SACD vs. CD - Unfair competition?

[Sir Terrence the Terrible]

I am also curious of how SirT can explain how mixing should counter some of the effects caused by downsampling when the critical point, as he claims, is recreation of the waveform and not the high-frequency content (and when reducing the HF content to 18 kHz i.e. inaudible, the waveform suffers, how can THAT be possible???).

I am afraid the question you are asking is not well formed. You can counter the loss of percussion transients partially through eq (note I said partially) You can pan some instruments further out into the soundfield to counter image shifting towards the middle. All of this is far from perfect, but (in some cases) can partially deal with some of the effects of downsampling. I never mentioned reducing anything to 18khz, and the waveform suffers because it is not complete, it has been altered through the downconversion process (i.e cymbals, and percussion which have significant high frequency content sounds audibly blunted)

What kind of mixing is that? Can he show e.g. the result from such type of mixing e.g. what it does to a waveform?

Your apparent lack of experience betrays you. We don't mix waveforms, we mix music.

Is something understood in this process? IF and I say IF there was an audible difference how does SirT know that this was not caused by any flaws during conversion, e.g. decreased level of high-frequency content up to 22,050 kHz?

Well, I think it is obvious that any information over 22khz is gone, any flaws during conversion, well I use Lavry digital products, and they are considered amoung the best and most transparent in the world.

How does he know that the previous high-frequency content >22,050 kHz did NOT cause any peculiar distorsion in e.g. the speakers that was audible?

The tweeter in the speaker is flat to 50khz, it will not distort, and distortion stays well below audiblity as long as we don't exceed more than 100db's. This is not really relevant since I do not hear much above 18khz anyway. What we heard wasn't any distortion artifiact anyway. What we heard was a shifting of instruments in the soundfield, a digital glaze that covered the mix, percussion losing its punch and timbre, and the total inability for the downsampled audio to accurately play back a muted trumpet(it sounded nothing like the master tape).

[Sir Terrence the Terrible]

No,

I did not participate in this particular test, but I've attended other listening tests blindly and double-blind in the studio. So I am very familiar what kind of methods they use during testing and what listeners that usually can reveal differences. And I say that most the times, differences they or I hear or feel between different kinds of equipment are VERY subtle.

After all the many tests with 16/44.1, they have concluded that the format as such is better, yea, much better than what most people BELIEVE, when the most transparent equipment for encoding and decoding are used.

T

Unfortunate Thomas, the most transparent equipment for downconverting audio is VERY expensive and out of the reach of most DIY mixers, and a great many studios. It is also unrealistic to think that all equipment and processes are completely transparent, and without problems. This why I do not like the downconversion process.

I would also like to offer to you that most people who do most of their mixing and mastering at 44.1khz will not even know what to listen for when listening to a downconverted digital file. If the test included well recorded cymbals crashes, high hat claps, or muted trumpets and you still considered it transparent, then I would call you a liar to you face. These are tortute tests for redbook CD, I have yet to hear in over 20 years experience these cleanly and accurately reproduced at 44.1khz.

Something strikes me funny here. In your previous postings you never made mention to the fact that you participated in "all of these" listening test. You just mentioned you were a layman. When did it occur to you that you had went from layman, to experienced listener?

You did your test, I did mine. We both came to different conclusions. You don't work in audio and admit you are a layman, I work in audio, mostly in with DD, Dts, CD, DVD-A and SACD. I have learned and trained my ear over the years on what to listen for with high to low downconversion. If you have not trained your ear on what to listen for, then I'll bet good money you won't hear any difference. If you think that you can walk in off the street and be able to indentify any audible differences, then you are totally fooling yourself.

Did "they" meaning the other people make this opinion. Or did you with your laymans experience make an opinion of their own? Which is it?

[E-Stat]

If the test included well recorded cymbals crashes, high hat claps, or muted trumpets and you still considered it transparent, then I would call you a liar to you face. These are tortute tests for redbook CD, I have yet to hear in over 20 years experience these cleanly and accurately reproduced at 44.1khz.

Not to mention attempting to "float" a triangle over a symphony as one hears in the live event. Redbook at its best is very good. Perfect reproduction or "transparent" as Thomas would say is naive.

rw

[Thomas_A]

Unfortunate Thomas, the most transparent equipment for downconverting audio is VERY expensive and out of the reach of most DIY mixers, and a great many studios. It is also unrealistic to think that all equipment and processes are completely transparent, and without problems. This why I do not like the downconversion process.

I would also like to offer to you that most people who do most of their mixing and mastering at 44.1khz will not even know what to listen for when listening to a downconverted digital file. If the test included well recorded cymbals crashes, high hat claps, or muted trumpets and you still considered it transparent, then I would call you a liar to you face. These are tortute tests for redbook CD, I have yet to hear in over 20 years experience these cleanly and accurately reproduced at 44.1khz.

You did your test, I did mine. We both came to different conclusions. You don't work in audio and admit you are a layman, I work in audio, mostly in with DD, Dts, CD, DVD-A and SACD. I have learned and trained my ear over the years on what to listen for with high to low downconversion. If you have not trained your ear on what to listen for, then I'll bet good money you won't hear any difference. If you think that you can walk in off the street and be able to indentify any audible differences, then you are totally fooling yourself.

Did "they" meaning the other people make this opinion. Or did you with your laymans experience make an opinion of their own? Which is it?

They made the conclusion, not me. Now to a small experiment. If you take a 16/44.1 recording which naturally do not contain high frequency content, convert it to higher sample rate and down-convert back, there should be no audible difference between the original and the processed one if the converter is of good quality. Have you done this experiment with the equipment at hand? Just asking, because if there is a difference, how do you know that the reason for the "audible difference" is caused by reducing high-frequency content and waveform accuracy during down-conversion of signals with wider bandwidth?

[E-Stat]

Have you done this experiment with the equipment at hand? Just asking, because if there is a difference, how do you know that the reason for the "audible difference" is caused by reducing high-frequency content and waveform accuracy during down-conversion of signals with wider bandwidth?

Not only that, what is the airspeed velocity of an unladen swallow?

rw

[Sir Terrence the Terrible]

They made the conclusion, not me. Now to a small experiment. If you take a 16/44.1 recording which naturally do not contain high frequency content, convert it to higher sample rate and down-convert back, there should be no audible difference between the original and the processed one if the converter is of good quality. Have you done this experiment with the equipment at hand? Just asking, because if there is a difference, how do you know that the reason for the "audible difference" is caused by reducing high-frequency content and waveform accuracy during down-conversion of signals with wider bandwidth?

My mother always told me this. Since you mentioned it, you can do it. I am passing this on to you. If you take ANY signal, and subject it to that kind of processing expecting a transparent result, then I have a island in San Francisco bay for sale. Since no one does this in real life, this test is unrepresentative of any real life situation and means nothing..

This test does not prove converter transparency at all, do you have anything else in your laymans handbook that does apply?

[Thomas_A]

And...SirT you mix music which do not consists of waveforms? Is it lack of experience to think that music does not consist of waveforms?

[Sir Terrence the Terrible]

And...SirT you mix music which do not consists of waveforms? Is it lack of experience to think that music does not consist of waveforms?

Just like I thought, you spin your point when you cannot come up with answers. Music does consist of MANY waveforms. However do you see engineers staring at scopes, or level indicators? Gone to majoring in minors huh?

[Thomas_A]

My mother always told me this. Since you mentioned it, you can do it. I am passing this on to you. If you take ANY signal, and subject it to that kind of processing expecting a transparent result, then I have a island in San Francisco bay for sale. Since no one does this in real life, this test is unrepresentative of any real life situation and means nothing..

Why should I be surprised...you claim working professionally with audio, and you don't know that testing the equipment needs to be done also in real life?

[Thomas_A]

Just like I thought, you spin your point when you cannot come up with answers. Music does consist of MANY waveforms. However do you see engineers staring at scopes, or level indicators? Gone to majoring in minors huh?

Waveform does not mean solely pure sinewave. A waveform may consist of any combination of waves, in music hamonically related, changing over time to create melody.

[Thomas_A]

Not only that, what is the airspeed velocity of an unladen swallow?

rw

African or European?

[E-Stat]

African or European?

Both for you.

Ni !

rw

[Thomas_A]

"Something strikes me funny here. In your previous postings you never made mention to the fact that you participated in "all of these" listening test. You just mentioned you were a layman. When did it occur to you that you had went from layman, to experienced listener?"

Why should I mention everything I've done during 25 years of music/audio hobby? I've been involved in listening test, e.g. one which we performed years ago and where we had invited one person from AudioReview to listen for CD players differences. The link has been posted several times on audioreview. I participate in listening tests and sometimes write articles in Swedish Audio Technical Society's journal. Last, I joined a test with 30 different headphones.

But what has this to do with the central question of this thread, i.e. whether 16/44.1 is audibly flawed or not? The claims of difference comes from you and not me.

[hifitommy]

if he desires to answer tohse questions, he will. i take him at his word that it was necessary, and that there was a bandaid approach he used that he didnt feel should be necessary. that is why he prefers the higher res formats.

i suppose if ansel adams were alive, he would have experimented with digital cameras and then return to film just as film buffs prefer film over the newer digital media.

[mtrycraft]

No bother, just curiosity. Never met someone who expends the amount of energy you do into an interest without ever experiencing it.

rw

Now you have met someone. First time for everything.
Curiosity is good.

[mtrycraft]

if you have been reading all along, you will know thqat he is a professiional recordist and has 'been there, done that' and knows these things by experience and has had to deal with the consequences. he is paid to do his job correctly, he aparently EARNS the money by doing so.

Oh, yes, the almighty 'professional' label take care of it. Immunity and expertise unlimited, right? I have exchanged with so called 'professionals.' Refund anyone?
Now you are hanging your hat on authority. Never question it. Maybe enogh for you. Not for others.

[hifitommy]

i know thats a word youre unfamiliar with. get a dictionary, and then try getting some.

[Sir Terrence the Terrible]

"Something strikes me funny here. In your previous postings you never made mention to the fact that you participated in "all of these" listening test. You just mentioned you were a layman. When did it occur to you that you had went from layman, to experienced listener?"

Why should I mention everything I've done during 25 years of music/audio hobby? I've been involved in listening test, e.g. one which we performed years ago and where we had invited one person from AudioReview to listen for CD players differences. The link has been posted several times on audioreview. I participate in listening tests and sometimes write articles in Swedish Audio Technical Society's journal. Last, I joined a test with 30 different headphones.

But what has this to do with the central question of this thread, i.e. whether 16/44.1 is audibly flawed or not? The claims of difference comes from you and not me.

I personally don't think we are going to settle this question on this board. And it seems that we have been reduced to discussing small irrelevant issues. You and Mtry seem to think that 16/44.1khz is just fine, me and many in my line of work don't. I am not here to prove anything to anyone, so I think this discussion has run its course. If you believe that redbook CD is fine, keep using your CD player. For those of us that love hi rez multichannel DVD-A and SACD, we'll keep using our DVD players. Your happy, I am happy.

[mtrycraft]

I personally don't think we are going to settle this question on this board.

That is correct. But, even if we did, would the rest of the audio world listen? Of course not.


And it seems that we have been reduced to discussing small irrelevant issues. You and Mtry seem to think that 16/44.1khz is just fine, me and many in my line of work don't.

Yes, at the consumer end, not mastering, it is fine until there is evidence, credible evidence, for audible differences solly due to the sampling and bit differences, not some other factor introduced in the chain by someone.
Oh, we are not alone by a long shot, here, maybe.

I am not here to prove anything to anyone, so I think this discussion has run its course.

But, being in the inductry I would think you of all people would have the necessary credible evidence to support what you have been claiming.

If you believe that redbook CD is fine, keep using your CD player.

Unfortunately I prefer multi channel, even for music.

Your happy, I am happy.

How can anyone argue this?

[Smokey]

I can't belive this. After 218 responses and 3,487 views, we are still back in square one as far as CD vs higher resolution format such as DVD-A or SACD

How about this argument.

Even if we agree that higher sampling rate probably don't improve sound quality due to Nyquist Criterion, then what about improved ratio of noise and dynamic improvements for DVD-A/SACD?

Brick wall noise filtering of CD at 22 kHz is most definitely audible on some recordings, and moving of noise filtering to 44 kHz for DVD-A definitely improve the harsh product of brick filtering. Not to mention improved Dynamic headroom

[Thomas_A]

In studio apps yes, higher sampling and bits make sense due to easier work.

I don't agree that SACD always have higher dynamics than standard CD. Looking at the highest octace 10-20 kHz noise starts to increase in SACD, at least in all measurements of players I've looked at on Stereophile's homepage. Also I have hard time of how transients can be accurately be reproduced by SACD since the HF content will be buried in noise. Technically, DVD-A would be superior. Also 1-bit systems will have problem to accurately reproduce (sinx)/x pulses.

Picture taken from Stereophile's homepage showing difference in DSD vs 24-bit PCM.



T

[Sir Terrence the Terrible]

In studio apps yes, higher sampling and bits make sense due to easier work.

I don't agree that SACD always have higher dynamics than standard CD. Looking at the highest octace 10-20 kHz noise starts to increase in SACD, at least in all measurements of players I've looked at on Stereophile's homepage. Also I have hard time of how transients can be accurately be reproduced by SACD since the HF content will be buried in noise. Technically, DVD-A would be superior. Also 1-bit systems will have problem to accurately reproduce (sinx)/x pulses.

Picture taken from Stereophile's homepage showing difference in DSD vs 24-bit PCM.



T

24bit PCM is not the redbook CD standard. Try 16bit PCM. We are not argueing DVD-A here, we are talking about redbook CD's transparency. You are getting off of the issue once again. What you should be comparing in terms of noise is 16 bit vs 24bit.

Using this graphic as a way to discredit SACD is a very misleading thing. If you look very carefully the noise is mostly above 20khz. What is not shown is the masking effect of actually program material(music+harmonics). The noise at 20khz is well below the threshold of hearing, and will most likely be masked by the level of the program material. What would have made this graphic more true to life, is to compare that noise with the level of the music. That would have allowed one to determine its audiblity in relationship to the program material itself. A filter inserted at 50khz(which can be inserted from the encoding end) would bring all of these noise levels much down in level, and still capture all of the high frequency energy in muted brass, cymbals, triangles, glocks, and other instruments rich in high frequency energy. The transient response of percussion(a problem with redbook) would be preserved.

This graph looks like the 50khz filter was not engaged during encoding.

[Sir Terrence the Terrible]

[QUOTE=mtrycraft]I personally don't think we are going to settle this question on this board.

That is correct. But, even if we did, would the rest of the audio world listen? Of course not.

I do not think you are qualified to make a decision in advance for a whole industry. You have no idea whatsoever if they would listen of not. Now that is fact!


And it seems that we have been reduced to discussing small irrelevant issues. You and Mtry seem to think that 16/44.1khz is just fine, me and many in my line of work don't.

Yes, at the consumer end, not mastering, it is fine until there is evidence, credible evidence, for audible differences solly due to the sampling and bit differences, not some other factor introduced in the chain by someone.
Oh, we are not alone by a long shot, here, maybe.

Well since we know(at least some know) that downconverting from 96khz sampling to 44.1khz is problematic, and even doing simple downcoversion from 88.2khz to 44.1khz produces some degredation, that in and of itself would make redbook inadequate if a higher sampling rate was used for recording and post production. So there is no benefit of mastering at a higher level if redbook is the final destination.

I am not here to prove anything to anyone, so I think this discussion has run its course.

But, being in the inductry I would think you of all people would have the necessary credible evidence to support what you have been claiming.

Listening is my credible evidence. Since you refuse to do so, and think a white paper can tell someone what they hear, then there is nothing further to add to this, right?

Since you pick and choose what you deem is credible, then whether I am in the industry or not is irrelevant in term of coming up with credible evidence. You think every supporter of higher sampling rates are hearing things, prone to bias, and various other bull, yet you listen to the music they mix. Interesting.....

[Thomas_A]

24bit PCM is not the redbook CD standard. Try 16bit PCM. We are not argueing DVD-A here, we are talking about redbook CD's transparency. You are getting off of the issue once again. What you should be comparing in terms of noise is 16 bit vs 24bit.

Using this graphic as a way to discredit SACD is a very misleading thing. If you look very carefully the noise is mostly above 20khz. What is not shown is the masking effect of actually program material(music+harmonics). The noise at 20khz is well below the threshold of hearing, and will most likely be masked by the level of the program material. What would have made this graphic more true to life, is to compare that noise with the level of the music. That would have allowed one to determine its audiblity in relationship to the program material itself. A filter inserted at 50khz(which can be inserted from the encoding end) would bring all of these noise levels much down in level, and still capture all of the high frequency energy in muted brass, cymbals, triangles, glocks, and other instruments rich in high frequency energy. The transient response of percussion(a problem with redbook) would be preserved.

This graph looks like the 50khz filter was not engaged during encoding.

Well if you look at the previous poster, it was more dealing with DSD vs. PCM in studio applications. And if you truly want to capture the nature of the music such a closed-miced trumpet (or symbals) which may have HF content up to 100 kHz and where the 30-50 kHz energy is 50-60 dB below the highest level harmonics of the trumpet signal, SACD will have problems. If the highest harmonic of a trumpet signal would be at -10 dB in a song, the HF harmonics between 30-50 kHz would be at at -60 to -70 dB, close to the noise of SACD. This will also go for transient information. Applying filters is not nessecary if you have PCM 24/192 - it will capture all of it.

Looking at the region 10-20 kHz, SACD is no better than standard redbook, only below 10 kHz.

As you said before, the issue about audibility is not going to be solved here, so why should I continue to talk about redbook CD vs. higher resolution media? You have not and don't want to provide any other data than your own experience, and there is not much backing up your claims. Thats fine, but what is there to discuss? Do you want to continue or not?

[Woochifer]

As you said before, the issue about audibility is not going to be solved here, so why should I continue to talk about redbook CD vs. higher resolution media? You have not and don't want to provide any other data than your own experience, and there is not much backing up your claims. Thats fine, but what is there to discuss? Do you want to continue or not?

Well, if this issue means that much to you, then why not setup your own test? I typically opt out of these discussions because they're meaningless unless you have access to a high res master tape for comparison. Discussing the merits and dismerits of a format based strictly on listening to commercially distributed CDs and other music discs yields no meaningful conclusions unless you have a master available for comparison, or you produce your own master and run it through the downsampling cycles. Asking for proof as if everybody who works with digital audio equipment for a living has the time and inclination to go write their own academic papers is an equally meaningless spin job.

I used to do my own blind tests for various cassette tapes so that I could adjust the tape bias level to best match the source material. This was for my own dubbing purposes, and not meant for publication or peer reviewed scrutiny. Setting up my deck for the most transparent sound to the source was my goal, not trying to satisfy the demands of naysayers.

Like I said, if getting to an actual answer, as opposed to raising doubts about someone else's conclusions, is your actual end goal, then why not setup your own test?. With a PC, a microphone, a soundcard, and piano or other instrument of your choosing, you can easily record your own high res digital master and then put it through the downsampling cycles. If you hear any difference in the downsampled playback, then you have your answer. If you can't detect any difference, then you have your answer. Sitting back and asking for proof is nothing more than giving the appearance of having an answer by spinning the same question over and over.