Is 96khz better than 48khz for multitrack digital recording?

[simonv]

Originally Posted by The Byre

Now, if you double the sample rate to 96kHz (or 88.2kHz for a CD) when making the original recording, then all that distortion taking place between say, 14kHz and 22kHz is moved to between 28 and 44kHz - i.e. out of hearing range.

I think that's it.
I'm not exactly sure about what I'm gonna say, but a few years back while studying Electroacoustics I had learned that if your sample rate is 44.1 and you pick up a frequency of 22+khz, it will be mirrored down in the audible range somehow..

For example, if you get a frequency of 30khz (which is 8khz above 22khz), it will be heard at 14khz (8khz below 22khz). This being caused by the converters misinterpretting the signal. I had tested it with a sine generator going through convertors.. I went up and up and up and up and up in the frequencies, and eventually the more I went up, the more the frequency went down.

By using 96khz, you increase your limit to a frequency of 48khz.

Maybe I'm wrong..
But to stay on topic, I definitely hear a difference between 44.1 and 96, and 96 wins!

[Ethan Winer]

Originally Posted by synthoid

It is pretty reasonable to think that you might need additional headroom during recording, in terms of dynamic range, in terms of distance between the frequencies you are recording and the Nyquist frequency, etc., even though the final result will not be in the same format.

Modern digital at 44/16 is infinitely better than even the finest $80,000 Studer tape recorder. What, that's not good enough for you?

mahasandi nailed it when he said "Its all bull. Make a good record however you do." It really is all bull.



--Ethan

[Ethan Winer]

Originally Posted by jamwerks

I'm sure there are other studies that "prove" the contrary.

Let's see them! And I mean real studies, not some audiophile magazine writer's opinion.

if they were played back through certain monitors that make everthing sound pretty (e.g. Genelec's) or on low end monitors. Or if the convertors were not top of the line, or even cableing ...

Yes, and the room you listen in matters far more than all this tweak-head stuff.

I'm still not convinced by just one artical

What exactly would it take to convince you?

--Ethan

[Ethan Winer]

Originally Posted by Melodioso

I remember once comparing two sound files, going back and forth between the two solo'ed files, and finding a subtle but yet present and noticeable difference... only to realize later the files were the same!

Yes, I've done that, and so has everyone else who's been doing this stuff for a while. This is exactly the point that needs to be hammered home again and again!

--Ethan

[wirerecording]

[quote=Ethan Winer;1723181]Modern digital at 44/16 is infinitely better than even the finest $80,000 Studer tape recorder. What, that's not good enough for you ?

ummmmm I guess if you define "better" as "more accurate" you could make your argument, but my experience tells me that a great studer 800 or even 827 sounds "better" (to me) than any digital. before digital i thought i wanted accurate, now it's clear that I don't always want maximum clarity.

If music were only science, science would be the final say, but luckily that is not the case. Intellect and music tends to be a boring combination

[BrianT]

Originally Posted by Ethan Winer

Modern digital at 44/16 is infinitely better than even the finest $80,000 Studer tape recorder.

Now, Ethan.........

That's a hyperbolic statement, and subjective to boot.

As a creative tool, tape's nonlinearity has a great deal of merit in the real world. There are just too many successful producer, engineers and artists in agreement there. "Better" is an inherently subjective evaluation in the case of music.

So if you drop the word "infinitely" and replace the word "better" with the word "more accurate", you have yourself an objectively supportable statement.

Just tryin' to help you maintain your street cred, bro.



P.S. I can't believe the post above me said the same thing as me while I was typing. Cool.

[Quantumphysics]

For the past week I've run several test songs using mixing ITB - OTB - multitracking at 44.1 - 96K - anyway you get my drift. I appreciate all the great information (perceptions) on these types of threads. It is all subjective - and really boils down to what does the individual prefer. I read people buying and selling gear just to find the mojo. The other day me and a friend were going through my archives of the past decade of my recordings. Now - most of us know arrangment - dynamics/mix is critical to the mojo factor. But even in some of the rough mixes the old (cheese I know) Tascam TSR8 and mackie 1604 vlz (not xdr) came up with the most depth and girth. Willing to spend the money to do major upgrades - first I wanted to see if I could find the mojo in the digital rig that I have in the TSR rig. This motivated me to to do the tests.
Music type - rock - funk - real drums and amps - no DI except bass guitar.

Digital rig - Samplitude 7.23a - RME ADI-2 clocking a delta 1010
Analog - Peavey VMP2 - FMR RNP - RNC - RNLA and Mix Wizard 2
AKG 451B's - AT4033 - and sm57's and 58's (I said it was modest)
Monitors NS10m's - Celstion SL600si (audiophile speakers) - Sony MDR7506 - car stereo - and various other systems.

Before going out and buying tons of new gear (I'm not in this for the money - maybe 4 clients a year - mostly my own material) I painstakingly ran the tests and the results were as follows (probably no suprise for some of you)

Low level multitracking at 96K 32 bit float (the 1010 sounds pretty good at 96K clocking from the RME)
Mix into 10 channels of the Mix Wizard (coming in at the insert - still 10Kohms - bybassing the mic pres)
Print hard (not slammed) to two inside tracks (no DBX) of the TSR8.
Bring back into Samplitude via RME at 44.1K 32 float.
Master in the object - pwr3 dither when burning cd.
The major drawback (which I'll overcome the more I do this) is I can't monitor in real time whats coming off the tape. But in time I get used to this analog processor called tape.
It's seems every change that is made in digital takes aways some of the mojo.
For my type of mind - mixing while not looking at a computer screen is more fun and the outcome is more musical (I listen more - one sense compared to two uses less "intuitive resources in the brain)

So when stacking tracks - is 96K "better" than 44.1K - it's a matter of opinion - and though the western mind loves exacting formulas on paper - there is a reason why some recordings don't fatgue the listener. I firmly believe our BRAIN can percieve frequencies well above 20K and it knows what's up there - so I'll take 96K OTB to tape (infinite sampling rate) any day.

Now I just need to get a good 1/2 inch two track mix down machine (see I knew I'd still want to buy something) It's all about the chase - us humans.

[not_so_new]

Originally Posted by Quantumphysics

For the past week I've run several test songs using mixing ITB - OTB - multitracking at 44.1 - 96K - anyway you get my drift. I appreciate all the great information (perceptions) on these types of threads. It is all subjective - and really boils down to what does the individual prefer. I read people buying and selling gear just to find the mojo. The other day me and a friend were going through my archives of the past decade of my recordings. Now - most of us know arrangment - dynamics/mix is critical to the mojo factor. But even in some of the rough mixes the old (cheese I know) Tascam TSR8 and mackie 1604 vlz (not xdr) came up with the most depth and girth. Willing to spend the money to do major upgrades - first I wanted to see if I could find the mojo in the digital rig that I have in the TSR rig. This motivated me to to do the tests.
Music type - rock - funk - real drums and amps - no DI except bass guitar.

Digital rig - Samplitude 7.23a - RME ADI-2 clocking a delta 1010
Analog - Peavey VMP2 - FMR RNP - RNC - RNLA and Mix Wizard 2
AKG 451B's - AT4033 - and sm57's and 58's (I said it was modest)
Monitors NS10m's - Celstion SL600si (audiophile speakers) - Sony MDR7506 - car stereo - and various other systems.

Before going out and buying tons of new gear (I'm not in this for the money - maybe 4 clients a year - mostly my own material) I painstakingly ran the tests and the results were as follows (probably no suprise for some of you)

Low level multitracking at 96K 32 bit float (the 1010 sounds pretty good at 96K clocking from the RME)
Mix into 10 channels of the Mix Wizard (coming in at the insert - still 10Kohms - bybassing the mic pres)
Print hard (not slammed) to two inside tracks (no DBX) of the TSR8.
Bring back into Samplitude via RME at 44.1K 32 float.
Master in the object - pwr3 dither when burning cd.
The major drawback (which I'll overcome the more I do this) is I can't monitor in real time whats coming off the tape. But in time I get used to this analog processor called tape.
It's seems every change that is made in digital takes aways some of the mojo.
For my type of mind - mixing while not looking at a computer screen is more fun and the outcome is more musical (I listen more - one sense compared to two uses less "intuitive resources in the brain)

So when stacking tracks - is 96K "better" than 44.1K - it's a matter of opinion - and though the western mind loves exacting formulas on paper - there is a reason why some recordings don't fatgue the listener. I firmly believe our BRAIN can percieve frequencies well above 20K and it knows what's up there - so I'll take 96K OTB to tape (infinite sampling rate) any day.

Now I just need to get a good 1/2 inch two track mix down machine (see I knew I'd still want to buy something) It's all about the chase - us humans.

Oh no... here it comes. The great "We can hear / we can't hear + 20K frequencies debate..."



I disagree with a lot of what you posted Quantumphysics, sorry. I am not even going to touch the magical "we can hear above 20 K" argument even when there are no substantiated scientific research findings to prove it and none of your other capture or playback gear can reproduce these frequencies in the first place.

Your assertion that a tape machine has an infinite sample rate is a little misleading. An ATR DOES in fact have a high frequency cut off, there is a top end to what can be captured and played back.

You are implying in your argument that the sample rate is about frequency response (which is not the whole picture). With that are the basis for your argument, comparing frequency response between digital and analog, the sample rate does not matter because both machines are limited on what they can reproduce above 18 k or so.

A tape deck using your argument is NOT a infinite sampling rate device. I know what you are getting at and I know the argument but your definition is not making that argument.

On top of that, if analog tape decks are all about capturing this mystical high ultrasonic information why do so many 60's, 70's and 80's recordings sound so good but have little to no high end content above 16 k or so?

Hummmm?

It seems to me that the people complaining about digital (and how did this turn into a digital vs. analog thread again) are 1) using bad digital 2) using analog techniques in the digital world 3) are fooling themselves.

Look, I LOVE analog tape but the reason for that is because it is NOT accurate. I feel people complain about digital and lack of depth (which I fight with as well) because digital is too good, not because it is limited. That is why some people choose 44.1, it helps knock back that "perfection" in digital.

Anyway, I don't want to rant but there are so many things I disagree with in your post there Quantumphysics. It's all good, we all have different opinions about this stuff, that's what makes the world go 'round.

[Quantumphysics]

- not so new - no need to apologize for disagreement - that's what these forums are about for me - learning from others and sharing ideas. I am not attached to the results of my most recent tests but it was fun (and I'll probably change my mind next major upgrade) and now I can get back to music.
I will add that science has been proven wrong (the earth is flat) more times than proven right. So I always go with my gut before science or scopes.

[not_so_new]

Originally Posted by Quantumphysics

- not so new - no need to apologize for disagreement - that's what these forums are about for me - learning from others and sharing ideas. I am not attached to the results of my most recent tests but it was fun (and I'll probably change my mind next major upgrade) and now I can get back to music.

I will add that science has been proven wrong (the earth is flat) more times than proven right. So I always go with my gut before science or scopes.

Errrrr.... are you saying that science is wrong and the Earth IS flat?




Or are you saying that science said the Earth was flat at one time?

Because if it is the latter then there are 2 points for you to consider.

1) "Science" never said the Earth was flat. People who didn't use proper scientific methods may have said the Earth was flat but that was not a scientific finding. The Earth was only proven to be sort of round (more pear shaped actually) when proper scientific methods were actually applied. Chalk one up to science, it's not a case of science being "wrong" at all.

2) There ARE cases when science is wrong about something. BUT, the BIG mistake most people make (and the mistake you are making here) is that science is self correcting. Among other things the scientific method asks that the test in question be repeatable and verifiable. If the test is not repeatable the findings are suspect. If new evidence comes to light that changes the original findings science is MORE than willing to change it's findings based on the new evidence.

The problem is, people see this as flip flopping. It's NOT, it's all about creating a body of base knowledge and building upon it. Science is not like some politician that sticks with a stance because it sounds good on a news clip. Science is about finding empirical evidence to back up the exploration of our universe, mistakes will be made along the way but they will be cleaned up. If that were not the case all science would fall in on it's self because nothing would work as expected.

Sorry for the rant. In today's anti-science environment I feel the need to fix incorrect assumptions about the field.






Oh and.... as far as science being proven wrong more that it is proven right... can you please back that up with some statistics? Because I don't believe that at all (it goes back to that verifiable evidence thing I was talking about above).

[Quantumphysics]

dude - I don't want to use my time to argue with you (or dilute this thread) - there is good music to be written.
If you have time check out "What the Bleep do we Know" it's a cool movie - something different.
We all have beliefs - and then there is the truth - and we all have our own truths.

[barefoot]

Originally Posted by High Fidelity

Many parameters can biased one way or another a listening test:

- Having visual cues of materials being played, equipment usde, etc.
- Not using A/B/X listening procedure ... double blind setup
- Listening with already an answer in mind about the listening 'dilemma'
- Untrained listeners or lack of listening experience
- Lack of objective vocabulary for subjective listening qualification
- Levels
- Low resolution or biased audio equipment in especially poor room acoustics and setup
- Who has the highest authority*

Unfortunately a lot can go wrong during audio tests...

If you get a chance to read the actual Journal article, you'll see they conducted double-blind (A/B/X) tests with 60+ audiophiles and recording professionals using a variety of high-end equipment in different facilities. All they tested for was whether or not listeners could identify a sound difference when a 16/44.1 A/D/A unit was switched into the signal path from high-resolution sources (SACD/DVD-A) playing high-resolution recordings. They got 276 correct answers out of 554 trials. Statistically that's the same as if participants had flipped a coin. They tried to parse out the data to see if they could find any statistical correlations with age, sex, listening experience or upper frequency hearing limit. None were found.

The only clear difference they found was when they raised the gain +20dB above the nominal level with no music signal present. Then the participants could consistently identify the higher noise floor associated with the 16/44.1 A/D/A unit. However, listening to actual music at those levels was uncomfortably loud.

I wouldn't say this is a definitive experiment. But it certainly puts a heavy burden of proof on anyone who claims that higher resolution playback is significantly different from 16/44.1.

Originally Posted by Quantumphysics

If you have time check out "What the Bleep do we Know" it's a cool movie - something different.

Saw it. That was a great comedy! I laughed the whole way through! Seriously, the connections that movie tried to make between science and spirituality were silly at best.

Thomas

[not_so_new]

Originally Posted by Quantumphysics

dude - I don't want to use my time to argue with you (or dilute this thread) - there is good music to be written.

Cool.... I was only commenting on what you brought up.

If you have time check out "What the Bleep do we Know" it's a cool movie - something different.

Oh, that.... errrr..... "movie" huh? There are so many glaring wrong assumptions and outright untruths to that film it amazes me that people point it it as a source of logical debate. But it does point out the "anti-science" thing I mentioned above. People don't actually take the time to research what they believe to be true they just believe it because someone said it was so.

Here, read this if you get a minute.

What the bleep. - tribe.net

We all have beliefs - and then there is the truth - and we all have our own truths.

You don't want to get into it and that is cool. I will just say that your above point is fundamentally flawed and leave it at that.

Rock on dude.

And so I don't get blamed for taking a thread that really has no answer off topic, I like to record at 96... but I said that already above.

[miqer]

If you use a lot of plugins, like waves ssl eq etc... plugs that add coralation etc., does it then make more sense to track at 96? For the plug has a more accurate signal to work with?

miqer

[Dan Lavry]

Originally Posted by MickeySmid

Thanks, Dan!!... Is it not also then wizer to record higher than 44 if we run a lot of plugins in our DAW? These can then also procces the 96khz recordings and add things like corolation (vintage warmers etc), and get a better result like you state above...

So if we use a lot of plugs it's def. best to record above 44?!

(This is an endless discussion... I know. Somehow I trust mr. Lavry... !!!

thanks,
Mickey

Hi Mickey,

Thank you for the nice comment.

As a rule, digital processing is very transparent for linear processing. As a rule, alias distortions show up when doing digital non linear processing, such is the case unless the process was designed to be free of aliasing, and the AD10 digital alias free emulation ™ is the only unit around to offer that. To achieve alias free distortions, one needs to work at higher sample rate, but that is only one little requirement. Unless the process is specifically designed to be alias free, going to higher sample rates may help some, but will not eliminate the aliasing.

IN ANALOG, any non linearity generates high frequencies, and such non audible energy can be filtered out (to prevent it from impacting gear and speaker performance).

IN DIGITAL, any non linearity “tries” to generate high frequencies, but as we know, when the frequencies exceed Nyquist, they “fold back” into the audible range. Aliasing can not be filtered out. By the time it happened it is already too late to do anything about it. Unlike analog, doing something "later" is not an option.
At first one may think that moving Nyquist up will solve the problem, but such approach proves very disappointing. Why? While the higher frequency harmonics do tend to have reduced amplitude, the amplitude reduction is pretty small for the ear, which is near logarithmic in its response to amplitude. A 1% on a scope is difficult to see, but for the ear it means only -40dB…

So digital is great for accurate summing, EQ, reverb… and all linear processing. But digital is terrible for limiters, tube and transformer emulation and other non linear processing (unless it is customized to work without aliasing).

What is the definition of linear process? For someone with math background it is a simple equation. I will tray to explain it without math, in a somewhat loose manner:

Say you wish to run some process on some some entity.
You can process the whole entity and get some outcome, lets call it outcome A.
Now lets break the entity into two or more pieces, and run the same process on each of the pieces. After doing so, lets combine (add) all the individually processed pieces, and we end up with an outcome B.
If the outcome A is the same as outcome B the process is linear. If, however, the outcome A is different then B the process is not linear.

Examples:

Say I have 2 channels and I boost the gain of each channel by 3dB, then I sum them into one channel. The result will be the same if I added the channels first, then boosted the combined single channel by 3dB. Thus the process of gain and attenuation is a linear one, and works well with digital processing.

Now lets take 2 channels and run them through a tube. The combined signal has larger range (think of the peak to peak span), causing the combined signal to “bent” more. The same with a transformer, where the impact is greatest when you “push the device” towards full magnetization. When running each channel at a time, the levels are lower and the coloration is much lower as well. When you add 2 channels with low coloration in each, the total will have less coloration compared to the single combined larger signal.

So a linear process is when the “sum of the processed parts” is the same as the “processed sum of the parts”. That includes EQ and reverb. Tube and transformer emulation is non linear and so on.

Most DAW processing is linear (mixing, EQ, reverb …) so DAW work is very fine, be it at 44.1KHz or higher. But regarding digital non linear processing, unless the process was very deliberately crafted to avoid or greatly reduce aliasing, I would look for an analog method.

Regards
Dan Lavry

[Ethan Winer]

Originally Posted by BrianT

replace the word "better" with the word "more accurate", you have yourself an objectively supportable statement ... Just tryin' to help you maintain your street cred, bro.

Heh, I got too much street cred. Want some of mine?

Seriously, from my perspective more accurate is better every time. That's why they call it High Fidelity!

Hey, you want color? Get a fuzz box. And no need to spend thousands on some boutique device with toobs and/or or transformers. Glue and color in the form of distortion is easy to create.

Of course I knew my post would ruffle some feathers, but that's my job.

--Ethan

[Duardo]

If you use a lot of plugins, like waves ssl eq etc... plugs that add coralation etc., does it then make more sense to track at 96? For the plug has a more accurate signal to work with?

Tracking and mixing at 96 kHz may sound better than doing so at 44.1 kHz, but not necessarily because you're working with a more accurate signal...whether the signal is more accurate or not depends a lot on the specific converters used, as has been pointed out several times in this thread already...but there certainly are processors that sound better at higher sampling rates than lower.

Still, that doesn't mean that 16/44.1 isn't entierly sufficient as a delivery format...I believe that it is as well, and while I wouldn't use words like "incontrovertable" myself I think the test mentioned certainly is a telling one. I still don't understand why everyone wants higher-resolution formats to sound better. I want 44.1 kHz to keep getting better and better, as it has done thus far...

[scott petito]

Well I have an interesting story to tell..... an Audiophile Label recently licensed one of my recordings from the late nineties... at the time my main studio was setup remote to make a recording with the "Band" So I tracked the record in question on 3 blackface adats with a few gml pres at 16 bits, 48K in my spare time........When the company asked to do the license I explained that it was a 16 bit recording 48K mostly recorded in my living room and mixed to dat on a First generation o2R... they did not care about that...
.in truth it is a very good sounding recording that several mastering engineers have told me they use as a reference CD to this day... who know why but all the stars lined up and it just came out really great....

anyway the Audiophile label took the master on 16 bit CD bumped to both DSD and SACD and are selling a boatload of them to audiophiles.....( in fact they just went to a second pressing) the reviews of the sonic quality in all the audio mags have been stellar... I just kinda laugh at the whole thing knowing the origin of the source material..... So get the best players, great singer, good arrangements, and do careful work and forget about all this sample rate crap....

cheers
SP

[barefoot]

Originally Posted by barefoot

But it certainly puts a heavy burden of proof on anyone who claims that higher resolution playback is significantly different from 16/44.1.

I forgot to add one caveat:

This experiment is obviously subject to the limitations of current playback technology - especially loudspeakers. Perhaps if we develop loudspeakers that can effortlessly reproduce 120dB SPL at the listening position with very low distortion (0.001% ish), then high-resolution recordings will be readily discernable from regular old 16/44.1. We'll see..... emm... hear.


Btw, the authors of that article don't say that high-resolution recording isn't useful on the front end of the recording process, considering that signal processing and such can degrade things considerably. Hi-res files are common in graphic design and digital photography for just this same reason, even when they know the end result will be low-res computer images.

Thomas

[mdme_sadie]

I think a more interesting question is - What impact does the higher frequency have on summing (and how that affects the overall result)?

After all we're talking abut not just 2 tracks, left and right here - but potentially upwards of 50 or 100 tracks, mixed and phasing with each other. That's perhaps where the article may not completely cover the whole debate as it seems to be mostly just talking about the final output.

[mmcfarlane]

Originally Posted by barefoot

...Perhaps if we develop loudspeakers that can effortlessly reproduce 120dB SPL at the listening position with very low distortion (0.001% ish), then high-resolution recordings will be readily discernable from regular old 16/44.1. We'll see..... emm... hear. ...

Perhaps this will fix many of the arguments in this forum because none of us will be able to hear anything after a few months at 120db

[BrianT]

Originally Posted by scott petito

Well I have an interesting story to tell..... an Audiophile Label recently licensed one of my recordings from the late nineties... at the time my main studio was setup remote to make a recording with the "Band" So I tracked the record in question on 3 blackface adats with a few gml pres at 16 bits, 48K in my spare time........When the company asked to do the license I explained that it was a 16 bit recording 48K mostly recorded in my living room and mixed to dat on a First generation o2R... they did not care about that...
.in truth it is a very good sounding recording that several mastering engineers have told me they use as a reference CD to this day... who know why but all the stars lined up and it just came out really great....

anyway the Audiophile label took the master on 16 bit CD bumped to both DSD and SACD and are selling a boatload of them to audiophiles.....( in fact they just went to a second pressing) the reviews of the sonic quality in all the audio mags have been stellar... I just kinda laugh at the whole thing knowing the origin of the source material..... So get the best players, great singer, good arrangements, and do careful work and forget about all this sample rate crap....

cheers
SP

What was it mixed through and to, if you don't mind?

[synthoid]

Originally Posted by Ethan Winer

Modern digital at 44/16 is infinitely better than even the finest $80,000 Studer tape recorder. What, that's not good enough for you?

I have a 16-bit / 44.1KHz Tascam Portastudio 788 with original packaging and power supply that I will happily trade you for your outdated Studer. PM me.

-synthoid

[innesireinar]

Originally Posted by scott petito

Well I have an interesting story to tell..... an Audiophile Label recently licensed one of my recordings from the late nineties... at the time my main studio was setup remote to make a recording with the "Band" So I tracked the record in question on 3 blackface adats with a few gml pres at 16 bits, 48K in my spare time........When the company asked to do the license I explained that it was a 16 bit recording 48K mostly recorded in my living room and mixed to dat on a First generation o2R... they did not care about that...
.in truth it is a very good sounding recording that several mastering engineers have told me they use as a reference CD to this day... who know why but all the stars lined up and it just came out really great....

anyway the Audiophile label took the master on 16 bit CD bumped to both DSD and SACD and are selling a boatload of them to audiophiles.....( in fact they just went to a second pressing) the reviews of the sonic quality in all the audio mags have been stellar... I just kinda laugh at the whole thing knowing the origin of the source material..... So get the best players, great singer, good arrangements, and do careful work and forget about all this sample rate crap....

cheers
SP

I still have the StudioSound review where Roger Nichols remixed on the 02R a SteelyDan song he did on a Neve VR72 and by comparing the two mixes he said that the one done on the Yamaha was more clear and more detailed. At time I read many comments about how many $$$ Yamaha gave to Roger for saying this.
BTW I bought the 02R achieving better mixes compared to those done on analog stuff. It was 1996.

[bob katz]

Originally Posted by severe

Why not the whole word?

"Meyer and Moran do not say that 14 or 15 bits in a truncated CD are just as good as 20. What they say is that spot-on 16-bit/44.1-kHz processing is as good as it gets, audibly."

Love that part.

What these tests seem to ignore (whether they are valid or not is another subject) is the cumulative losses that occur in a production chain. Using a final SACD disc as a source for their tests is potentially questionable as it is not an original source. Potentially, every step in a production chain involves losses, each one of which may be very subtle to the ear, but when auditioned cumulatively, produce an obvious loss (at least to critical listeners).

Therefore, even if for all practical purposes, a listening test seems to show that copying an SACD to a 44/16 medium seems to sound the same*, THIS DOES NOT MEAN that if you take an original recording, mix it at 4416 (dithered), send it to a mastering house, have it processed and mastered, and end up at 4416, there will be no audible degradation! There is a big difference here. And my contention (proven to me many times over many years) is that the HIGHER THE RESOLUTION OF THE SOURCE, AND THE HIGHER THE RESOLUTION OF THE PROCESSING SYSTEM, THE LESS THE PERCEIVED LOSS OVER CUMULATIVE GENERATIONS OF PROCESSING.

And the correlary to this:

Even if each step in a cumulative chain seems to pass the transparency test to the listener, it is amazing to discover that at the end of that chain, there are audible losses. This seemingly contradictory conclusion can be explained by realizing that if step 1 in the chain has a loss which is not noticeable or audible to the listener, and step 2 has a similar loss, the two together can add up to an audible loss!

This "accumulation" is the main justification for using high resolution (e.g. 96 kHz/24 bit) in the origination medium if transparency is desired for the final result. This works well for types of music (such as acoustic jazz, classical, etc.) that likes transparency and little distortion. Especially if performing any dynamics processing in the digital domain. But the same goes for analog processing. Even without any analog processing in between, around here, with matched levels, we can HEAR a subtle difference between the source and passing that source through the world's best D/A/D chain. Even if Brad Meyer "proves" that he cannot, I wager that in a cumulative chain of such "tiny" or even "inaudible" losses----there will be a meaningful audible loss at the end of that chain.

*I'm not saying I agree with this statement, not having taken the blind test, but I'm certain that if carried out properly, with matched levels, if there is a difference, it will be extremely subtle.

[Ethan Winer]

Originally Posted by bob katz

Potentially, every step in a production chain involves losses, each one of which may be very subtle to the ear, but when auditioned cumulatively, produce an obvious loss (at least to critical listeners).

Absolutely Bob. Of course, the cumulative loss can be measured and assessed. I've read of reports where a signal was passed through ten or twenty A/D/A conversions as a test. Wasn't there something like that at the recent AES show in New York?

--Ethan

[Ethan Winer]

Originally Posted by scott petito

who know why but all the stars lined up and it just came out really great.

Agreed, and I think this is why many people buy into the various hi-res BS claims. When I used to record professionally there would be groups that sounded great immediately, and groups that sounded lame no matter what I did. One of my favorite memories is of a big band session - they played their butts off and sounded amazing in the studio and in the control room.

So Joe Blow buys a new outboard word clock or whatever and the next group he records happens to come out great. Joe concludes the word clock improved the sound.

--Ethan

[severe]

Originally Posted by bob katz

What these tests seem to ignore (whether they are valid or not is another subject) is the cumulative losses that occur in a production chain...

Thanks for that, Bob.

Originally Posted by Ethan Winer

So Joe Blow buys a new outboard word clock or whatever and the next group he records happens to come out great. Joe concludes the word clock improved the sound.

Agreed. (including all the other variables that we speak of on a daily basis, from mic placement to humidity)

Originally Posted by innesireinar

BTW I bought the 02R achieving better mixes compared to those done on analog stuff. It was 1996.

And now you're here and everything's gonna be o.k.

[barefoot]

Originally Posted by mmcfarlane

Perhaps this will fix many of the arguments in this forum because none of us will be able to hear anything after a few months at 120db

Well, ideally that would just be used as headroom for extremely dynamic recordings.

.... ideally

[Priscilla]

Originally Posted by Ethan Winer

Absolutely Bob. Of course, the cumulative loss can be measured and assessed. I've read of reports where a signal was passed through ten or twenty A/D/A conversions as a test. Wasn't there something like that at the recent AES show in New York?

--Ethan

Hello Ethan,

The manner of accumulation is not all that simple. If someone tells you that they ran conversion 10-20 times, the "casual listner" would expect to have 10-20 times worse result. That is not all that simple. For example, lets take dynamic range limitations, which are due to a random noise floor. Say we use one conversion as a reference point, calling it the 0dB referance point. Then the degradation in dB for 2-20 converters is:

2 converters -3.01dB
3 converters -4.77dB
4 converters -6.02dB
5 converters -6.99dB
6 converters -7.78dB
7 converters -8.45dB
8 converters -9.03dB
9 converters -9.54dB
10 converters -10.00dB
11 converters -10.41dB
12 converters -10.79dB
13 converters -11.14dB
14 converters -11.46dB
15 converters -11.76dB
16 converters -12.04dB
17 converters -12.30dB
18 converters -12.55dB
19 converters -12.79dB
20 converters -13.01dB

So the biggest loss is the first convertion. The second converter brings about a -3dB, which is the same as the COMBINED last 10 conversions in the chain. In fact, the difference between say convertions 19 and 20 is tiny. Now, if you compare one converter to say 16 converters, you have 12 dB "jump", which is 2 bits loss. And if you just add one converter at a time, starting at 10 converters and ending at 20 converters, the difference is 1/2 a bit.

To add "insult to injury", if the converter is say and AD10 with 120dB "A" weighted, 117dB unweighted, and the signal presented to it is say 100dB noise floor, even 20 converters will be less noisy (117-13=104dB) then the signal, and the total loss is tiny (1.5dB). However, if the signal presented to the converter is 117dB noise floor, then even one converter will cost you 1/2 bit...

When it comes to distortions, the accumulation may or may not be faster, it does depend on many factors, beyond this post. It is very unlikely to have 20 conversions yield 20 times the distortions (26dB higher distortions). For that to happen, you must have "all the stars lined up", and that does not happen in the real world.

But my argument was regarding dynamic range (in dB and bits). If you do one conversion, it is that conversion you want to pay attention to. If you have more then one, it is the worst conversion in the chain that will tend to dominate the limitations. To be fair, my above list does assume that the noise is random, or at least different between converters. If all converters share some "noisy tone", the accumulation will be faster. It is not easy to cover all cases which are gear and setup dependent.

But I do agree with you that if the music is bad, nothing will help.

Regards
Dan Lavry