PlayClassics test files to compare file formats

[徐中銳]

I fail...

 

You have and will in a few second be on my Ignore List. Because there's only so much time. There's other people...

 

Here's a parting gift, Spanish :

En la tabla central, El Bosco recrea un proverbio flamenco: “El mundo es como un carro de heno y cada uno coge lo que puede”. Todos los estamentos, incluido el clero -censurado por vicios como la avaricia y la lujuria-, quieren coger ese heno y subirse al carro. Para lograr su objetivo no dudan en cometer todo tipo de atropellos y pecados, incluso el asesinato.

[semente]

Okay, have had some time to spend on this under reasonably good conditions.

 

I listened to only the 24 96 and the 44 16. MP3 is known not to be fully transparent so I had no interest in it.

 

After a few listens I thought there was a difference at times. Especially in the decay in the hall of the piano. When I would go directly back and forth the difference wasn't really there. One afternoon when doing other things I put these two into random play mode so I wouldn't know at any given time which was playing. No differences were apparent to note.

 

So I ran these through an ABX test. 16 trials done twice. 10 of 16 and 9 of 16 done a couple hours apart. Had I gotten positive results or heard differences in the sighted part of the test with conviction I would have tried some of the in between rates and bit depths.

 

After all listening I upsampled the 4416 file to 2496. Dumped it into a sound editor. Flipped polarity and mixed it with the original 2496 file. You get noise that wiggles slightly around the -88 or 89 db level. What you would expect from 16 bit dithering. I boosted this by 40 db, and listened. You hear faint swishing noise. There are some points in the latter portion of the track where you get low level clicks. They coincide with places the original clipped or very nearly did. I imagine the up and down sampling may have caused some intersample overs to show up this way. I then took this file and slowed it down by a factor of 4. This would make it easy to listen to any ultrasonic effects one normally doesn't hear. This was still boosted by 40 db mind you. Really nothing to report, a bit higher pitched tone to the swishing hiss much like FM radio hiss between channels.

 

Here is an FFT of one of the louder (though unclipped) sections using this mix between original and upsampled 4416. Note this is not boosted in level.

 

[ATTACH=CONFIG]22533[/ATTACH]

 

You can read the RMS level for the 0-40 khz band is -89 db. I used shaped dither in the upsampling which is what gives the noise floor the shape you see.

 

I don't see any reason to think these would be audibly different.

 

Hi Dennis,

 

Your first impressions seem to indicate that there is more apparent "resolution" in the decay of the higher res file (and they coincide with my own).

 

In what manner were you trying to characterise that possibility with your technical evaluation?

 

From your previous comments in another topic I get the impression that you are focusing mainly in frequency response differences and noise floor.

 

And how could we measure the effect of artifacts that produce say "grain" in high frequencies?

 

Cheers,

Ricardo

[Jud]

I've had a very interesting listening experience with the test files. I did not go through the same process that Ricardo did because I am too lazy. However, I am very curious (curiosity and laziness are a bad combination!), so I am very glad Ricardo has done this and will be telling us about it.

 

Meanwhile, I have taken another path and had a very interesting experience.

 

My first listening experience was summed up well by Sal:

 

2. The 24/96 file sounded beautiful, clean, smooth, and without edge throughout the entire cut. Also seemed to present a better representation of the differing timbre and harmonic structures of the notes.

3. Jumping to 16/44 RedBook the file seemed to marginally lose that detail of the timbre and harmonics that I heard on the 24/96. Not day/night and it would be very hard to pick out in any blind test if at all.

 

Then I got curious(!). I thought to myself about my listening chain (similar to the chain most of us are using), and wondered how it affected what I was hearing. Because none of us are listening to Mario's files, we're listening to the entire signal chain from analog musical performance to analog reproduction of that performance. As background, and to be thorough, here are the steps in the chain (those I can think of - I could have missed something):

 

Performance -> sigma-delta modulator in ADC -> decimation filter(s) in ADC and/or software takes the bitstream from ~2.8 or ~5.6MHz to a 96 or 44.1KHz sample rate -> my custom filter in iZotope 64-bit SRC interpolates the bitstream to 24/352.8 or 24/384KHz -> sigma-delta modulator in DAC takes bitstream to ~5.6MHz -> final analog filter in DAC converts bitstream to music.

 

I have software (Audiophile Inventory) that does offline sample rate conversion, as opposed to the inline sample rate conversion done by my iZotope filter. Audiophile Inventory also has the huge advantage that its filtering is designed by someone who actually knows what he is doing, rather than me. One additional advantage of Audiophile Inventory is that it will convert to DSD, eliminating my DAC's internal sigma-delta modulator step. So the chain becomes:

 

Performance -> sigma-delta modulator in ADC -> decimation filter(s) in ADC and/or software takes the bitstream from ~2.8 or ~5.6MHz to a 96 or 44.1KHz sample rate -> Audiophile Inventory converts to sigma-delta modulated bitstream at ~5.6MHz -> final analog filter in DAC converts bitstream to music.

 

How would using professional offline filtering and conversion affect what I was hearing? Would it make both the RedBook and 24/96 versions sound so good that they would sound pretty much the same to me? (I say "to me," because these are just my subjective impressions.) Or would the higher quality filtering/conversion highlight differences in the files that my custom filter and the DAC's internal sigma-delta modulator had failed to bring out?

 

I listened to the DSD128 conversion of the RedBook file first (just the Albeniz Iberia). Very nice, indeed better than my recollection of its rendering through my custom filter.

 

Then I listened to the DSD128 conversion of the 24/96 file. It was absolutely gorgeous, probably as close to a real piano sound as I've heard in my system. I was sad when it ended, and I resolved to make a donation and download the recording (which I've done). I even went on to listen to what I assume is the Chopin. I hadn't listened to much of it previously, because classical/operatic vocals are seldom my thing (with a number of notable exceptions). But this time the vocal quality was quite beautiful, and I didn't mind it. (Though I must say even this rendering did not quite persuade me to purchase the recording.)

 

So my impression listening to the files that were professionally filtered/converted to DSD is that both sound good, but somehow the converted 24/96 file crossed some subjective boundary to real emotional enjoyment and musical beauty.

 

(Edited to correct embarrassing error in DSD sample rates.)

[odelay]

Nice post Jud.

Good to get your perspective on the tangled web of A-D and D-A filters. I have enough trouble counting my toes. Glad there are some bright minds out there who manage to turn music into 1s and 0s... and back again!

[Sal1950]

I fail to see the relevance of the linked article.

Could you please elaborate on your point, making an effort to write it in a more intelligible English?

 

Cheers,

R

 

You have and will in a few second be on my Ignore List. Because there's only so much time. There's other people...

 

Here's a parting gift' date=' Spanish :[/font']

 

semente, Please don't feed the trolls!

[kumakuma]

semente, Please don't feed the trolls!

 

I don't think he's a troll but I must admit that I have a hard time understanding the significance of the hay wagon.

[Sal1950]

I don't think he's a troll but I must admit that I have a hard time understanding the significance of the hay wagon.

 

??? Either there is a sever language block or he's just jerking our chains. Every post I've read by him is like that, nonsensical direction to it that could not be responded to ?

[esldude]

Hi Dennis,

 

Your first impressions seem to indicate that there is more apparent "resolution" in the decay of the higher res file (and they coincide with my own).

 

In what manner were you trying to characterise that possibility with your technical evaluation?

 

From your previous comments in another topic I get the impression that you are focusing mainly in frequency response differences and noise floor.

 

And how could we measure the effect of artifacts that produce say "grain" in high frequencies?

 

Cheers,

Ricardo

 

I listened and thought noise floor was lower with the hirez file in the decay after notes were struck. Then I would AB short sections and find it wasn't so.

 

So I blind tested that notion with a short snippet of just such a decay. Those results failed to confirm I was hearing a difference.

 

So I measured the differences. Was there anything there that could be heard? Response imbalance, high noise, etc. etc. Down sampling to 44.1 and 16 bit meant lost information vs the original. Upsampling to compare can't put back what was lost. You subtract one from the other. What was lost is what is left.

 

Quick rule of thumb that won't lead you wrong very often. It is possible in such a residual difference to faintly hear things that will not be audible listening to the original files. If you play that difference file at your normal playback levels and hear absolutely nothing, you can be about as sure as it gets the difference is inaudible. I heard nothing in that difference file at normal levels nor at somewhat elevated levels. Meaning complete silence in a quiet room and house. When boosted 40 db, which is a lot, I heard only slight hissing noise.

 

My sighted opinion all considered was no difference, an ABX comparison gave no evidence of difference, technical measures showed a difference of a kind so small and so even throughout it almost surely cannot be the source of an audible difference. Plus it was exactly what you expect of good dither on 16 bits.

 

Now under the right conditions humans are surprisingly sensitive to noise floor modulation and a simple note decaying into the near silent noise floor with quality recordings is just where you might hear such a thing. Let me give an example. If you have a clean recording with low noise from the recording venue, and added white noise at -40 db parts of the recording will decay into that added noise. Comparing it to the clean copy the difference will be most apparent on things like struck piano note with time to decay to silence. In louder or busier parts of a recording you may not hear a difference as the noise is masked. Even more noticeable is when the noise floor is modulated. Rises with level drops with level or perhaps is modulated in ways not related to level.

 

As for grainy sound, that usually is some sort of IM distortion. It isn't noticeable just in the quiet decay of notes. The small noise difference in these two versions of these excellent recordings wouldn't result in perceiving a grainy quality.

 

Another way to look at the noise of 16 bit vs the 24 bit is the level of it. It was -89 or 90 db from max level. If I played this loud, with peaks of 105 db, the noise of 16 bit is around 15 db above our hearing threshold. Of course few rooms have less than 30 db noise levels. We can hear into noise some 15 or maybe 20 db, but with music playing even as notes decay it almost surely would be masked by ambient noise. Such a noise level would be on the edges of being perceived at best. So even if it just barely was noticeable on a few very quiet areas of decay it wouldn't amount to enough to change the basic character of the music.

[徐中銳]

kumakuma, let's simply say those Classics I know re-play naturally by involuted alluding...

 

The Spanish was as much for Mario—I can quote an apt Gracián or two too. But, how do you regard this PlayClassics test ? And why shouldn't I ignore Ricardo ? Must I for ever reply to his foolery (you've noticed) ?

[Sal1950]

kumakuma' date=' let's simply say those [i']Classics[/i] I know re-play naturally by involuted alluding...

 

The Spanish was as much for Mario—I can quote an apt Gracián or two too. But, how do you regard this PlayClassics test ? And why shouldn't I ignore Ricardo ? Must I for ever reply to his foolery (you've noticed) ?

What the h-ll are you talking about. Every thing you write makes no sense and is unintelligible. You proved with the last sentence you are capable of intelligent conversation, why this constant gibberish?

[徐中銳]

What...

 

Dolt as you're proving yourself, you've missed my citing the Haywain in my first Post into this Thread ?

 

And do you recall Ricardo writing « Saludos » ?

[semente]

I've had a very interesting listening experience with the test files. I did not go through the same process that Ricardo did because I am too lazy. However, I am very curious (curiosity and laziness are a bad combination!), so I am very glad Ricardo has done this and will be telling us about it.

 

Meanwhile, I have taken another path and had a very interesting experience.

 

My first listening experience was summed up well by Sal:

 

 

 

Then I got curious(!). I thought to myself about my listening chain (similar to the chain most of us are using), and wondered how it affected what I was hearing. Because none of us are listening to Mario's files, we're listening to the entire signal chain from analog musical performance to analog reproduction of that performance. As background, and to be thorough, here are the steps in the chain (those I can think of - I could have missed something):

 

Performance -> sigma-delta modulator in ADC -> decimation filter(s) in ADC and/or software takes the bitstream from ~2.8 or ~5.6MHz to a 96 or 44.1KHz sample rate -> my custom filter in iZotope 64-bit SRC interpolates the bitstream to 24/352.8 or 24/384KHz -> sigma-delta modulator in DAC takes bitstream to ~5.6MHz -> final analog filter in DAC converts bitstream to music.

 

I have software (Audiophile Inventory) that does offline sample rate conversion, as opposed to the inline sample rate conversion done by my iZotope filter. Audiophile Inventory also has the huge advantage that its filtering is designed by someone who actually knows what he is doing, rather than me. One additional advantage of Audiophile Inventory is that it will convert to DSD, eliminating my DAC's internal sigma-delta modulator step. So the chain becomes:

 

Performance -> sigma-delta modulator in ADC -> decimation filter(s) in ADC and/or software takes the bitstream from ~2.8 or ~5.6MHz to a 96 or 44.1KHz sample rate -> Audiophile Inventory converts to sigma-delta modulated bitstream at ~5.6MHz -> final analog filter in DAC converts bitstream to music.

 

How would using professional offline filtering and conversion affect what I was hearing? Would it make both the RedBook and 24/96 versions sound so good that they would sound pretty much the same to me? (I say "to me," because these are just my subjective impressions.) Or would the higher quality filtering/conversion highlight differences in the files that my custom filter and the DAC's internal sigma-delta modulator had failed to bring out?

 

I listened to the DSD128 conversion of the RedBook file first (just the Albeniz Iberia). Very nice, indeed better than my recollection of its rendering through my custom filter.

 

Then I listened to the DSD128 conversion of the 24/96 file. It was absolutely gorgeous, probably as close to a real piano sound as I've heard in my system. I was sad when it ended, and I resolved to make a donation and download the recording (which I've done). I even went on to listen to what I assume is the Chopin. I hadn't listened to much of it previously, because classical/operatic vocals are seldom my thing (with a number of notable exceptions). But this time the vocal quality was quite beautiful, and I didn't mind it. (Though I must say even this rendering did not quite persuade me to purchase the recording.)

 

So my impression listening to the files that were professionally filtered/converted to DSD is that both sound good, but somehow the converted 24/96 file crossed some subjective boundary to real emotional enjoyment and musical beauty.

 

(Edited to correct embarrassing error in DSD sample rates.)

 

Hi Jud,

 

Thanks for your description.

My D/AC doesn't support DSD but your post has shown that offline upsampling offers other possibilities.

 

In my case I listened to the sample files by upsampling them to 192KHz with the music player in order to bypass the ASRC in my D/AC which "sounds" worse.

 

R

[skolis]

 

I have software (Audiophile Inventory) that does offline sample rate conversion, as opposed to the inline sample rate conversion done by my iZotope filter.

 

So my impression listening to the files that were professionally filtered/converted to DSD is that both sound good, but somehow the converted 24/96 file crossed some subjective boundary to real emotional enjoyment and musical beauty.

 

 

Nice tip, thanks. was not familiar with Audiophile Inventory.

Question: Can a PCM file be converted to DSD and then just be played back

as a DSD file with no other "intervention" or manipulation?

Would like to try it, but would like to know that it's reasonably simple to work with.

thx.

[Jud]

Nice tip, thanks. was not familiar with Audiophile Inventory.

Question: Can a PCM file be converted to DSD and then just be played back

as a DSD file with no other "intervention" or manipulation?

Would like to try it, but would like to know that it's reasonably simple to work with.

thx.

 

In any DAC that accepts DSD, no further user intervention is required.

 

(Note: DACs with ESS chips will internally oversample even when fed DSD input.)

[kumakuma]

Dolt as you're proving yourself' date=' you've missed my citing the [b']Haywain[/b] in my first Post into this Thread ?

 

And do you recall Ricardo writing « Saludos » ?

 

徐先生, I must be a dolt as well because I still don't understand how the Hay Wagon picture or your quote to Ricardo about corrupt clergymen relates to this thread.

[徐中銳]

Okay—quickly, this Thread, who is Mario Martinez ? What's his « PlayClassics, the art of true music » ? Indeed, for what ends did he take time to provide these test tracks ?

 

As Mario wrote :

We are dedicated to chamber music. We are purposely recording this music in a way that resembles its original setup. Our auditorium is a small hall, much the size of the palace chambers used to perform chamber music back in the XIX century.

Hieronymus Bosch's Haywain Triptych (which may well have already featured on a Classical/chamber music record sleeve) was painted ca. 1516 and been in Spain since Philip II's acquisition in 1570, possibly inspiring many artists, including musical...

 

In English, what I quoted from Museo Nacional del Prado :

In the central panel, Bosch recreates the Flemish proverb “the world is like a hay cart and everyone takes what he can”. All of the powers-that-be, including the clergy —censured for vices such as avarice and lust— want to catch that hay and climb onto the wagon. They have no qualms about committing all sorts of crimes to do so, including murder.

It's not just your « corrupt clergymen » because « everyone » are motivated by something ?

 

Myself, I want to spend more time caring for the 7 or 9 people that are my dear friends—besides the constant responsibility of my own well-being. And, it's not quite a crime, I took my time back from having to deal with Ricardo ever again—because I can.

 

Incidentally, of PlayClassics' vaunted « Albniz Iberia » I heard, 3 months ago, a 35-minute introduction of the composer from New Zealand's own Concert Radio if some readers are interested in going deeper...

[kumakuma]

Okay—quickly' date=' this Thread, who is Mario Martinez ? What's his « PlayClassics, the art of true music » ? Indeed, for what ends did he take time to provide these test tracks ?

 

You seem to be implying that Mario has some sinister or even <gasp> commercial motive in providing these files.

 

In regards to his motive, I am pretty sure that he provided them at the request of members of this forum when we were discussing high resolution file formats.

 

Mario has been extremely generous with both his time and his company's product, providing free music to anyone who asks for it.

 

I, for one, will be very happy if this leads to more commercial success for him and his company because this will allow him to continue to provide this wonderfully recorded music to those who enjoy it.

[sandyk]

I, for one, will be very happy if this leads to more commercial success for him and his company because this will allow him to continue to provide this wonderfully recorded music to those who enjoy it.

 

Likewise.

[Sal1950]

Likewise.

 

+1 I wish him all the success.

[Mario Martinez]

In regards to his motive, I am pretty sure that he provided them at the request of members of this forum when we were discussing high resolution file formats.

 

Here is where this thread comes from: http://www.computeraudiophile.com/f8-general-forum/we-dont-need-no-stinking-hi-rez-26580/index16.html#post486937

 

Maybe I should have placed that reference at the OP

 

Thank you !

[徐中銳]

Good for you who are retired and have time to write. I'm very busy and will try writing when I can.

 

In short, I requested many things when I was young but my late parents had the good sense to deny me.

 

Regular readers of Darko and Blake, please summarise for me what motivates them. And what has and is transpiring since I abandoned that abominably « stinking » Thread after my #52

 

Incidentally, an antecedent example here that reeked of Ricardo's habitual fallacy :

The purpose of these files is to evaluate the difference between formats of what, in mine and in others' opinions, are very high technical quality recordings.

That quality is what makes them (more) suitable for the task.

Of « quality » I've quoted Blue Coast and issues with « second generation » (which R « fail to see the relevance of » !) from Andrew Everard :

 

So' date=' the DSD is the one to go for. then? Not so: the site says that ‘West Side Story was originally recorded to 9624 PCM. The 9624 WAV files are the original digital file generation sent to us. [u']The DSD and FLAC files are considered second generation and made from conversions using our Blue Coast conversion methods.[/u] DSF and FLAC will offer the convenience of metadata that the WAV and DFF files will not.

 

‘After several blindfold tests, it is our opinion that the 9624 wav files sound the best, followed by DSF or DFF and after that the FLAC 9624. The difference is minimal. We suggest you purchase files for your best performing home DAC. The DACwill make more difference than the file type.’

 

Plus Jud, and I too, have mentioned « delta-sigma » and DSD « upcoverting »

 

Well, earlier, what kumakuma thinks I « seem to be implying » reflects his own...

 

What I ask readers to consider is whether this Thread's tests are an (using kumakuma's word) « commercial » advantage for PlayClassics ? Participants align their experiences by the standards provided from a PCM recording and its subsequent « protools » processing ?

 

Perhaps Forum members connected to Native DSD should unashamedly voice their production and delivery views, for example, Jared Sacks :

 

To me' date=' DSD's superiority has to do with emotion, depth, and how the sound leaves the speaker. It's not a block anymore in the way it dissipates. When you listen to PCM, you can literally hear it as a block of sound coming out of the speaker. That doesn't happen with DSD. There's air around the sound. At the end of the day, we are talking about the air around the sound.

 

In our business, we have to do post-production, but not all the time. I always make a mix-down into stereo. The surround channels go directly to an A/D converter, so they don't go through a mixer, and I try to leave them like that. Then I make a master without going through post-production (without going through the sigma-delta converter again).

 

The moment I have to change levels or do some EQ, I have to go through the mixer, and that means going through the sigma-delta again, which lowers the quality. Of course, it's all high DSD, but you have to go into DXD if you do post-production, and there's really no way around it. This problem will be solved in the future. But we are talking about further research, which costs money, at a moment when there is not much to be made selling to recording companies.

 

When you listen to my raw data, and you compare it to the post-produced recording, there's a difference in the air around the instruments and the depth. There's a degradation of sound. It's slight, but it's there. It's unfortunate, but there's nothing we can do about it, because we have to go into the sigma-delta processor again. As with any other audio signal, if you have to keep on processing, it will change.

 

You may ask, given that, if there is a difference between the sound of 192 and DSD? You have to have a really good system, and it also depends on the repertoire, to hear the difference. I still do, especially because of the dynamic range. When I down-sample to 192, you can hear that it's PCM, absolutely.[/quote']

[esldude]

I don't know if Mario uses ProTools or not. The 24/96 are said to have been made with no processing not even level adjustment. That should be as pure as you will get in terms of literally no processing. None. I, and others applaud his effort and his results are equally worth praise.

 

Now yes he needed some software to provide the downsampled versions which appear to have been processed no other way at all other than processing required for new sample rates and bit depths.

 

The picture in your post with the Jared Sack's comment shows some Grimm Audio LS1 speakers which use Hypex amps. Yet despite a sort of processing in those, DSD comes thru flowing as music rather than as PCM blocks. Laughable is what that is.

[PopPop]

I don't know if Mario uses ProTools or not. The 24/96 are said to have been made with no processing not even level adjustment. That should be as pure as you will get in terms of literally no processing. None. I, and others applaud his effort and his results are equally worth praise.

 

. . . . .

 

+1 These files appear to me to be the best opportunity I have had (to date) to evaluate my system and my listening ability. What I am able distinguish between the sound of these files (if anything) will certainly guide me in selecting future equipment upgrades and music downloads.

 

Thanks, Mario!

[The Computer Audiophile]

Hi guys - Lets try to stay on topic. There is clearly a large language barrier at play here. I suggest ignoring a post or two or three if you've been offended by someone who is more comfortable writing in a language other than English. Plus, nobody is going to make this thread better by dragging out the hard-to-decipher conversation already going on between a few people.

 

Not pointing fingers. Just want to keep this from going too far south.

[esldude]

Hi guys - Lets try to stay on topic. There is clearly a large language barrier at play here. I suggest ignoring a post or two or three if you've been offended by someone who is more comfortable writing in a language other than English. Plus, nobody is going to make this thread better by dragging out the hard-to-decipher conversation already going on between a few people.

 

Not pointing fingers. Just want to keep this from going too far south.

 

I agree there is a considerable language barrier. And nothing to be ashamed of as for myself I have no ability to speak of in any other language. The style of posting with pictures and references to others may be an attempt to bridge that barrier by means of overall concepts illustrated outside of language. Seems like a creative idea. In fact it might work better with less language and more symbolic content.

 

Yet, I am having a very hard time understanding what the person is trying to say. I am no dolt. Nor has anyone posting in this thread deserved that title in my opinion. I am pointing fingers at Wilhelm. Not finger pointing as accusing him of doing something wrong. Just pointing fingers in the sense I am not understanding him and he might be someone worth understanding.

 

Might I suggest the person whom we are having a hard time understanding perhaps shoot some ideas via PM's to someone who might can help. I don't know who it might be. If nothing else if he sent them to me, I could say "I don't get it", or this is what I think you are saying, and that person could adjust as more than likely what I think he is saying is not what he really is trying to say. Perhaps if he shot a post in his native language through Google translate and then PM'd some people for help cleaning it up.

 

I know it is a lot to ask, and were it me I would probably not bother. Otherwise we have very little communication going on. Whether someone is much much smarter than someone else or much less smart, if you can't manage to communicate little worthwhile will come of it.