from his posts is this is just an informal thing between acquaintences...I didn't and don't get the feeling he's purporting official scientific results...more like water cooler type conversation.

Anyways, I learned something about myself, and I know of one other person who did too...I'm not quite as shy about making a fool of myself as some http://www.audioforums.com/forums/smile.gif

I'd hate to see this thread follow the same progression as so many have recently on this board...good points have been made by all, I think...no need to insult anyone's education or little sister, is there? As soon as that stuff starts we no longer get any good information...mebbe I'm still naive and idealistic, but a forum like this could be a great tool (again) if we make it so. Hate to see this turn into yet another pissin match, that's all.

Happy holidays.

[Dark Phader]

Originally posted by windowman:
Listen kid, I've got nothing more to say to you. Get it? Go pick on your little sister and come back when you've got at least something halfway reasonable to say. You're only hurting yourself. Nobody ever won a debate with childish lies.

I see, so your concern over possibly creating a heated thread was just feigned, as now you're trying to make it personal.
For a person with nothing more to say you've certainly exited with the same abhorrence of truth and objectivity that you came in with.

[beetlefan]

You may prove that multiple conversions of a 16-bit file back and forth from an mp3 is transparent but you still did not prove your original claim that there is no difference between 24-bit and 16-bit.

[windowman]

Hey Beetle;

Long time no talk. Remember the long boring discussion we had with some others last year about the sound quality of various MP3 encoders? This thread is really conjuring up some bad memories... http://www.audioforums.com/forums/smile.gif

Try and look at it from my point of view one more time will you? digital recording has been around for 30 years. Most of the improvements that have been made were things along the lines of oversampling for error correction and that sort of thing, most of which has remained basically unchanged since the early 80's. The more dramatic changes in digital audio since then have come from the compression specialists. They've gotten incredibly good at masking out quantization noise and so on, dumping bits that are to low in amplitude to be audible, and placing more bits where you need them i.e. in the 3k range for presence and so on. Now you have to ask yourself a couple of key questions:

1. If CD's are going to stay at 16.44.1 then would recording at anything higher than that only to have to downsample it later anyway be of any real value? The only 2 sensible reasons for doing this would be:

A. Better error correction in the AD/DA conversion. (Did you have any noticeable errors when tracking at 16/44.1? I know I don't.)

B. Having more bits to keep the sound in tact after doing conversions. (The problem here is that most of us do very little destructive editing on our multi-track waves; it's the final mixdown stereo wave that gets most of the generation loss so to speak. It's that stereo mix that we normalize, cut off beginnings and endings, compress or limit, and so on. It would be much more cost efficient to record at 16/44.1 and then upsample to 24/?, make you edits, and then transform back to 16/44.1. Am I wrong?)

2. If it's possible to make a better recording than we already have at 16/44.1 but impossible to hear any difference in the final product without the consumer having spent a thousand bucks on speakers alone then is it at all sensible? (And I'm not convinced that anyone could hear any difference even with the best hi-fi listening equipment anyway.)

3. Have you heard that sample I put up of the Steve Vai tune? Have you been able to make better sounding recordings at 16/44.1? In theory you certainly should. If you can't then it's other avenues of the recording process that you need to work on and not higher bit rates. (I'm not singling you out here Beetle but all of us.) At 16/44.1 you should already be able to produce recordings at a much higher quality level than Vai did. But who do you know that has? I've listened to quite a few samples people have put up of their music here and at other websites. By and large the sound quality isn't very good. People need to work on their engineering techniques and not worry about higher bit rates until they've proven they can make the most of 16/44.1 IMHO.

This is something Roger Nichols said once that I got a kick out of:
*******************

Bits, Bits, Bits

"My wife says all I do is bits. It’s 20bits this, 24bits that. I mean what’s the floating point? Just when you thought it was safe to abandon your trustworthy analog recorder for digital audio, the bit wars begin. It is as bad as the ballot recounting in Palm Beach County.

I look at bits like automobiles. The nice 16bit models cruise along pretty good, top out at the 44.1k speed limit, and get very good mileage. A tank of bits (hard disk space) can get you through a whole album’s worth of mixes including all of the various versions with vocals up and down.

You go to the showroom to check out the new hot rods. 24bits of power with a top speed of 96k. You pop the extra cash for this baby and drive it home. After a week of driving it around the block you invite your buddies over and take it for a spin. Running at top speed with all the extra bits of power the tank empties in 1/3 the time it did with your old machine. You didn’t count on that extra expense, but you go out and buy extra hard disks to avoid running out of road. Then you notice little burps every once in awhile. It turns out that your bit pump can’t keep up with the new demand and you have to spend more money on a hopped-up SCSI accelerator.

On your last project you only needed 57 tracks to record everything. Way under the top end limit of the 16bit 44.1k machine. You check the page in the manual for the 24bit 96k setup and find out that you have a maximum track capacity of 24 tracks at 96k. There is no back seat in this hot rod. You will have to leave some of your buddies behind. Not enough tracks.

Now you are a little wiser. You had your fun in the new speedster, and you are looking at pictures of next year’s 192k 32bit models. You ponder the shelf full of awards you won with your new album and wonder if it was all worth it. Yup. Been there. Done that. You go out to the garage and pull the bicycle off of your old 16-bit beauty and dust it off. “Well,” you think out loud, ”we’ve had some good times together. I’ve learned a lot from this machine. The new 24bits machine is nice, but maybe the 96k is pushing it. I could save 50% on my hard disk bill by just keeping my foot off the (SCSI) accelerator. And I will have room for all of my buddies on the next project.” You walk back into the house smiling."
****************

Thought that summed it up pretty well. http://www.audioforums.com/forums/smile.gif In all fairness he does record at 24/96 quite often and records all of Steely Dan's live gigs this way. But if you've followed his articles at all in the past he makes it very clear that it's all rather vain and almost in fun. Like "I'm rich and famous—what the hell!"

See ya, Bill

[windowman]

Also, like I said earlier, it seems a bit odd that we've "cruised along", as Nichols puts it, at 16/44.1 for 15 years and nobody's complained or suggested that going to a higher bit rate (which they could have done 15 years ago too) would be of any benefit. All of the sudden sales are down for audio workstations and suddenly the guys in the suits have decided that we need higher bit rates after all. Right.… Sorry, I don't buy it.

BTW;

I realize most of us will be going out tonight and maybe out of town for a couple of days so I'll go ahead and tell you now—Swoopie was the original file. You can make your own bastardized brother (Swampie) by following the directions on the web page for doing so. Also, I said there that I had done something to try and keep people from being able to tell which file was which by spectral graphs, etc. I was bluffing. http://www.audioforums.com/forums/wink.gif

One last thought about this stuff: Has it ever dawned on any of you that the designers of AD/DA converters could have followed along the lines of the folks at Fraunhofer et al, by coming up with a algorithm/codec for wave files that would have taken 16/44.1 and instead of evenly spreading it across the frequency spectrum placed more data bits where they're more necessary and masked the dead space noise where they weren't? Not that I think the sound could be improved anyway but that would certainly have gone a lot further toward improving the sound than simply throwing 1/3 more data bits into the pile evenly dispersed. And it wouldn't have cost us anything but maybe an updated chip on the sound cards we already have. By giving us more bits instead, we now have to buy new sound cards, bigger HD's, more RAM, and new updated software for taking advantage of the so-called improvements. The guys who already have digital workstations will need to buy new AD/DA converters and bigger HD's for the extra info storage or better yet (in the mind's of the guys in the suits) new workstations altogether.

But to each his own; life's to short for me to worry about whether you buy into that stuff or not. Let's just have some fun doing what we love.


[This message has been edited by windowman (edited 12-31-2001).]

[Dark Phader]

Originally posted by beetlefan:
You may prove that multiple conversions of a 16-bit file back and forth from an mp3 is transparent...

He may do that someday. He may yet do that in the specific case of the one example that has been proffered but so far the tally of those reporting back to the thread is 1 claim of no difference, and 4 claims for a difference. I suggest that if there's any proof it is to the the contrary. Unless I'm missing somebody as of this writing 400% more people have made statements indicating a difference than those not.

[Dark Phader]

Originally posted by windowman:
Also, like I said earlier, it seems a bit odd that we've "cruised along", as Nichols puts it, at 16/44.1 for 15 years and nobody's complained or suggested that going to a higher bit rate (which they could have done 15 years ago too) would be of any benefit.

That's simply not true. There have been complaints about 16/44.1 since the introduction of the CD almost 20 years ago, and not just from audiophiles.
Around 9 years ago Sony introduced SBM or Super Bit Mapping that claimed 20 - 24 bits of resolution in a 16 bit word as one method of providing higher quality audio on a CD and opened the doors for others to find ways to cram more significant information into the not quite enough 16 bits.
And even though there were many quality complaints 15 years ago, the technology to use any more than 16 bits sampled up to 48k was prohibitively expensive and 24/96 was virtually unfeasible.

[Bajoran]

So boys and girls, what have we learnt after reading windowman's long posts on this and other threads? Well, according to his own admission, he cannot perceive the differences between a song recorded and mixed at 16-bit and one recorded and mixed at 24-bit.

So what does this tell those of us who have no trouble perceiving the increase in clarity, transparency and openness of 24-bit compared to the closed-in sound of 16-bit?

I would say it boils down to one or the other of these:

1) His sound card, converters or monitoring system is not of sufficient quality to enable him to perceive these differences

2) He has never directly A-B'ed the same mix (from a 24-bit recording) done at 24-bit and at 16-bit on a high-quality system (e.g. Masterlink)

3) He has some kind of perception problem (not trying to be rude here but if it's not 1) or 2), then I don't know what else it would be)

Maybe I missed something but I suspect that it is one of these 3. 1) and 2) are not at all uncommon.

Baj

[windowman]

Gee, a new member with 2 whole posts and maybe a proxy server to come in thru?

Do you really think you're fooling anybody? That's just sad.

[Dark Phader]

Originally posted by windowman:
Gee, a new member with 2 whole posts and maybe a proxy server to come in thru?

Do you really think you're fooling anybody? That's just sad.

I don't know who are addressing in this post but if you are trying to say something you should just have the balls to say it. In my opinion you're borderline libeling which is about as bad a thing as can be done in a group such as this.

For my own part I take part in these discussions straight up. And although I use a handle (Dark Phader), it's the only one I use and my profile information, which has from day one, contained the URL to my main website and email address and is available to all. As of this writing, your profile contains neither.

When you start intimating that others might by lying or creating false personae because you can't force everyone to accept your point of view is somewhat paranoidal.

Apparently your sense of self is as small as your capacity for logical thought.

[JohnBussoletti]

The progression from 16 bits to 20 to 24 and higher has followed the development of hardware (A/D and D/A converters) of improved quality at (relatively) low prices. Same thing goes for sample rates. It's not an industry conspiracy to earn more money. It's simply taking advantage of the same technological developments that have led to 1 GHz processing speeds when a 100 MHz processor would have been more than enough for, say word processing applications.

The perceptive differences in the recording technologies are subjective and so dependent on not only individuals, but their particular sound cards, amplifiers, speakers, listening environment etc. Science says that higher bit resolution and higher sampling rates offer a quatitatively more accurate representation of the physics. But the choice of 16bits/44.1 for CDs was made at the time because it provided a clarity and accuracy of sound that was never achievable by vinyl or tape. Remember, vinyl was the standard for music distribution some years ago. (For the nostalgic, there are even plug-ins that allow you to introduce hiss, pops, wow and all the other wonderful characteristics that accompanied music recorded on vinyl when Windowman and I were growing up.) Now with DVD technology and the beginnings of DVD-RW standards and hardware, you can expect to see an attempt to develop a consumer market that will be based on 96kHz sampling and who-knows how many bits. So if this consumer market develops, we better be able to generate recordings that are compatible with the technology.

That said, good music doesn't always need good (or near perfect) technology. Bad technology doesn't necessarily (always) mean bad music. If I'm listening to music as I drive down the road in my car, especially if the windows are open on a nice sunny day, I'm not going to have a chance to hear any significant differences between a 16 bit or 20 bit or 24 bit recording, nor a difference between reasonable mp3 encodings or 44.1 or 96 kHz sampling. Same goes for listening to music on a boom box at the beach.

So what does this rambling add up to? Just know your market and do what's important for the market. There are no magic technology solutions. There aren't even all that many technology requirements. You have to create something that a large enough mass of the rest of us recognize as pleasing or enjoyable. That's all. (And that's quite enough!)

I think that was behind Windowman's arguments more than anything else. At least, that's what I got out of it all.

Flames away!!!!

[Jazz E Bob]

I want to see the vote tally for the test...

[Bajoran]

JohnBussoletti, I actually agree with much of what you say except for, "But the choice of 16bits/44.1 for CDs was made at the time because it provided a clarity and accuracy of sound that was never achievable by vinyl or tape."

At the time one of the reasons for the outcry against CDs was that the quality of the converters was very poor and tended to be innacurate in the highs and sound quite harsh. 44k also chops off the upper harmonics above 22k while analog tape at the time was reproducing frequencies and harmonics up to 30k. 1/2 inch masters running at 30 ips sounded very clear and noise-free and 16-bit 44k was a considerable degradation in sound quality.

Since that time the quality of converters has improved considerably and the jump to 24-bit was a major step. In "The Mastering Engineer's Handbook", while there are people who are insisting that the only way to fully get a similar clarity to the best of analog is to go to 192k, a number of the mastering engineers comment that current 24-bit 96k technology has for them reached a quality level comparable to the best of analog.

Just as there is good and bad analog quality, so there is good and bad digital quality (jitter being a major bugbear in digital).

Baj

[windowman]

Here's a piece of an article I picked up somewhere about digital audio and mini disc recorders. Thought it was funny.

History has - surprisingly - repeated itself yet again with the invention of the Sony Minidisc. Minidisc was also deemed not to be for professional use, but with so many sound engineers recording (eg) interviews on location in the broadcast industry, coupled with the relative mechanical frailty of DAT machines, Minidisc has become an acceptable standard for such applications. Music Sound Engineers steer well clear of it though, as they are usually deeply suspicious of the way it modifies the sound in order to compress it onto a small disc. This suspicion is probably a little bit unfair, considering that Dolby-SR noise reduction (which most music engineers love) does a very similar thing and no-one ever complains about that. Actually the similarities are such that Sony inadvertently breached a Dolby patent on sound processing when they invented Minidisc, and a royalty is due to Dolby on every Minidisc machine sold. That's why Dolby is mentioned on the back of all Minidisc machines. Clever chaps, those Dolby people.

Something to remember about Minidisc is that - being software driven - the sound quality has improved with each successive generation. The first generation Minidisc players were criticized for sounding "metallic", but now we're onto revision 4.5 of the ATRAC encode/decode algorithm the sound quality is pretty spectacular actually. Speaking personally, I LOVE the sound of my Sony MZ-R5ST Minidisc recorder. It seems to put back in the "transparency" that gets lost through the recording chain. Although I master onto DAT, I'm seriously considering looping future mixdowns through one generation of Minidisc to capture that great Minidisc sound! I can't help but wonder if the Minidisc ATRACK algorithm - designed as it is to throw away what it considers to be superfluous background junk - throws away certain intermodulation distortions that otherwise "muddy" the mix.

Of course the general public will probably not be blown away by this technology - they simply don't have the heritage or experience to know how to use it. Give the Korg Portastudio to an experienced Record Producer and they'll be able to produce something approaching production quality with relative ease. Give the same machine to a 16 year-old and they'll produce something that sounds exactly as bedroom recordings always sound - crap - and what's more, they will blame the machine for it...
***********

Ya gotta love it! http://www.audioforums.com/forums/smile.gif

I'll tell what I think is amusing; since the first DAT's hit the market it's been one review after another about how, "This machine is incredible. You'll get out of the back end exactly what you put into the front end!" And then only to find the exact same self-proclaimed audio expert coming back to review some new digital recorder or ad/da converters the very next year and saying how the new generation is better yet! This continues year after year. Well, if you got exactly out the back end what you put in the front end with the first generation machines (hey, that's what they were all saying), then how can you improve on "exactly the same"?

Ah those wondrous, human ears. Everyone thinks they have golden ones. Oh well, I imagine my old 16 bit card will break down eventually and by the time it does I'll be forced to buy a 24 or maybe even higher one by then since 16 bit will likely be a thing of the past someday. What are ya gonna do though? Ya gotta float where the water travels.

It's hell being a jellyfish....

But like Phil Keaggy once said, "As soon as they're in the car all that fidelity goes right out the window as soon as they crack it." I thought it was a lot more fun when we all had little 4 and 8 track analog machines. It was fun to see how much you could get out of em and it was almost a contest to see who could make the best recordings on them. Now everybody's got easy access to decent digital recording with almost unlimited tracks. But instead of making great music on them it's like everyone wants to see how bad they can make them sound. Have you heard any of the alternative, techno junk that's out there now? There's no accounting for taste it seems. But if your end product is suppose to sound like crap warmed over, why not start with a 4-track cassette to begin with. Guess I just ain't hip enough to understand it.

WeeeLLL, I've been power lifting most of my adult life. I like any sport that doesn't require me to chase a ball around in the grass on a hot summer day. http://www.audioforums.com/forums/wink.gif Unfortunately my joints can't handle the heavy squats anymore and I'm forced to sit in the leg press now along side the other sissies. I'm too old to compete, but I like going to tournaments now and then. We got a couple of guys in St. Louis that are closing in on the thousand pound mark in the squat. When I was in college very few people used steroids and human growth hormone. You did the best you could with what the good Lord gave you. Now guys are shattering records with the stuff. Back in the 50's only 2 people in the world could bench press 600. At last years bench-press championships in Tennessee, your starting lift had to be no less than 700! Three guys started at 740!

But you know, I never once sat around wondering how much better I might have been if I'd have used drugs. I wonder how good they'd be without them.

Well, that's the end of another lengthy spiel, (as someone said.) My little associates degree was in computer repair but as a weird quirk of fate I ended up the owner of a commercial window cleaning company as my main gig since 93. And it's been a bit too cold around here too work the past week so I've had time to play on the net for a bit. Gotta get back to the grindstone though. I LOVE to work! See ya's.

(For my threatening little nemesis)

Bill Seper /owner
We Do Windows
Box 23135
Belleville, IL. 62223
618-277-1989

Drop by sometime. We'll have French Vanilla coffee and I'll show ya how to change some capacitors in the tank tone circuits in your old Fender Bandmaster that it'll make it blow the doors off a Boogie Mark III!

[Heavens to Betsy 2]

This entire thread is a disappointing rope-a-dope, apparently designed for the loudest gentleman to antagonise people and lure them into an argument. And all for what--a ridiculously subjective decision.

Who is being hurt by the usage of sampling and bit rates higher than 44.1/16? Nobody--not even the engineer if he's sure he can hear a difference. If he has any sense, he will have weighed the purported benefits of such a resolution against the additional requirements for storage (and hardware purchase); so when he does decide to use the higher resolution, it's based on a reasoned decision.

Not long ago I was skeptical of any benefits of sampling beyond 44.1 or 48, discarding any perceptions as psychological tics. Having unintentionally heard an obvious difference in clarity very recently, I have abruptly changed my opinion. If I waste disk space and throughput, so be it. I bought my rig! http://www.audioforums.com/forums/smile.gif

And finally: who appointed Nyquist master of reality, especially when human hearing and sonics are so scantily understood? Newtonian physics was indisputable in its adherence to perceived reality--until Einstein clocked it.

My ears can tell a difference and it's because of sampling rates. What would Nyquist say? I dunno. Go dig him up and ask him.

[Tim Z]

Hey, right on Heavens !! Too many of these forums have become a place for "the loudest gentleman to antagonise people and lure them into an argument" and I (and many others) are getting VERY tired of it.

Personally, I could care less what others think about this issue and all the number crunching. I know for a fact that my 50 year old ears can hear the differences on my system, with my converters and my monitors; especially on cymbals and reverb tails. Slight as they may be, I am going to use my 24/96 bit converters to their max. Do you need better than 16/44.1 to make great recordings? Absolutely no, but everyone should already know that. We don't need to keep harping on this. Ten years from now this whole debate is going to be a joke because there will likely be a whole new technology that we'll be saying "can you believe we were recording at xx/yy"

Anyway, I just wanted to say that after seeing most of your posts Heavens to Betsy, you and I appear to have similar thoughts on the whole recording scene.

Take Care

Tim

[Heavens to Betsy 2]

Thank you, Tim! Honestly, it's good to know that there's some patience and wise perspective around.

You couldn't have hit it more squarely: how could anyone possibly think that the book has been written for digital audio? I agree, too, that a good song can be captured without the cutting-edge; and when we want fidelity and something approaching (or succeeding) the glories of analog, we look to technologies such as 96/24 or whatever may take its place.

Incidentally, it was cymbal crashes in 96/24 that dropped my jaw! Smooth as can be.

Well met, my good fellow! If you're not a participant already, you may wish to check out the forums at http://forums.syntrillium.com as it's well-mediated and populated with some great minds.

[Dark Phader]

[QUOTE]Originally posted by windowman:
(For my threatening little nemesis)

<snip>

Drop by sometime. We'll have French Vanilla coffee and I'll show ya how to change some capacitors in the tank tone circuits in your old Fender Bandmaster that it'll make it blow the doors off a Boogie Mark III!/QUOTE]

I guess you missed the point. I care not for your personal information. The point was that you, through false accusations and innuendo, tried to defame another poster as being fraudulent but yet provided no valid information about yourself.
As to why you call me threatening, when the record shows I have not threatened you, is as puzzling as your various "proofs".
As for being your nemesis, that's a bit silly, I just wanted to set the facts straight. Apparently you wanted to interact on different level; not what I was looking for, but I can hold my own.
As to your kind offer, thanks, but no thanks. I don't like sweetened or flavored coffee, nor do I have a Bandmaster. And I certainly wouldn't want my Deluxe Reverb Reissue to sound anything like a Boogie.
I do regret that I am unable to find the corollary between power lifting and audio editing, but if you're really interested in a good discussion on the subject you might try posting in the alt.sport.weightlifting and misc.fitness.weightlifting newsgroups.

[BoBDolez]

I didn't bother to take the test. Obviously it is not a scientific study and obviously Windowman didn't try and pass it off as being one. He didn't try to publish his study in the New England Journal of Medicine did he? What is a scientific study anyway? Double blind controlled randomized placebo? Doesn't matter really - every study ever done has bias. Generally you need to just look at the source of funding for a "scientific" study to see where the potential bias is coming from. When it comes to studies, you can find or design one to back up whatever point you are trying to peddle or push. Every statistic can be manipulated and marketed for a purpose.

I'm not an audiophile, I'm a consumer and a hack hobbyist musician-wannabe. I value the advice of Windowman who is simply trying to save the little guy some money. Let's face it - this forum is nothing but little-guys. There are no elite audio-gods or high-end users here. I will agree with him that the push to higher level formats is primarily for financial gain and secondarily for audio quality. Who the hell is the music being made for anyway? It sure isn't being made for audiophiles. Perhaps 1-2% of the population can tell a difference between the 2 formats. Perhaps more? So some of you experts can tell the difference between the two formats - well be thankfull for your rare gifted ears - you obviously have genetically superior ears or know exactly what to listen for. Does that make a difference to something as subjective as music? As an analogy we consider 20/20 vision as perfect vision - yet there are people out there with better vision such as: 20/15, 20/13, and even 20/10 vision(1% of the population). Can those with 20/10 vision "see" the Mona Lisa better than you or I? Do you even think the person with 20/10 vision could spot the difference between the original Mona Lisa and a computer generated copy?

It's a logical progression of technology to some degree. Technology will continue to rise exponentially whether we need/like it or not. Expecially if someone can make a buck out of it. 24/96 will be surpassed and become old hat to even the lowly hobbyist some day. The cost of providing 24/96 to the masses is obviously dropping and like it or not - it will eventually become the next standard as we all know. Competition within the marketplace will necessitate all vendors to eventually provide this format. I don't need 24/96 but if I have to choose one and they are the same price - well that really isn't a tough decision is it. But IF I am not a millionaire and thus have limited funds I think his message is highly valuable regardless of his scientific methods. Windowman is obviously trying to provide practical advice to musicians and wannabees at the bottom of the musical food chain (ie. me and you) albeit in a non-"scientific" manner.

Personally I would like to stay with 16/44 for a little while longer so I don't have to upgrade every damn bit of my signal path and pump up the power of my cpu/harddrive/etc for the ****** music I'm going to make that only my friends and family are going to hear. If I was producing Janet Jackson's newest CD I sure wouldn't use 16/44 even though it would be good enough - you know what I mean. I still think the current push to consumers for that standard is primarily financial.

[This message has been edited by BoBDolez (edited 01-04-2002).]

[Heavens to Betsy 2]

I don't need 24/96 --"BobDolez"

Of course! Nobody needs 96/24 or any greater resolution that offers better sound for the price of less throughput and storage ability. 44.1/16 does the trick--and even a four-track will do if equipment is limited. Lo-fi can carry a good song in many situations.

But the difference is real and the alternative is available for people who would like to use it. It's especially tempting to try for those of us who have sound cards capable of either 44.1/16 or 96/24.

[hennyco]

24/96 is stupid and it takes to much room anyway. no one can tell teh difference because there is none.

[beetlefan]

If I do more than two changes to a 16-bit file I can hear the degredation in the form of loss of clarity and slight smearing of the transients, and that's just on my home stereo through speakers.

The difference between 16-bit and 24-bit is that the 24-bit sound has more openness or spaciousness.

I can hear the differences. I'm not imangining them.

BUT, if you record to 16-bit without any changes, it is better than going to 24-bit and dithering, maybe with noise shaping. Noise shaping changes the sound.

96k sampling rate? All marketing hype so far as I can tell, as I have not yet heard it.

People always tend to be wary of anything that asks them to part with their $$$.

My .02 on the subject.

[hennyco]

anyone who says they can tell the diff is lying or making it up. there is no diff.

[Heavens to Betsy 2]

Interesting, Beetle. I did some experiments myself and found that I heard less difference between bit rates than with sampling rates, keeping in mind that I avoid work in 16-bit like the plague.

As far as 96kHz--I can't say this too many times. A few weeks ago and I could have been just as skeptical...but I heard what I heard and there is a smoothness inherent to the higher sampling rate. Go figure!...though now that I think about it, Beetle, most of your work is vinyl transfer and the difference in sound may not be as stark as with recordings laid down on your own rig (sort of like the way that analog recordings still retain their magic when transferred to CD).

And "hennyco," lighten up. Nobody's been claiming superiority because of a higher resolution. Some of us have the extra storage space, have heard a difference, and are interested in playing around with the stuff. It's our time and money.

Don't take it personally, okay? [ http://www.audioforums.com/forums/smile.gif]

[This message has been edited by Heavens to Betsy 2 (edited 01-04-2002).]

[JohnBussoletti]

At the risk of flaying the proverbial horse, Bajoran's comments regarding my earlier statements point out the heart of the matter:

Bajoran said:

'JohnBussoletti, I actually agree with much of what you say except for, "But the choice of 16bits/44.1 for CDs was made at the time because it provided a clarity and accuracy of sound that was never achievable by vinyl or tape."
At the time one of the reasons for the outcry against CDs was that the quality of the converters was very poor and tended to be innacurate in the highs and sound quite harsh. 44k also chops off the upper harmonics above 22k while analog tape at the time was reproducing frequencies and harmonics up to 30k. 1/2 inch masters running at 30 ips sounded very clear and noise-free and 16-bit 44k was a considerable degradation in sound quality'

Well, yes, in the studio 1/2 inch tape at 30 ips give very good results. But the end product in that era was a vinyl recording or perhaps a cassette tape. Yes, in the studio, with the best of equipment and the experience and training from working with that equipment, you can tell a difference. But for the mass market, CD audio 44.1kHz/16 bits is a far cry "better" technically (S/N ratio, frequency response, etc) than vinly or cassette tape.

Again, the point is know your market. There may be a few audiophiles with the equipment and/or prejudices to distinguish how things were mastered. But as far as the mass market (and isn't that where the money/fame/prestige etc lies???), you'll be hard pressed to find anyone who can tell the difference in subtle audio effects that have generated such discussion under this topic.

Create the art, the best way you can, but know that the art itself is the product that leads people to give their hard-earned dollars to share in it in whatever way they can, not the technology to produce the art. That said, as far as things electronic and computer-related, recent history has shown and continues to demonstrate that technically superior and lower cost products continue to emerge that supplant what was wonderful levels of capability two to three years previously.

So do the best that you can for the art, exploiting the best technologies available, but don't sweat the bits/kHz, unless you're really sure the market requires it.

Which I guess isn't really saying anything at all. Sorry for wasting your time (sic).

[Bajoran]

At the risk at of stirring up another hornet's nest, one of the reasons pro engineers and producers like to record with 24-bit is that there is considerable agreement in the recording and mastering community that a 16-bit 44k CD well dithered from a good 24-bit recording/mix sounds better at the consumer's end than one that was recorded/mixed originally at 16-bit.

Baj

[beetlefan]

Originally posted by hennyco:
anyone who says they can tell the diff is lying or making it up. there is no diff.

Anyone who says they CAN'T hear the difference needs their hearing tested. And, how many people here are going to admit to attending too many rock concerts without hearing protection?

[Sonic Valley]

"anyone who says they can tell the diff is lying or making it up. there is no diff."

You would hear the difference if you had a good gear and a good ear and a good understanding of the recording process http://www.audioforums.com/forums/smile.gif

Any-hoo...I'm off. Have fun in 16bit degradation land...lol.

[Kihoalu]

Here is a good "degradation" test I have done which will let you hear at least what the mp3 conversion process does. This, of course, does not directly address this 24/96 vs 16/44 argument but it is fun to do anyway. Take a good source file with subtle instruments (like a violin solo) and also with some percussion (like cymbals/ hihats). You may need to piece together things to get this. Save the original .wav and then comvert to mp3. Convert the mp3 back to .wav. Get the original .wav back and the double-converted .wav and match levels AS CLOSELY AS POSSIBLE. Then invert one of the .wavs and add (mix) the two together. If you matched the levels well you will MOSTLY hear the differece, which is the stuff that the mp3 THREW OUT. I was impressed that the difference was around -20db (i.e. not that much was actually thrown-away) and that the character of what WAS thrown-away was very "noise-like". If you process the processed file AGAIN you will find out that in the next generation VERY LITTLE more stuff is thrown-away. On the first pass thru the mp3 converter the data becomes more "mp3 compatible" and the encoding process throws away less stuff than before.

...K