Does High Resolution Audio sound better

[wakibaki]

You could have a little man inside the computer who takes all the numbers and draws a graph through them, and then reads the value off the graph at the midpoints between the existing points and inserts it between the existing numbers, but you could take away all those extra numbers, and the graph would still be there. That's the meaning of redundancy.

[Daudio]

So you prefer friendly, but factually wrong?

 

Factually ??

 

I said nothing, nor implied anything, about ideas or facts, yours, mine or others. Just pointed out your bad attitude.

 

But if you want to talk about facts, maybe it was about time you listened to some other folks facts, not just your own. Way I see it, is you only have half of the facts, having thrown away any data that doesn't come from a machine or flawed DBT. Just because you don't trust them doesn't mean they aren't true. I don't find your arguments credible, since you are only working with half a tool kit, hopelessly twisted logic, and a big chip on your shoulder, too (expectation bias ?).

 

We can see a lack of flexibility of your mind in your dealing with facts, just now, in your treatment of Jud's discussion of Shannon's information theory. He was talking about the core meaning of information, and said so. You ignored, or didn't understand, the level of abstraction and misrepresented the discussion to be about specific audio instances, which had little to do with his point.

 

I really wish you would take advantage of the wealth of knowledge of your fellow CA members to learn more, instead of trying to be THE EXPERT WHO KNOWS ALL AND IS GOING TO TELL US ALL ABOUT IT !

 

 

So, to your question: I prefer friendly, and factually correct, but get little of either from your posts

[RHSims]

What you have here is a comparatively small number of hi-res enthusiasts congregating locally and drowning out any rational discussion in threads like this.

 

You have to remember that statistically you are only entitled to a tiny fraction of the bandwidth. Your views are not representative of the population as a whole. As long as it is an open forum, you will get people coming along from time to time and trying to reconnect you to reality. And you wonder why you have a problem attracting membership?

 

I'm not a big fan of polls, but this assertion by wakibaki suggests the need for one. After "prot" & his ilk made coherent conversation on this site all but impossible, I quit following and now just drift by by to see how the flagellation was going down, and if I could learn anything.

 

I suspect statistically, the all bow-down-to Nyquist-Shannon theorem are a tiny minority, but I could be wrong. So maybe a poll could sort out the curious, open-to-possibility types that listen to music, and those that seem to believe that theorems are all we need.

 

The funny thing is that when CDs (16.44) were first introduced in the early 1980s, most anyone with an ear was not impressed unless all they cared about were no tape hiss and no record pops. Since then the sound of CDs has improved dramatically, the theorems have not changed.

 

So I guess I'm saying theorems are great, but listening is too. All these criticisms about not believing in science are mostly bogus since the proponents arguing theorems (math), and very little listening.

 

We need a poll. The losers get a small ice berg and a new forum of their own creation. The rest of us, I'm guessing, get to wonder, ponder, speculate without constant reminders from the theorem scolds that nothing is possible and that everything sounds the same.

 

I hope the theorem school will lose their crappy headphones, crappy DACs and crappy sources, and spend some time listening to music. I'm not optimistic since certainty, not curiosity seems to be their tagline.

 

Thanks

 

Rob

[esldude]

I'm not a big fan of polls, but this assertion by wakibaki suggests the need for one. After "prot" & his ilk made coherent conversation on this site all but impossible, I quit following and now just drift by by to see how the flagellation was going down, and if I could learn anything.

 

I suspect statistically, the all bow-down-to Nyquist-Shannon theorem are a tiny minority, but I could be wrong. So maybe a poll could sort out the curious, open-to-possibility types that listen to music, and those that seem to believe that theorems are all we need.

 

The funny thing is that when CDs (16.44) were first introduced in the early 1980s, most anyone with an ear was not impressed unless all they cared about were no tape hiss and no record pops. Since then the sound of CDs has improved dramatically, the theorems have not changed.

 

So I guess I'm saying theorems are great, but listening is too. All these criticisms about not believing in science are mostly bogus since the proponents arguing theorems (math), and very little listening.

 

We need a poll. The losers get a small ice berg and a new forum of their own creation. The rest of us, I'm guessing, get to wonder, ponder, speculate without constant reminders from the theorem scolds that nothing is possible and that everything sounds the same.

 

I hope the theorem school will lose their crappy headphones, crappy DACs and crappy sources, and spend some time listening to music. I'm not optimistic since certainty, not curiosity seems to be their tagline.

 

Thanks

 

Rob

 

 

These kits really are endlessly fascinating.

 

So, let me see, was this new? Guys who don't hear the hires advantage, don't listen, don't have good equipment, aren't curious, and if they lose a poll we should put them on their own little restricted area. Gee how nice, and original of you.

 

Just as a note about those early CD players here is a quote from John Gordon Holt back in 1983 about digital.

 

For the record: Stereophile has tested two digital recorders, the Sony PCM-1 and PCM-F1, and found (as have many others) that both did an extraordinary job of copying analog originals and an excellent job at live recording compared to similarly priced analog units. We have good reason to believe that digital recording is not inherently injurious to recorded sound.

 

JGH did go on to say they had listened to 40 of the very first CD's available. Only 4 or 5 were excellent, and could hold up to comparison with the best of analog replay. Having done his own recordings with the Sony equipment he felt it outclassed any analog recording. So the problem with CD was not inherent to digital. It was a problem of everything else, and remains so to this day. Mainly a problem of mixing/mastering of CD. If hires holds any value in extra sound quality it is in absolute terms small. Were it not quite small we would have no arguments. CD recorded and mastered well leaves little on the table. Maybe nothing, maybe a tiny bit. The difference in 192 and 44 is nowhere remotely close to the difference between HD video and standard video.

[sandyk]

I notice that you appear to be steering well clear of RB CD vs. double DSD, which is also a more recent High Res. format and has quite a following with many C.A. members.

 

BTW, I also owned a Sony CDP 101, and the quoted figures were meaningless, as it was far removed from a good sounding CD player.

The Sony CD player and it's clones suffered from a frequency dependent transfer time.Philips and Philips derived systems did not suffer from this effect.

The effect can be negated by introducing a delay in the transfer time of the frequencies below 4...6kHz which equalises the delay over virtually the entire audio range - Elektor magazine published such a circuit.

[mayhem13]

In the real world, without the idealizing assumptions necessary for the sampling theorem to work, you are only correct in one specific case, something you referred to previously: "Zero padding," where you turn a 16-bit file into a 24-bit file simply by padding the 8 least significant bits with zeros.

 

However, without the idealizing assumptions of the sampling theorem, in the real world where time domain performance and frequency domain performance of a filter are conjugate variables, upsampling/oversampling a file does make an actual difference in terms of minimizing distortion/artifacts in the eventual analog result, compared to converting a file with a lower sample rate. Thus the additional information, even though it does not help to further specify the signal in terms of the sampling theorem, is in fact not redundant.

 

There you go again.......real world examples seperating time and frequency. How does one do this exactly if one is a function of the other? Analog CSD of tone burst and sweeps have been around a long time and can capture the ringing or energy storage that is so called created by poorly executed digital filters. Where are the graphs to support these claims? Right there, we get the best window of looking at your so called seperation of time and frequency (which is better expressed as amplitude but for the sake of discussion) by viewing three instead of two axis.

 

I really can appreciate your considerations that these things are possible. But what I can't support is the notion of considering theory over existing, proven science. You've got to prove the existing model is broken first...........and that starts with verifying the experiences are consistent. NO ONE has been able to do this to date. This is why we have the reverse scientific process arguements that exist here on CA. No offense, but sorry......science isn't a courtroom open to interrpretations of physical laws. It may work with 12 regular guys who's goal is to get back to work ASAP ( the judicial system's one most critical flaw) but it's not going to fly with people who actually care.

 

Doesn't this get tiring for you? As of late, it's probobly become evident to you I post less and less here for the simple reason I can't reason with some of these folks here. But if I reply to YOU, it's purely out of respect for the reasonable arguements you propose. This isn't one of them and I'm pretty sure you haven't convinced yourself of that either. That's why it's referred to as 'practice'.

[esldude]

I notice that you appear to be steering well clear of RB CD vs. double DSD, which is also a more recent High Res. format and has quite a following with many C.A. members.

 

BTW, I also owned a Sony CDP 101, and the quoted figures were meaningless, as it was far removed from a good sounding CD player.

The Sony CD player and it's clones suffered from a frequency dependent transfer time.Philips and Philips derived systems did not suffer from this effect.

The effect can be negated by introducing a delay in the transfer time of the frequencies below 4...6kHz which equalises the delay over virtually the entire audio range - Elektor magazine published such a circuit.

 

If I recall right, the CDP101 used one DAC that was switched between right and left channel. I didn't say the CDP101 sounded great, but it was capable of low distortion. Sony tried to fix this half sample delay between channels with something of a delay in the filtering. It was close, but not quite right. I believe that is what the fix in the Elektor circuit was likely for, though I haven't seen the article. Just a guess which seems likely. That multiplexed DAC was quite an odd choice to me. Yes, Philips never did such a thing and therefore had no issue with it.

 

As for RB and DSD double or otherwise I don't care much what following it has. As I don't use DSD capable DACs currently it simply isn't an issue. I have software which converts DSD to PCM if I need music available in DSD.

[sandyk]

As for RB and DSD double or otherwise I don't care much what following it has. As I don't use DSD capable DACs currently it simply isn't an issue.

 

Nevertheless, according to your reasoning about a wider audio bandwidth being pointless, Double DSD improvements over RB CD must be a figment of the imagination too ?

[Don Hills]

... Sony tried to fix this half sample delay between channels with something of a delay in the filtering. It was close, but not quite right. ...

 

The result was about 22 degrees difference between the channels at 20 KHz.

[semente]

If you are going to use numbers in your argument, you'd best get the correct numbers and use them correctly, lest you demonstrate your lack of knowledge.

 

I have commercially produced recordings that show dynamic range of sustained solo instrument note played ppp vs. 100 person orchestra tutti playing fff of 55 dB. However, as soon as the musicians stop playing the notes will decay through hall reverberations all the way down to the noise floor or listener's threshold of hearing. This is a natural property of live acoustic performances.

 

If you look at microphones such as the DPA4006 that are commonly used for recording classical music you will see that its equivalent noise is 15 dB and it allows for peak SPLs of 137 dB without exceeding 1% distortion. This seems more like 120 dB dynamic range to me.

 

Most important, your argument is based on mediocrity. You are using assumed weaknesses in other parts of a system to justify skimping. The end result is a collection of mediocre components and processes that are self justifying and an end result that is mediocre. This may be a wise business decision if the goal is a value product, but is inappropriate when the goal is performance. We see noisy microphones which are justified based on tape machine S/N ratios, and then digital recording formats based on noisy microphones. This is the end result of this line of arguing.

 

One more point on the noise floor. To justify existing noise as masking additional noise, the additional noise has to be about 10 dB below the existing noise, so that the total noise changes by a small fraction of a decibel. If the two noise floors are equal the net result will be a 3 dB increase in noise.

 

For illustration purposes, here's a DR plot of the combined first 6 CD tracks of Berlioz's "Lélio" by Inbal on a 1987 Denon recording:

 

 

 

 

P.S.: this work's pieces alternate between an actor narrating a story and an orchestra with chorus and solo singers.

[Jud]

There you go again.......real world examples seperating time and frequency. How does one do this exactly if one is a function of the other? Analog CSD of tone burst and sweeps have been around a long time and can capture the ringing or energy storage that is so called created by poorly executed digital filters. Where are the graphs to support these claims? Right there, we get the best window of looking at your so called seperation of time and frequency (which is better expressed as amplitude but for the sake of discussion) by viewing three instead of two axis.

 

I really can appreciate your considerations that these things are possible. But what I can't support is the notion of considering theory over existing, proven science. You've got to prove the existing model is broken first...........and that starts with verifying the experiences are consistent. NO ONE has been able to do this to date. This is why we have the reverse scientific process arguements that exist here on CA. No offense, but sorry......science isn't a courtroom open to interrpretations of physical laws. It may work with 12 regular guys who's goal is to get back to work ASAP ( the judicial system's one most critical flaw) but it's not going to fly with people who actually care.

 

Doesn't this get tiring for you? As of late, it's probobly become evident to you I post less and less here for the simple reason I can't reason with some of these folks here. But if I reply to YOU, it's purely out of respect for the reasonable arguements you propose. This isn't one of them and I'm pretty sure you haven't convinced yourself of that either. That's why it's referred to as 'practice'.

 

This depends on the audibility of the decimation filtering at the recording end; the interaction between the ADC filtering/conversion and your DAC's filtering/conversion; on the particular interpolation filtering used by your DAC; and on the audibility of the sigma-delta modulation step in your DAC. (Or better, it depends at least on those things.) The interaction of all these factors is complex, and thus the answer will not be simple or uniform across recordings and DACs.

 

***

 

[Y]ou can try out hardware or software that offers a choice of filters. So there are audible/measurable differences. But you're not going to be in a position to hear them unless they override whatever other filtering/conversion is going on at the ADC and/or DAC.

 

Hi Anthony. Taking the last bit I quoted from my previous post first: Because of your DACs, any CSD (a/k/a "waterfall") plots are likely not going to show changes as a result of filter changes in PC software or hi res. This is because both your Benchmark and your ODAC use asynchronous sample rate conversion (ASRC) as an anti-jitter measure. Taking your Benchmark as an example, no matter what sample rate you feed it, it's converted to 110kHz internally. You can play around as much as you like with input resolutions, either using software or feeding higher res files to the DAC, and internally it's still going to be 110kHz files with the Benchmark's filtering.

 

Now turning to what you said: There are good measurements available showing these tradeoffs. This isn't "theory over existing, proven science," it's just math. The math shows that if we make idealizing assumptions (perfect filter applied to perfectly band-limited signal over infinite time), a sample rate just over twice the highest frequency in the perfectly band-limited signal will allow the perfect filter to perfectly reconstruct that signal, given infinite time in which to apply the perfect filter. That of course is the sampling theorem. The math also clearly shows that if we don't have a perfect filter, perfectly band-limited signal, and infinite time (which of course we never do in the real world), then as frequency domain performance of the filter gets better (as the filter is configured to better eliminate aliases), then the time domain performance (ringing) gets worse, and vice versa. The math also shows the interrelationship of these with phase behavior.

 

Regarding real world examples, I think the stuff at SRC Comparisons is very clear. Just bring up software like iZotope RX Adv 2, generally acknowledged to be some of the best SRC software available, and see how the passband, pulse and sweep tests interrelate for the "Intermediate Phase" and "High Steepness" parameter settings.

 

We can (and have) discussed audibility of what's shown in these test measurements, but it's quite evident that the measurements do show these filters acting precisely as the math says they must.

[bdiament]

...This is what Tony, JonP, and Barry were talking about - doing a conversion that is completely transparent is mathematically impossible, and according to people with a lot of experience, difficult to impossible to do inaudibly...

 

Hi Jud,

 

I'll speak only for myself. I never said what is attributed to me in the quote above.

 

There *was* a typo in an earlier post of mine, which I corrected in a subsequent post.

To be clear, I said that I would not (inadvertently left out the "not" the first time) attribute the issues I hear with the results of converting high res to Redbook to the sample rate conversion. To my ears, the biggest issues are inherent in the 16/44 format. They are still there, plain to hear, when recording directly to 16/44.

 

Of course some SRC gets out of the way more effectively than other SRC but I believe you know I'm a fan of one particular algorithm (and have a dozen or more in the "toolbox").

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

P.S. Neither would I agree that upsampling in a digital filter is interchangeable with the type of sample rate conversion I might do using iZotope's 64-bit SRC algorithm offline. These may be used in marketing speak but that is not all these things are. Again, I have mentioned a source for much better information on this than I'm able to supply.

[Jud]

Hi Jud,

 

I'll speak only for myself. I never said what is attributed to me in the quote above.

 

There *was* a typo in an earlier post of mine, which I corrected in a subsequent post.

To be clear, I said that I would not (inadvertently left out the "not" the first time) attribute the issues I hear with the results of converting high res to Redbook to the sample rate conversion. To my ears, the biggest issues are inherent in the 16/44 format. They are still there, plain to hear, when recording directly to 16/44.

 

Of course some SRC gets out of the way more effectively than other SRC but I believe you know I'm a fan of one particular algorithm (and have a dozen or more in the "toolbox").

 

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

P.S. Neither would I agree that upsampling in a digital filter is interchangeable with the type of sample rate conversion I might do using iZotope's 64-bit SRC algorithm offline. These may be used in marketing speak but that is not all these things are. Again, I have mentioned a source for much better information on this than I'm able to supply.

 

Hi Barry. Difficult to impossible to do inaudibly is not an accurate representation of what you've said, and I apologize. If I remember correctly, you have said iZotope's offline conversion and the internal conversion done by your Metric Halo both sound like your mic feed to you.

 

Edit: At least the conversion/filtering done on various upsampling and editing operations. Whatever the reason, and I acknowledge your thinking that it is not the conversion but the resulting format, decimation to 16/44 does not sound transparent to you.

[mayhem13]

Hi Anthony. Taking the last bit I quoted from my previous post first: Because of your DACs, any CSD (a/k/a "waterfall") plots are likely not going to show changes as a result of filter changes in PC software or hi res. This is because both your Benchmark and your ODAC use asynchronous sample rate conversion (ASRC) as an anti-jitter measure. Taking your Benchmark as an example, no matter what sample rate you feed it, it's converted to 110kHz internally. You can play around as much as you like with input resolutions, either using software or feeding higher res files to the DAC, and internally it's still going to be 110kHz files with the Benchmark's filtering.

 

Now turning to what you said: There are good measurements available showing these tradeoffs. This isn't "theory over existing, proven science," it's just math. The math shows that if we make idealizing assumptions (perfect filter applied to perfectly band-limited signal over infinite time), a sample rate just over twice the highest frequency in the perfectly band-limited signal will allow the perfect filter to perfectly reconstruct that signal, given infinite time in which to apply the perfect filter. That of course is the sampling theorem. The math also clearly shows that if we don't have a perfect filter, perfectly band-limited signal, and infinite time (which of course we never do in the real world), then as frequency domain performance of the filter gets better (as the filter is configured to better eliminate aliases), then the time domain performance (ringing) gets worse, and vice versa. The math also shows the interrelationship of these with phase behavior.

 

Regarding real world examples, I think the stuff at SRC Comparisons is very clear. Just bring up software like iZotope RX Adv 2, generally acknowledged to be some of the best SRC software available, and see how the passband, pulse and sweep tests interrelate for the "Intermediate Phase" and "High Steepness" parameter settings.

 

We can (and have) discussed audibility of what's shown in these test measurements, but it's quite evident that the measurements do show these filters acting precisely as the math says they must.

 

But the increased bitrate and improved filters audible improvements 'should' result in less ringing of the fundamental and associated harmonics. Correct me but we're not talking about jitter and related harmonic distortion are we? We can measure those as well but can't associate harmonic distortion with a reliable subject test that shows the threshold of human hearing. Heck, it's even been suggested that many an audiophile actually enjoy even order distortion products. Also on the analog side we have the inconvenient truth that baseline harmonic distortion numbers are already exceeding the small advantages you propose. Where's the value? Show me the money?

 

Well.......you already know my position on the small signal stuff. No need for us to go round and round about it. I'll still respect you in the morning! Lol

 

BTW.....did you catch my post where I indictated at mmerrils request I borrowed a Lynx Hilo( selfish reason actually....I had some analog stuff that needed archiving anyways) from a friend's studio? Dedicated an entire three day weekend to it actually.

[lamode]

Most important, your argument is based on mediocrity.

 

It's not that we should strive for mediocrity, but your cherry picking is becoming absurd. Even if you are a classical music lover, listening to a digital recording using the quietest of microphones, listening to a piece which has a dynamic range far greater than 99% of recordings, and you happen to have a high end audio system with >110dB S/N ratio, with volume set high for 105dB peaks, and you happen to live in a remote area far from traffic, in a house with all appliances turned off, at a time when there is no wind or rain, or anyone else in the house, and you have perfect hearing.... Even if all these conditions are true, the worst that can happen is that you might hear a little hiss in the final seconds of the recording as the final reverberation dies out completely. Tragic. My heart goes out to all 3 people that this ever happened to.

[sbgk]

It's not that we should strive for mediocrity, but your cherry picking is becoming absurd. Even if you are a classical music lover, listening to a digital recording using the quietest of microphones, listening to a piece which has a dynamic range far greater than 99% of recordings, and you happen to have a high end audio system with >110dB S/N ratio, with volume set high for 105dB peaks, and you happen to live in a remote area far from traffic, in a house with all appliances turned off, at a time when there is no wind or rain, or anyone else in the house, and you have perfect hearing.... Even if all these conditions are true, the worst that can happen is that you might hear a little hiss in the final seconds of the recording as the final reverberation dies out completely. Tragic. My heart goes out to all 3 people that this ever happened to.

 

Well he did say you weren't good with numbers

[mmerrill99]

...

BTW.....did you catch my post where I indictated at mmerrils request I borrowed a Lynx Hilo( selfish reason actually....I had some analog stuff that needed archiving anyways) from a friend's studio? Dedicated an entire three day weekend to it actually.

I missed this post too, Mayhem, got a link as I can't find it with a search?

[lamode]

Well he did say you weren't good with numbers

 

I'll just ignore the petty, ad hominem remarks.

[mmerrill99]

Just found it:

Well Jud, all I can add is that I have an ODAC, a Benchmark and for the purposes of entertaining another member here, a time with a LYNX. I AB'd them against one another and found it very difficult to prefer one over another. Yes, there were some very small discernable differences in detail when being extremely critical of well known tracks. Given the scale of costs, the ODAC is IMO an outstanding value.

 

So what is your point? You are comparing an ODAC (with built in ASRC in the chip) against your Benchmark (with ASRC) & Lynx Hilo without ASRC, AFAIK?

You heard "some very small discernable differences in detail" - is detail your criteria for judging audio differences or is there anything else?

[bdiament]

Hi Barry. Difficult to impossible to do inaudibly is not an accurate representation of what you've said, and I apologize. If I remember correctly, you have said iZotope's offline conversion and the internal conversion done by your Metric Halo both sound like your mic feed to you.

 

Edit: At least the conversion/filtering done on various upsampling and editing operations. Whatever the reason, and I acknowledge your thinking that it is not the conversion but the resulting format, decimation to 16/44 does not sound transparent to you.

 

Hi Jud,

 

To be clear, what I have had trouble discerning from my direct mic feed is the output of the Metric Halo ULN-8 when I record at 24/192. No conversions are involved. (Here we may see this differently. I've already said I do not see or hear upsampling in the filter as being the same as SRC.)

 

Further, it is not any decimation to 16/44 that sounds degraded to me ("not transparent" is an understatement), it is the 16/44 itself, without any decimation, that I find problematic. I've found this to be true with every converter I've ever tried, from the earliest ones in the early '80s to the best ones I've heard today, whether recording directly to 16/44 in the early days or starting with high res nowadays and using the finest SRC and dithering I know of, as offline processes, to convert the high res to 16/44.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

[jabbr]

In the real world, without the idealizing assumptions necessary for the sampling theorem to work, you are only correct in one specific case, something you referred to previously: "Zero padding," where you turn a 16-bit file into a 24-bit file simply by padding the 8 least significant bits with zeros.

 

However, without the idealizing assumptions of the sampling theorem, in the real world where time domain performance and frequency domain performance of a filter are conjugate variables, upsampling/oversampling a file does make an actual difference in terms of minimizing distortion/artifacts in the eventual analog result, compared to converting a file with a lower sample rate. Thus the additional information, even though it does not help to further specify the signal in terms of the sampling theorem, is in fact not redundant.

I think I am agreeing with you but the lengthly discussion of information theory I think is confusing the issue.

 

I am (loudly hearing a debate that seems to conflate information theory with "real world" practice.

 

By real world practice we are talking about currently available DAC chips which do various things to produce an analog output from a particular digital signal. I don't think there's a big need to invoke sampling theory here, rather that a particular class of DAC chips are known to behave in a certain way. They might oversample/upsample themselves alternatively may be provided with an oversampled/upsampled input. Similarly they might SDM themselves alternatively be provided with a digital signal already in DSD format. For PCM signals, it is well known that SDM DACs benefit from upsampling more than Ladder DACs. For DSD signals upsampling makes the filter's job easier (needs less slope

 

Again these are all implementation dependent issues.

[jabbr]

 

To be clear, what I have had trouble discerning from my direct mic feed is the output of the Metric Halo ULN-8 when I record at 24/192. No conversions are involved. (Here we may see this differently. I've already said I do not see or hear upsampling in the filter as being the same as SRC.)

 

Further, it is not any decimation to 16/44 that sounds degraded to me ("not transparent" is an understatement), it is the 16/44 itself, without any decimation, that I find problematic. I've found this to be true with every converter I've ever tried, from the earliest ones in the early '80s to the best ones I've heard today, whether recording directly to 16/44 in the early days or starting with high res nowadays and using the finest SRC and dithering I know of, as offline processes, to convert the high res to 16/44.

 

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

Your observations are very important. You are very careful to state that these are your personal observations. From a scientific point of view, if there were a scientist who wished to devote the effort to quantify this, your observations could be used as the basis for a study where a number of other observers could make similar observations and the observations could be tabulated and statistically analyzed.

[jabbr]

Sorry but that's rather pedantic and disingenuous to call 8x oversampling '3 conversions'. For one thing, the noise generated by the interpolation is tiny, less than -100dB, so the effect of the whole filter is still inaudible in the real world.

Shrug. Seems fine to me. 2^3 sampling.

 

For the life of me, I can't understand the meaningless fixation on hi-res audio, when every transducer we are using is deeply flawed, not to mention listening environments. Upgrading a speaker or listening room will transform your music in ways which HD audio can only dream about.

 

You are being silly now. It should be understood that audio formats are far far far more important than the details of your speakers and walls in your apartment, because formats form the basis for long term sound archiving.

[Jud]

But the increased bitrate and improved filters audible improvements 'should' result in less ringing of the fundamental and associated harmonics. Correct me but we're not talking about jitter and related harmonic distortion are we? We can measure those as well but can't associate harmonic distortion with a reliable subject test that shows the threshold of human hearing. Heck, it's even been suggested that many an audiophile actually enjoy even order distortion products. Also on the analog side we have the inconvenient truth that baseline harmonic distortion numbers are already exceeding the small advantages you propose. Where's the value? Show me the money?

 

Well.......you already know my position on the small signal stuff. No need for us to go round and round about it. I'll still respect you in the morning! Lol

 

BTW.....did you catch my post where I indictated at mmerrils request I borrowed a Lynx Hilo( selfish reason actually....I had some analog stuff that needed archiving anyways) from a friend's studio? Dedicated an entire three day weekend to it actually.

 

I didn't see your post about the Hilo, but I'll check it out.

 

You're correct that we're not talking about jitter effects.

 

You're also correct that there are some indications that people (not just audiophiles) enjoy some forms of distortion, particularly those that give a "hot" or energetic sound. Jimmy Iovine, who went on to be the less famous Beats founder, used a gizmo called the Aphex Aural Exciter on Born to Run. I don't think it's any accident that if you look at the settings people have come up with "by ear" for the iZotope filter bundled with Audirvana Plus, most of them don't cut much at the high end. But that would suggest whatever that high end is imparting to the sound (intermodulation distortion?) is something people are aware of, if perhaps only to the extent of a liking for the overall energetic "feel." (See how I snuck that in? OK, maybe I wasn't that sneaky....)

 

And yep, I already know your position, and think there's nothing wrong with prioritizing speakers (with amps that can run them) and room.

 

Be well, my friend.

[Jud]

Even if all these conditions are true, the worst that can happen is that you might hear a little hiss in the final seconds of the recording as the final reverberation dies out completely. Tragic. My heart goes out to all 3 people that this ever happened to.

 

I tend to disagree on the audibility of various filtering configurations, having heard non-"pathological" (the term in quotes being my word for filters configured by ear by amateurs like me) filters done by pros and having been surprised at the degree of difference to be heard between different filters. I really would not have thought there would be such a difference. (I do have a DAC that will accept 352.8/384 input and send it right to the SDM, and that will also accept DSD128 input, so any differences in filtering of the input are pretty evident.)

 

But even so, re the above - *that's* funny. Hey, we're audiophiles.