A couple of riffs on the notion of "bit perfect"...

[spdif-usb]

Here is another (IMO great) article to discuss the benefits of upsampling versus oversampling.

Upsampling vs. Oversampling for Digital Audio — Reviews and News from Audioholics

[Jud]

Here is another (IMO great) article to discuss the benefits of upsampling versus oversampling.

Upsampling vs. Oversampling for Digital Audio — Reviews and News from Audioholics

 

I think it provides firm support for the idea that one wants as high as possible an ADC sample rate relative to the sample rate at the DAC reconstruction filter, up to 1:1. The one advantage it cites for "upsampling" (interpolation) is reduction of jitter through asynchronous sample rate conversion, which is pretty well yesterday's jitter reduction scheme since Gordon Rankin came along with asynchronous USB input code. Otherwise, it's solidly in praise of "oversampling" (high sample rate at ADC).

[spdif-usb]

Please note this article was written in 2004. Upsampling was still fairly new then. Eight years later, we have DAC chips that can upsample 192 kHz material to 1536 kHz, apply noise shaping, are 100% jitter-immune, use internal DSP with a system clock as high as 40 MHz, and are already becoming mainstream very rapidly. Upsampling in sigma delta DACs does an excellent job when combined with oversampling. Would you rather be downloading FLAC files from HDTracks that are using a 1536 kHz samplerate? :lol:

 

I think it provides firm support for the idea that one wants as high as possible an ADC sample rate relative to the sample rate at the DAC reconstruction filter, up to 1:1. The one advantage it cites for "upsampling" (interpolation) is reduction of jitter through asynchronous sample rate conversion, which is pretty well yesterday's jitter reduction scheme since Gordon Rankin came along with asynchronous USB input code. Otherwise, it's solidly in praise of "oversampling" (high sample rate at ADC).

[Jud]

Please note this article was written in 2004. Upsampling was still fairly new then. Eight years later, we have DAC chips that can upsample 192 kHz material to 1536 kHz, apply noise shaping, are 100% jitter-immune, use internal DSP with a system clock as high as 40 MHz, and are already becoming mainstream very rapidly. Upsampling in sigma delta DACs does an excellent job when combined with oversampling. Would you rather be downloading FLAC files from HDTracks that are using a 1536 kHz samplerate? :lol:

 

You described the article as "great." Now that I like what it says, you're backing off? ;-)

 

I'd have no problem myself downloading high sample rate material. Whether it will one day be 1536 (can just see the adherents of 768 vs. 1536 going at it or something a little lower/smaller, who can say?

[spdif-usb]

You described the article as "great." Now that I like what it says, you're backing off? ;-)

I still think it's a great article, because it explains what the problems were when the tech was still relatively in its infancy. We have come a long way since then, and part of what makes this hobby interesting IMO is that digital audio is getting better and cheaper all the time.

I'd have no problem myself downloading high sample rate material. Whether it will one day be 1536 (can just see the adherents of 768 vs. 1536 going at it or something a little lower/smaller, who can say?

What would you do if the 1536 kHz files you just downloaded turned out to be upsampled from CD and / or have a dynamic range value of maybe 6? ^

[RobbieC]

Jud,

 

I've thought about your question as posed originally and I'd like to offer why I think it has been done the way it's been done. But first, please note that aside from formal mathematical training I have no inside industry knowledge and so am making an educated guess based on reason.

 

What we need to know is what kinds of ADC's are out in the wild, and moreover, which is the most widely used (by the studios). I don't know the answer to that but let's start with what an ADC stage needs to accomplish and that's sample an analog signal without violating Shannon-Nyquist.

 

The way it's done, as I'm sure you are aware, is that the analog signal must first pass through an anti-alias filter before being sampled to ensure spectral content above the Nyquist frequency is adequately attenuated. Since this must be done before sampling, this means that the filter is applied before the information is made digital and if it is, then it is applied to the analog signal meaning that this filter is an analog filter (capacitor, inductors, and such). Analog filters can be made very sophisticated with acceptable phase shifts...for a price.

 

Let's suppose the ADC manufacturer is now faced with the decision of whether to use a sophisticated analog filter or a rather simpler one. The manufacturer would naturally weigh out the pros and cons of both approaches. The performance parameters of greatest concern I suspect are adequate attenuation at/above Nyquist and phase shifts/group delay caused by analog filtering. The two are related: the greater the rate of roll off (dB/octave) the more phase shifting for a given filter. Please note: phase shifts are a reality of analog filters.

 

For the sake of discussion, let's say sophistication is concerned with limiting the extent of phase shifts for a given rate of roll off such that the more sophisticated (and more costly) the design, the less phase shifts for the given roll off rate. Since the concern is of course avoiding discernable phase shifts in the audible band there are now two ways to accomplish that:

 

(1) Use a sophisticated analog anti-alias filter

 

(2) Use a simpler analog filter but sample at a higher rate (oversample)

 

Obviously, the second approach is the one you've mentioned occurs. It's easy to understand why. At 8x, sampling typically works out to 352.8 kHz which provides over three octaves (i.e., 20 to 40, 40 to 80, 80 to 160, plus a little more since the Nyquist is 176.4 kHz) out beyond audible in which to roll off high frequencies. That allows a much gentler filter to be used than the one needed for 1x which must do its work over less than a seventh of an octave (i.e., 20 to 22.05 so take the seventh root of 2 and multiply 20 by that to show a little less than a seventh of an octave available). So that means the filter used in 1x case is about 21 times steeper for a given depth of attenuation (so say if 96 dB of attenuation is desired then we have 32 dB/octave for the 8x case and 672 dB/octave for the 1x case). In general, the gentler the filter the gentler the phase shifts.

 

Now, 32 dB/octave is still a fairly significant slope and some phase shifting will likely still occur but then there is no reason why the roll off must start at exactly 20 kHz. By starting the roll off later, phase shifts can be reduced further still in the audible band. Anyway, the 8x data still contains frequencies in violation of those allowed for the 1x case. No problem. The data is now in the digital domain (having been sampled at the 8x rate of 352.8 kHz) and can now be easily modified by digital filters, filters that btw can be made to have no phase shifts whatsoever. I mentioned how before in another thread called, "Can a recording be made indistinguishable from the original" IIRC. In fact, someone mentioned how above though not as in depth as I had in that other thread. After digital filtering out all unwanted frequencies it is an easy case of decimation down to the desired 1x sample rate. The decimation is very straightforward: keep every eighth sample, discard the rest.

 

But your question asks why not just leave the data set at 352.8 kHz? Why downsample to some rate only to upsample back to 352.8 kHz? (I've chosen to continue with the 44.1 multiple I started with but you can just as easily apply to a 48 multiple.) I believe the reason lies in the design goals as set out originally. Originally, the goal was to get a sample rate of 44.1 kHz to sound better. The early analog anti-alias filters were introducing phase shifts and group delays that were audible. The two approaches above were considered and the second approach won out since it probably performed better and was cheaper overall. Keep in mind that since the goal was a 1x rate perhaps some decision was made not to capture the 352.8 kHz stream at all. Actual recording of data perhaps only happens after the completion of all filtering including decimation with everything else discarded and lost. With 176.2 kHz and 196 kHz recordings just replace the design objective above with those rates. Still an objective and no reason to record above objective.

 

But still, why downsample only to upsample back? Wouldn't it just be better to keep the originally captured "bit perfect" data than to interpolate data back from a downsampled set? Not necessarily. The reason, I believe, is that the analog anti-alias filter used are not good enough (by design, you know cost consideration) and introduce unsatisfactory phase shifts that while outside the audible range really don't add any value to the data set either so why keep all that data? You just end up needing more storage capacity and the bandwidth to transfer the data to that storage with little real benefit.

 

If an ADC comes around with a truly marvelous analog anti-alias filter then your wanting to keep the original data set at the orignal sample rate, whatever it may be, would then make sense. But then someone could argue: "Hey, wouldn't it be even better if we oversampled and then applied filtering in the digital domain and downsampled back to your desired sample rate?"

[spdif-usb]

RobbieC,

 

If a signal is sampled at a rate much higher than the Nyquist frequency and then digitally filtered to limit it to the signal bandwidth there are the following advantages:

• digital filters can have better properties (sharper rolloff, phase) than analogue filters, so a sharper anti-aliasing filter can be realised and then the signal can be downsampled giving a better result
  
• a 20-bit ADC can be made to act as a 24-bit ADC with 256× oversampling
  
• the signal-to-noise ratio due to quantization noise will be higher than if the whole available band had been used. With this technique, it is possible to obtain an effective resolution larger than that provided by the converter alone
  
• The improvement in SNR is 3 dB (equivalent to 0.5 bits) per octave of oversampling which is not sufficient for many applications. Therefore, oversampling is usually coupled with noise shaping (see sigma-delta modulators). With noise shaping, the improvement is 6L+3 dB per octave where L is the order of loop filter used for noise shaping. e.g. - a 2nd order loop filter will provide an improvement of 15 dB/octave.
  

 

Analog-to-digital converter - Wikipedia, the free encyclopedia

[spdif-usb]

P.S. - Decimation filters used in sigma delta ADCs are not entirely that simple, as this (now 10 years old) document shows.

 

http://inst.eecs.berkeley.edu/~ee247/fa05/lectures/decimation.boser.pdf

 

More info on sigma delta ADCs here.

 

http://www.ti.com/lit/an/slyt423/slyt423.pdf

 

http://www.ti.com/lit/an/slyt438/slyt438.pdf

[RobbieC]

I should have noted that I was talking about R-2R ladder type DAC's and ADC's (since the Burr Brown PCM1704 was mentioned) and not sigma delta types. I was doing so solely for simplicity in relaying my idea to Jud. In the case of the R-2R ladder derived discussion, decimation would indeed be as simple as I said. I am very familiar with the math and am formally trained in it from several universities including Berkeley. Graduated in 2010 so studied some very cutting edge mathematics.

 

What you mentioned about digital filters repeats me some but also would add to another discussion, one on low-bit converters, sigma delta type converters, DSD, etc.

 

Also, I guess I should mention now that there really isn't any such animal as an R-2R ADC's per se. So, in the case of ADC, the closest thing is a Successive Approximation Register type ADC which actually incorporates an R-2R DAC within its design. So I want to note that it was with this in mind that my discussion followed for those who want to research the ideas brought forward.

[spdif-usb]

In the Wikipedia article I linked, the following suggestion has been made:

Most[citation needed] high-profile recording studios record in 24-bit/192-176.4 kHz PCM or in DSD formats, and then downsample or decimate the signal for Red-Book CD production (44.1 kHz) or to 48 kHz for commonly used for radio/TV broadcast applications.

Analog-to-digital converter - Wikipedia, the free encyclopedia

 

I am assuming sigma delta ADCs are the most commonly used in high-profile recording studios, because of the higher ENOB (Effective Number Of Bits) and resolution that can be obtained with them versus the total cost. The high amount of computational power required to minimize the latency problem in sigma delta applications once used to be an important limiting factor, but thanks to DSP growing faster and faster all the time, R2-R ladder has largely been abandoned by the audio industry.

[mayhem13]

Excellent discussion.

[manisandher]

... thanks to DSP growing faster and faster all the time, R2-R ladder has largely been abandoned by the audio industry.

 

Yep, and I think this is a shame. IME, the best digital sound comes from R2R ladder DACs. But not all of the audio industry has abandoned the technology. MSB certainly hasn't. In this review JA says, "In most ways, MSB Technology's Diamond DAC IV offers the best measured performance in the digital domain that I have encountered." (MSB Technology Platinum Data CD IV transport & Diamond DAC IV & D/A converter Measurements | Stereophile.com)

 

Mani.

[spdif-usb]

The problem with measurements is always the fact it's perfectly possible to build a system that measures well, but is not satisfactory. Human hearing is non-linear on alot of levels. There is not a linear relationship between a reduction in digital artifacts (or even analog types of distortions, for that matter) and the improvement in musical satisfaction this reduction in artifacts (distortions) engenders. Therefore, measurements can be used only as an indication, and those who still claim otherwise are just not fairly up-to-date with their knowledge of modern psychoacoustics.

Yep, and I think this is a shame. IME, the best digital sound comes from R2R ladder DACs. But not all of the audio industry has abandoned the technology. MSB certainly hasn't. In this review JA says, "In most ways, MSB Technology's Diamond DAC IV offers the best measured performance in the digital domain that I have encountered." (MSB Technology Platinum Data CD IV transport & Diamond DAC IV & D/A converter Measurements | Stereophile.com)

 

Mani.

[Jud]

The problem with measurements is always the fact it's perfectly possible to build a system that measures well, but is not satisfactory. Human hearing is non-linear on alot of levels. There is not a linear relationship between a reduction in digital artifacts (or even analog types of distortions, for that matter) and the improvement in musical satisfaction this reduction in artifacts (distortions) engenders. Therefore, measurements can be used only as an indication, and those who still claim otherwise are just not fairly up-to-date with their knowledge of modern psychoacoustics.

 

Just want to caution that we want to stick fairly close to the topic of sample rate conversion rather than going off on a number of other interesting and more or less associated topics. I haven't had enough time to send in a couple of responses to some very good posts here on the thread topic, but will try to get to them soon.

[manisandher]

Therefore, measurements can be used only as an indication, and those who still claim otherwise are just not fairly up-to-date with their knowledge of modern psychoacoustics.

 

Without being any sort of expert myself, I tend to agree with this. However, PeterSt has stated that he used both measurement and listening for the development of his Phasure NOS1 DAC (which utilizes eight PCM1704U-K R2R ladder chips by the way) and found a direct correlation between them.

 

Back on topic...

 

I think I've persuaded the guy to whom I sold my PMII to give it back to me for my vinyl digitization project. If/when I get it back, I think recording at 24/192 on the PMII and playing back at 24/192 on the Phasure NOS1 will indeed give me a path with absolutely no conversions whatsoever, as Jud is suggesting. Recording at 24/192, the PMII does NOT use an oversampling filter (the only 24/192-capable ADC I'm aware of that can do this). Playing back at 24/192, the NOS1 does NOT use upsampling or sigma-delta modulation. (Although I suspect interpolating the 24/192 up to 24/768 in XXHighEnd will sound better than feeding the 24/192 natively to the NOS1.)

 

Mani.

[manisandher]

I think recording at 24/192 on the PMII and playing back at 24/192 on the Phasure NOS1 will indeed give me a path with absolutely no conversions whatsoever, as Jud is suggesting.

 

So here's what I mean (in comparison to 'standard' PCM and DSD):

 

 

Mani.

[Jud]

Jud,

 

I've thought about your question as posed originally and I'd like to offer why I think it has been done the way it's been done. But first, please note that aside from formal mathematical training I have no inside industry knowledge and so am making an educated guess based on reason.

 

What we need to know is what kinds of ADC's are out in the wild, and moreover, which is the most widely used (by the studios). I don't know the answer to that but let's start with what an ADC stage needs to accomplish and that's sample an analog signal without violating Shannon-Nyquist.

 

The way it's done, as I'm sure you are aware, is that the analog signal must first pass through an anti-alias filter before being sampled to ensure spectral content above the Nyquist frequency is adequately attenuated. Since this must be done before sampling, this means that the filter is applied before the information is made digital and if it is, then it is applied to the analog signal meaning that this filter is an analog filter (capacitor, inductors, and such). Analog filters can be made very sophisticated with acceptable phase shifts...for a price.

 

Let's suppose the ADC manufacturer is now faced with the decision of whether to use a sophisticated analog filter or a rather simpler one. The manufacturer would naturally weigh out the pros and cons of both approaches. The performance parameters of greatest concern I suspect are adequate attenuation at/above Nyquist and phase shifts/group delay caused by analog filtering. The two are related: the greater the rate of roll off (dB/octave) the more phase shifting for a given filter. Please note: phase shifts are a reality of analog filters.

 

For the sake of discussion, let's say sophistication is concerned with limiting the extent of phase shifts for a given rate of roll off such that the more sophisticated (and more costly) the design, the less phase shifts for the given roll off rate. Since the concern is of course avoiding discernable phase shifts in the audible band there are now two ways to accomplish that:

 

(1) Use a sophisticated analog anti-alias filter

 

(2) Use a simpler analog filter but sample at a higher rate (oversample)

 

* * *

 

But your question asks why not just leave the data set at 352.8 kHz? Why downsample to some rate only to upsample back to 352.8 kHz? (I've chosen to continue with the 44.1 multiple I started with but you can just as easily apply to a 48 multiple.) I believe the reason lies in the design goals as set out originally. Originally, the goal was to get a sample rate of 44.1 kHz to sound better. The early analog anti-alias filters were introducing phase shifts and group delays that were audible. The two approaches above were considered and the second approach won out since it probably performed better and was cheaper overall. Keep in mind that since the goal was a 1x rate perhaps some decision was made not to capture the 352.8 kHz stream at all. Actual recording of data perhaps only happens after the completion of all filtering including decimation with everything else discarded and lost. With 176.2 kHz and 196 kHz recordings just replace the design objective above with those rates. Still an objective and no reason to record above objective.

 

But still, why downsample only to upsample back? Wouldn't it just be better to keep the originally captured "bit perfect" data than to interpolate data back from a downsampled set? Not necessarily. The reason, I believe, is that the analog anti-alias filter used are not good enough (by design, you know cost consideration) and introduce unsatisfactory phase shifts that while outside the audible range really don't add any value to the data set either so why keep all that data? You just end up needing more storage capacity and the bandwidth to transfer the data to that storage with little real benefit.

 

Hi, Robbie, been wanting to ask since I read this: Possible to apply a digital filter that neither interpolates nor decimates, but maintains the same sample rate, to counteract the phase shift introduced by an inexpensive analog filter?

[spdif-usb]

Just want to caution that we want to stick fairly close to the topic of sample rate conversion

...Which is exactly why I posted what I posted in the first place. The very goal of sample rate conversion is, or ought to be, to provide better sound quality (which alot of people say it does).

[Jud]

...Which is exactly why I posted what I posted in the first place. The very goal of sample rate conversion is, or ought to be, to provide better sound quality (which alot of people say it does).

 

There are lots of discussions/arguments regarding the merits of various ways to "provide better sound quality" available in this forum. I want to stick *very* close to the specifics of sample rate conversion and any audible effects of that alone, rather than opening up the thread to generalized discussions of statements like these:

 

measurements can be used only as an indication, and those who still claim otherwise are just not fairly up-to-date with their knowledge of modern psychoacoustics

 

Discussions of the measurement vs. audibility proposition have been had and can still be had in any number of threads on the forum, or in a new thread of your own should you so desire.

[spdif-usb]

Discussions of the measurement vs. audibility proposition have been had and can still be had in any number of threads on the forum, or in a new thread of your own should you so desire.

If your point is to suggest audibility doesn't matter and that therefore sample rate conversion should be avoided, I'll be on my way out of this thread.

[Jud]

If your point is to suggest audibility doesn't matter and that therefore sample rate conversion should be avoided, I'll be on my way out of this thread.

 

Therefore, measurements can be used only as an indication, and those who still claim otherwise are just not fairly up-to-date with their knowledge of modern psychoacoustics.

 

Without being any sort of expert myself, I tend to agree with this. However, PeterSt has stated that he used both measurement and listening for the development of his Phasure NOS1 DAC (which utilizes eight PCM1704U-K R2R ladder chips by the way) and found a direct correlation between them.

 

Back on topic...

 

Mani gets it; so does RobbieC; so does pretty well everyone posting in the thread, except, it seems, for you. Now I don't assume these folks are that much smarter than you. (Some, maybe, but not that much.)

 

Of course I'm not saying audibility doesn't matter. I'm saying for purposes of this particular thread, the only aspects of audibility I'm interested in are those directly related to sample rate conversion, i.e., those caused by such conversions or enabled by the lack of them. I am definitely not interested in the thousandth discussion of the same old argument re subjective/objective, measurable/audible, etc., etc., etc., etc., ad infinitum, ad nauseam.

 

If understanding the simple direct statements in the previous paragraph is beyond your ability, then yep, you'll want to take whatever discussion you're interested in to another thread. (Don't cost nothin'.) If you've got something of relevance and interest re sample rate conversions and the goodness of them or their lack, then by all means say it here.

[Jud]

Excellent manufacturer paper with a portion discussing why we have the oversampling and filtering that we do, originally recommended by a commenter in another thread (thanks again, Nigel): Overview of New Filters: Version 2.1 and later

[mitchco]

Excellent manufacturer paper with a portion discussing why we have the oversampling and filtering that we do, originally recommended by a commenter in another thread (thanks again, Nigel): Overview of New Filters: Version 2.1 and later

 

An oldie but goodie: http://www.mlssa.com/pdf/Upsampling-theory-rev-2.pdf

[spdif-usb]

Yes. But do you know which audible aspects are directly related to sample rate conversion and which ones aren't? OK, let me rephrase that one. Do you know which aspects of an audio signal, whether they be audible or inaudible aspects, might change which audible aspects are and which ones aren't directly related to sample rate conversion?

 

Mani gets it; so does RobbieC; so does pretty well everyone posting in the thread, except, it seems, for you. Now I don't assume these folks are that much smarter than you. (Some, maybe, but not that much.)

 

Of course I'm not saying audibility doesn't matter. I'm saying for purposes of this particular thread, the only aspects of audibility I'm interested in are those directly related to sample rate conversion, i.e., those caused by such conversions or enabled by the lack of them. I am definitely not interested in the thousandth discussion of the same old argument re subjective/objective, measurable/audible, etc., etc., etc., etc., ad infinitum, ad nauseam.

 

If understanding the simple direct statements in the previous paragraph is beyond your ability, then yep, you'll want to take whatever discussion you're interested in to another thread. (Don't cost nothin'.) If you've got something of relevance and interest re sample rate conversions and the goodness of them or their lack, then by all means say it here.

[PeterSt]

... but alas ...

 

As I understand it, most DACs do their decimation filtering to convert to analog on "8x" bitstreams, 352.8 or 384kHz. (Here, as anywhere else in this post, I welcome correction of any inaccuracies I may have unintentionally committed.) Why then has more attention not been devoted so far to maintaining this sample rate from ADC on, thus avoiding the need for sample rate conversion?

 

Because recordings never have been in 352.8/384 maybe ?

:-)

Btw, is this named "decimation filtering" ? perhaps. But if so, it's quite confusing because "decimation" as such is downsampling. "Making less" in my poor english.

 

Most DACs these days have inputs restricted to 192kHz. Is this a limitation of USB2/3, optical, or coax? (At least for USB2 and thus 3, it doesn't appear to be. Otherwise PeterSt's DAC would have problems.)

 

There is no limit anywhere. Not even with SPDIF. I think it may be rated like that (IIRC even i2s is), but electrically (or protocol wise) it makes no sense.

USB2 has its limits, yes, but virtually only for multi channel audio. For 2ch it can be 1536K samples at least.

Btw, the predecessor of the NOS1-USB would output 2ch 384 over SPDIF.

 

I don't think that all D/A chips are based upon a speed of 384 internal sampling speed or something. But, possibly in the old days it was thought that this was enough. Otherwise it's just electrical limits we'd run into and what has been said by others in this thread (similarly) : unlinearities may be our part. I am not sure anymore whether someone said that "jitter precision" (or something like that) would be better on the higher sample rate, but altough mathematically this can be true, practically it works out differently. Think about DSD and its very high sampling rate, which now goes by accuracy of the bit (opposed to a 32 bit or so word). Not everybody will agree (maybe nobody) but I say that jitter has to be 30-36 dB better in order to represent the higher sample rate. This is not easy to see (and let's say I can be as wrong as that I can be right) because there's also the aspect of "time smearing". Okay, too difficult. But think about the high ferquency the "1 bit" square implies and how it needs to be filtered out. Well, in my view, when jitter anticipates the 32 bit audio word (x 2) then there will be 32 times more "infrequencies" at the 1 bit level implying *lower* frequencies. Now these have to be filtered out too.

Ah, ok, can't understand it myself anyway.

 

What are the common ADC sampling rates available in equipment for recording studios?

 

192 I'd say. 384 exists somewhat longer though through DAD and a derival (or predecessor ?) of it which I think is/was used by 2L.

Today some more exist, but I do not know of the A/D chips used. With good THD I mean.

But isn't it all moot since no recording label of fame is providing Hires anyway (today) ?

 

Assuming the "8x" sampling rates aren't common, is it a matter of not bothering because of the DAC input limitation referred to above, or are there technical limitations on the ADC equipment?

 

Well, what about the size of the files ? One hour of 24/192 takes about 3GB. And believe it or not, but at least .WAV can't be stored larger than 4GB. So that's a sort of technical reason too.

 

Do DSD recordings played back on DSD-capable DACs avoid the sample rate conversion problem, or are there sample rate (or other significant) conversions involved in this chain as well? More particularly, is the change of "1-bit" to, e.g., "6-bit" essentially a lossless change of digital "container" format, or are there potential audio problems that may apply to this stage?

 

Much has been chewed upon this one. But when it's reasoned out for the whole chain (including the process of recording and mixing) it makes no sense to think it will ever (have) be(en) native DSD throughout.

And if it was, no mixing whatsoever can have been in order. The solution to this (at least that's what a few come up with) ? mix in analogue and make DSD of it. Well, maybe. But not of course.

Ehm, my view, this time in very brief.

 

Regards,

Peter

 

 

PS: As you explained further down in the thread, this is obviously about recording in 352.8 (384) and when that is played back directly, not any filters have been around anywhere. Well, notice that this is the selling point of "Peter" at 2L. You may not know it, but they explicitly leave out ADC filtering. "It is just not necessary". FWIW, but it at least confirms what you're implying to be a good thing, Jud. One thing I can 100% guarantee you : what 2L does with this means of recording really works out. So, me as Anti-Hires #1 on the globe ... not so for 352.8 2L. It's audible in a fraction of a second and I never could do that at the length of full "Hires" normal albums.

So, very good subject ? Yes.