Hi-Res - Does it matter? Blind Test by Mark Waldrep

[gmgraves]

21 hours ago, sandyk said:

 George

 There are numerous albums available from HDTracks etc.that have obvious HF info well above 22kHz, not just noise.

 Regards

Alex

I’m sure there are. The music world has been recording at 24/96 or higher for close to two decades now. 

[Rexp]

2 hours ago, gmgraves said:

I’m sure there are. The music world has been recording at 24/96 or higher for close to two decades now. 

I think this is the problem, alot of redbook stuff is poorly downconverted from 24/96 masters. 

[lucretius]

33 minutes ago, Rexp said:

I think this is the problem, alot of redbook stuff is poorly downconverted from 24/96 masters. 

 

I don't think that's the case. I've heard a lot of bad or mediocre 24/96 files. I'm willing to bet that a bad redbook file is spawned from a bad 24/96 master.

[Rexp]

5 hours ago, lucretius said:

 

I don't think that's the case. I've heard a lot of bad or mediocre 24/96 files. I'm willing to bet that a bad redbook file is spawned from a bad 24/96 master.

And I'm willing to bet you are wrong haha

[Steve B]

On 10/29/2019 at 5:34 PM, Rexp said:

Why, is there a problem? 

I'm not sure there's a problem, but hoping someone can confirm what I see comparing the X and Y files in Deltawave.

 

The first screen shot is the original x and y files.  The white line is the 96 kHz file.  The blue line is the 44.1 kHz file.  The blue line should be hidden behind the white line from 0 to about 21-22 kHz.

 

The second, third, and fourth screen shots are the original 96 kHz file (white) and sample rate conversions (blue) that I have done from the original 96 kHz file to 44.1 kHz 24 Bit with SoX, then converted to 44.1 kHz 16 Bit with (2)Foobar2000 dither, (3)Reaper dither, and (4)no dither.

 

The results are not what I expected, so I hope someone can confirm.

 

I'll try other dithers.  

 

(I've varied from the Deltawave default corrections a little, using only the phase drift correction,  But using the default corrections makes little difference in this case any way).

 

[esldude]

1 hour ago, Rexp said:

And I'm willing to bet you are wrong haha

You greatly over estimate the influence of poor downsampling.

[Rexp]

3 hours ago, esldude said:

You greatly over estimate the influence of poor downsampling.

And you think RTR and CD sound the same, nuff said

[botrytis]

All things matter. I have good sounding vinyl and tapes and good sounding digital. Digital is way less forgiving on bad sounding recordings so the flaws are more inherent in the music. With analog, much of the flaws hide behind the noise from the instrument of production, TT or tape player.

[esldude]

3 hours ago, Rexp said:

And you think RTR and CD sound the same, nuff said

Someone has a reading  comprehension problem along with a listening issue.

[ralphfcooke]

Elsdude, that does sound like an 'ad hominem' comment to me, care to rephrase it?

[esldude]

4 hours ago, ralphfcooke said:

Elsdude, that does sound like an 'ad hominem' comment to me, care to rephrase it?

Reading comprehension.  I posted comments that comparing RTR, CD, and LP, that RTR, and CD sounded similar in general balance, and that LP was an odd man out.  This was on three different high quality systems with friends using a couple dozen recordings available in all three formats.  I never said that RTR and CD sounded the same, only that we were surprised they were more similar than LP.  That LP was in every case very different while RTR and CD has similar balances in FR and general impression of the recording.  Obviously RexP didn't comprehend what he read.  My other choice would be to assume he intentionally misrepresented it. 

 

Rexp seems to have this thing that LP is the optimum and the reference by which to judge and everything else is inferior.  I'd call that a listening issue.  He even has commented that if you've only listened to digital versions you don't know how something is supposed to sound like.  Again a listening issue.  Especially as LP's are almost always in the past made from tape, and listening to tape has a good chance of being the correct sound vs a medium we know requires considerable alteration to even make it onto the LP disc and be playable.  So if I hear quality tape, and I then listen to LP and CD with LP sounding odd and CD not, well which is the right sound? Rexp would say the LP.  A listening issue. 

[esldude]

12 hours ago, Steve B said:

I'm not sure there's a problem, but hoping someone can confirm what I see comparing the X and Y files in Deltawave.

 

The first screen shot is the original x and y files.  The white line is the 96 kHz file.  The blue line is the 44.1 kHz file.  The blue line should be hidden behind the white line from 0 to about 21-22 kHz.

 

The second, third, and fourth screen shots are the original 96 kHz file (white) and sample rate conversions (blue) that I have done from the original 96 kHz file to 44.1 kHz 24 Bit with SoX, then converted to 44.1 kHz 16 Bit with (2)Foobar2000 dither, (3)Reaper dither, and (4)no dither.

 

The results are not what I expected, so I hope someone can confirm.

 

I'll try other dithers.  

 

(I've varied from the Deltawave default corrections a little, using only the phase drift correction,  But using the default corrections makes little difference in this case any way).

 

Everything looks right to me.  Sox is a very good resampling algorithm.  Foobar may use Sox or something else depending upon how you have it set up.  Reaper has a mediocre resampling algorithm built in.  Secret Rabbit code I think it uses.  

[sandyk]

Perhaps a small part of the audible differences ?

 

Zero crossing - Wikipedia

https://en.wikipedia.org › wiki › Zero_crossing

A zero-crossing is a point where the sign of a mathematical function changes represented by a ... At audio frequencies,
such as in modern consumer electronics like digital audio players, these effects are clearly audible, resulting in a 'zipping' sound
when rapidly ramping the gain, or a soft 'click' when a single gain change ...

[Steve B]

1 hour ago, esldude said:

Everything looks right to me.  Sox is a very good resampling algorithm.  Foobar may use Sox or something else depending upon how you have it set up.  Reaper has a mediocre resampling algorithm built in.  Secret Rabbit code I think it uses.  

Thanks esldude.  Independent of the resampler, the strange thing I'm finding is that everything (including other program material) that I convert from 44.1 k 24 bit to 44.1 16 bit using Foobar2000 16 bit dither introduces a difference in the spectrum from 17 khz to 22 khz.  And it looks to be about a 5 dB difference.  I don't find this with any other dither, or even with Foobar2000 24 bit dither.  And I never expected dither to change the frequency response (I thought any problem would probably be the resampler), but I find this consistently with the Foobar2000 16 bit dither.  I really love Foobar2000, but this is very strange.

 

And I see the similar thing in the original x and y files (picture #1).

 

I'm really sorry the charts are not labeled boldly, but I'll be sure to do that next time.

 

They are:

 

1) original x and y files from sandyk (96k vs 44.1k)

2) 96k vs 44.1 Foobar2000 16 bit dither

3) 96k vs 44.1 Reaper 16 bit dither

4) 96k vs 44.1 16 bit no dither

[fas42]

1 hour ago, esldude said:

Rexp seems to have this thing that LP is the optimum and the reference by which to judge and everything else is inferior.  I'd call that a listening issue.  He even has commented that if you've only listened to digital versions you don't know how something is supposed to sound like.  Again a listening issue.  Especially as LP's are almost always in the past made from tape, and listening to tape has a good chance of being the correct sound vs a medium we know requires considerable alteration to even make it onto the LP disc and be playable.  So if I hear quality tape, and I then listen to LP and CD with LP sounding odd and CD not, well which is the right sound? Rexp would say the LP.  A listening issue. 

 

More usually, the degree of tweaking that's been applied to the source components that deliver the tracks ... it's been an interesting journey, over many years now, listening to the audio friend's rigs up the road - he applies a great deal of care to both formats, and the "which is better?!" aspect has seesawed back and forth, over that time. Recently, vinyl got a major spruce up - it had fallen behind, because so much more attention had gone into the digital side of things.

 

This is not about FR and all the obvious technical measures - it's whether the subjective sense of the musical event fully falls into place; and overrides any irritations due to the limitations of the format - when attended to sufficiently, both LP and CD do a fine job - and there is no point in picking "winners".

[sandyk]

Re X and Y

 Note also the much higher noise level right from the start of track X, which undoubtedly continued right through the  track 

 Y noise level is also shown at a much higher magnification level which can be readily seen in the attached.

 

Click on the image several times for a much larger image.

[Confused]

I tried the infamous X & Y files today.  To me, they both sounded like reasonably well recorded pieces of music.  I had a very slight preference for Y, but certainly found nothing that I would consider irritating or offensive with version X.  Does anyone actually think there is a big / significant difference between the two, or would actually be irritated by listening to one particular version?

[copy_of_a]

16 hours ago, Steve B said:

Independent of the resampler, the strange thing I'm finding is that everything (including other program material) that I convert from 44.1 k 24 bit to 44.1 16 bit using Foobar2000 16 bit dither introduces a difference in the spectrum from 17 khz to 22 khz.  And it looks to be about a 5 dB difference.  I don't find this with any other dither, or even with Foobar2000 24 bit dither.

Apparently Foobar2000 16bit dither is agressively shaped by default.

Too agressive... IMHO.

I'd suggest to use TPDF (flat) or a moderately shaped dither.

[FredericV]

On 11/1/2019 at 10:10 PM, sandyk said:

NOBODY has said anything about downloading X twice and hearing a difference between both copies . Several prominent members have also questioned your methodology ,even Dennis on this occasion who was also readily able to hear the differences.

 

My files are different, and some indeed hear differences.
 

On 11/1/2019 at 10:10 PM, sandyk said:

If you are talking about the checksums being identical, then that is outdated  dogma and easily proved to be incorrect, just like Martin Colloms verified by way of 6 separate DBT sessions with my supplied .wav files from Dire Straits-Love Over Gold several years ago and published the results.

This applies to both A and V files ,which I am able to generate readily from the same source file, yet can both SOUND and LOOK a little different, especially when using high resolution monitors with in the recent case as one member did, side by side on his 4K monitor.

Please send those files to me for further analysis.

[FredericV]

On 11/1/2019 at 6:01 PM, esldude said:

So did you do the down sampling or download each sample rate from tidal? Or more generally how were each of these created.

 

 

[sandyk]

4 hours ago, FredericV said:

Please send those files to me for further analysis.

 

!. I no longer have the exact same files that were evaluated in the 6 DBT sessions several years ago.

2. More recent examples have been checked by al.fe's parent Asian company and a 50 page report prepared.

( In case you don't already know, al.fe designed my internal LG GGW H20L Blu Ray writer and has several patents in the Optical storage area.)  They found no differences in the Digital realm, just like Mansr found none in the samples used in the sessions both he and Mani participated in. The files would need to be checked in the Analogue realm at the output of a DAC.

 I have far better comparison files available  these days, where you still will not find any differences using current Digital techniques.

 They are Music Video files where you can both see and hear differences at the same time, if you use a high quality Monitor and play them through a revealing audio system.

One member here viewed a pair of them side by side on his 4K monitor and reported this :

"Wow Alex.  I looked at the first two links...big quality difference in the video.
How are they different?  Or, what did you change when recording them?" 

[danadam]

22 hours ago, sandyk said:

Note also the much higher noise level right from the start of track X

That is the shaped dither. Here's the spectrogram of the beginning of that track and the spectrum of the first 0.6s:

 

For comparison, a track with non-shaped dither has visually lower noise:

 

but of course the spectrum of that noise is quite different (note a difference in dB scale):

[Steve B]

16 hours ago, copy_of_a said:

Apparently Foobar2000 16bit dither is agressively shaped by default.

Too agressive... IMHO.

I'd suggest to use TPDF (flat) or a moderately shaped dither.

Thank you copy_of_a.  That's exactly what makes the difference in the spectrum.  Shaped dither.  Since the files were apparently made and processed in SoX (not Foobar2000), that would have required a command of dither -s, which is very common in most of the examples in the SoX manual, and produces exactly this difference in the spectrum.  In SoX, If dither is not specified and the file being made is 16 bit from higher, flat TPDF dither is applied automatically. That produces no visible difference that I can see.

 

Foobar2000 has no dither options that I can find, and appears to apply shaped dither when dither is selected.  The spectrum difference looks very similar to the SoX shaped dither.

 

But I see from the other discussion that fas42 saw exactly this difference much earlier than I did.

[rvb]

The worst thing ever happened in music was mp3. 

 

Second worst : the compact disc.

 

If they made just a vinyl disc 24/96 then everything would be fine.

[marce]

12 hours ago, sandyk said:

 

!. I no longer have the exact same files that were evaluated in the 6 DBT sessions several years ago.

2. More recent examples have been checked by al.fe's parent Asian company and a 50 page report prepared.

( In case you don't already know, al.fe designed my internal LG GGW H20L Blu Ray writer and has several patents in the Optical storage area.)  They found no differences in the Digital realm, just like Mansr found none in the samples used in the sessions both he and Mani participated in. The files would need to be checked in the Analogue realm at the output of a DAC.

 I have far better comparison files available  these days, where you still will not find any differences using current Digital techniques.

 They are Music Video files where you can both see and hear differences at the same time, if you use a high quality Monitor and play them through a revealing audio system.

One member here viewed a pair of them side by side on his 4K monitor and reported this :

"Wow Alex.  I looked at the first two links...big quality difference in the video.
How are they different?  Or, what did you change when recording them?" 

This couldn't be two identical files are different, could it...

As the files are identical if you sent them me and I renamed them how would you tell which is which... Want to try it.