Real or fake PCM63?

[irgendjemand]

what do I want more than immediate recognition of the concert hall where I made a recording....

You are so right with this Herb!!!!

I always take a short recording when the hall is nearly empty, only with some personel who talks quitely to each other or move a chair on the stage....

This might sounds strange, but in the concert halls we are always being depended on the beaming systems, air conditioning, etc. We might have LESS noise at home.

The recognition is because of the akkoustics close to the noise floor.... Yes I understand very well the difference between 'background noise in a room' and 'electronical noise' but sometimes it seems so theoratical to me when the differences in level are so large. In all publications (dither and so on) there should be a great advantage applying their solution. For me it sounds terrific after using a very low noise clock, reclocking of the signals before the DAC, using a good IV-convertor on a PCB with the right lay out. I still think that these subjects are the most important. However, I want still learn things of you and others!

One should be able to hear the 3-4 following steps/ measures at home while using an excellent CD Transport / DAT Machine, DAC, Class A Amps, transparent Cables & Laudspeakers:

1. Absolute silence (HIFI System is ON).
2. Now the Player is in Play Mode, but no signal yet can be heard (one can hear the player rotations, eventually)
3. Now one hear that the sound engineer was letting the "Master Tape" to start its run, includes the capture of the hall ambient (but no music yet).
4. The music starts.

This should be relatively easy to hear, like on some "Adagio" tracks, when the music comes relatively "late". Now, if one can adjust his HiFI-Volume so that the most laud places (fff) on the recording would not be yet hurting /damaging the ears – but "almost" (!!), then, if one can still hear the above mentioned order/steps - one is listening in an environment that enable him to go behind the 60 db hall limit.

Might be easy to agree on this. Regarding your wonderful comments (above), I have NO DOUBT that this is exactly what we all are looking for & actually already having it. 🙂

Greetings.

 

[PA0SU]

This might sounds strange, but in the concert halls we are always being depended on the beaming systems, air conditioning, etc. We might have LESS noise at home.

you must be a conductor, right?

One should be able to hear the 3-4 following steps/ measures at home while using an excellent CD Transport / DAT Machine, DAC, Class A Amps, transparent Cables & Laudspeakers:

with all my effort done, I think I have an excellent CD transport. I agree with you, however....
there is NO need for class A amps as long as the THD is better than 80 dB,
cables could be very simple IF THEY ARE EMC (see my web site),
my loudspeakers are a combination of ESL and MFB (see my web site)
The last three remarks are outside this thread !!!!

In general I AGREE with you.

Thanks for your comments. They are encouraging !

 

[spzzzzkt]

Hi PA0SU,

I just wanted to make sure we were all on the same page. 😀

I also do some recording, but the "concert halls" are usually quieter, and the "performers" often far less obliging.

http://www.mactrix.com.au/files/recordings/LDDG.mp3


I posted a link to a page discussing relationship between differential non-linearity (DNL) and THD earlier in the thread. The article talks about ADC but I believe the same relationship exists in DAC's (see second link).

http://www.hit.bme.hu/~papay/edu/DSP/inl.htm

http://www.hitequest.com/Hardware/a_dac.htm

Here is what Doug Rife says in an article written in 2002:
(see page 4: http://www.mlssa.com/pdf/Upsampling-theory-rev-2.pdf )

Differential non-linearity, like jitter, is a uniquely
digital form of non-linear distortion. Just as very small amounts of jitter can degrade
digital sound quality so too can very small amounts of differential non-linearity.
In mutibit DACs, distortion due to differential non-linearity tends to increase with signal
level. This occurs because the more significant bit transitions show more differential
non-linearity error than the less significant bit transitions. The reason has to do with the
architecture of most mutibit DACs and is beyond the scope of this paper. What’s
important here is that the most significant bits are in use only during loud musical
passages. Therefore, distortion due to differential non-linearity in multibit DACs tends to
increase during loud musical passages, which can obscure low-level detail and result in
a subjectively grainy, harsh and sterile sound.

So if our low level tests identify a dac chip with significant levels of THD at low levels, which also indicates the dac chip has significant DNL, Rife's comments would seem confirm our listening tests that the chip will also have poor audio performance.

This is certainly the case with your tests of the CD624 and Tent-DAC. The CD624 displays significantly higher levels of THD than the Tent-DAC, and also has worse audio performance.


cheers
Paul.

 

[Bernhard]

Differential non-linearity, like jitter, is a uniquely
digital form of non-linear distortion. Just as very small amounts of jitter can degrade
digital sound quality so too can very small amounts of differential non-linearity.
In mutibit DACs, distortion due to differential non-linearity tends to increase with signal
level. This occurs because the more significant bit transitions show more differential
non-linearity error than the less significant bit transitions. The reason has to do with the
architecture of most mutibit DACs and is beyond the scope of this paper. What’s
important here is that the most significant bits are in use only during loud musical
passages. Therefore, distortion due to differential non-linearity in multibit DACs tends to
increase during loud musical passages, which can obscure low-level detail and result in
a subjectively grainy, harsh and sterile sound.

This is true only for colinear DACs like the PCM63 where it could make sense to adjust for best performance around full scale. Normally the MSB transition occurs at digital zero crossing and produces the low level distortion and should be adjusted.

 

[spzzzzkt]

What’s important here is that while differential non-linearity is found in all DACs, it has
no counterpart in the analog domain. Differential non-linearity, like jitter, is a uniquely
digital form of non-linear distortion. Just as very small amounts of jitter can degrade
digital sound quality so too can very small amounts of differential non-linearity.

I'd omitted the first line of the quote.

I'm assuming you mean that the increasing distortion with increasing signal level is a feature of co-linear dacs?

I've done some rough tests on my existing Y's with a 997hz 0dBFS test tone, and the results show very high levels of harmonic distortion - peaks of 55dB below the fundamental. This is with a pair of chips in a balanced config. Not really what I expected to see, and this could well be an issue with the HDP2 convertors. I'll have to do some more checks to confirm.

cheers
Paul

 

[PA0SU]

Paul,

This is certainly the case with your tests of the CD624 and Tent-DAC. The CD624 displays significantly higher levels of THD than the Tent-DAC, and also has worse audio performance.

Could you explain? I do not know what you mean. Is the CD624 the one I called CD624-U?

 

[spzzzzkt]

There are significant harmonic "spikes" on both CD624 spectrums, extending futher up the frequency scale and of greater magnitude than displayed by the tent-dac.

CD624
http://www.diyaudio.com/forums/attachment.php?s=&postid=1512157&stamp=1210976602
CD624-U
http://www.diyaudio.com/forums/attachment.php?s=&postid=1515065&stamp=1211276548

Tent-DAC
http://www.diyaudio.com/forums/attachment.php?s=&postid=1512602&stamp=1211032450

What I can see in those spectra is significant harmonic distortion in both CD624, both showing a worst harmonic 50dB below fundamental. The Tent-Dac shows worst harmonic distortion of something like 58dB below the fundamental. If you take THD as being the difference between the fundamental and the average of the first 3 to 9 harmonics then the Tent Dac has significantly lower THD than either CD624 even from simply eyeballing the harmonic peaks. The rising noise floor between 16-20khz and general low level "grunge" on the U is apparent compared to the other two units.

The measured figures of this unmodified CD624 are better or at least equal to those of my PCM63P's end/or PCM63P..K's in the TantLabs-DAC-construction, but the sound from it is UNCOMPARIBLE WORSE !

The measured figures of the CD624-U are significantly worse than the tent dac, so there is no surprise it sounds worse. On the basis of your tests I rank the results from worst to best: 624-U, 624, Tent-Dac.

 

[spzzzzkt]

I was going to edit previous post but timed out.

A revision (after some image editing)
The U has a worst harmonic 52dB down from the fundamental which is 2dB better than the 624, however the noise floor is 2-3dB higher than the 624. Based on my listen tests the higher noise floor will sound substantially worse than the effect of 2dB difference in worst harmonic.

PA0SU, I'm interested in how you characterize the difference between the Tent DAC and your modded 624? I'd expect that they would be very close, the 624 perhaps slightly less focused and slightly less precise in imaging compared with the Tent DAC?

 

[spzzzzkt]

Scaled to match....

Yes, it's a bit of a hack job. I've shifted the scale so that the peak of the fundamental is at the same point on both graphs, then resized the Tent-DAC pic so it matches the CD624U to within a pixel or so.

This should make the difference between the two absolutely clear.

 

[PA0SU]

[QUOTE
I've done some rough tests on my existing Y's with a 997hz 0dBFS test tone, and the results show very high levels of harmonic distortion - peaks of 55dB below the fundamental.
[/QUOTE]


I generated me a new test disk (977 Hz) with signals:
silence, -78 dB with 6 dB-steps up to 0 dB.
It is amasing how different the cicuits behave on the different levels. It may well be that the figures with -60 dB are excellent and with -54 dB or -66 dB about 10 dB worse.
So, sorting out chips should be done with a lot of levels !!!

 

[spzzzzkt]

I suspect this is something to do with the converters in the HDP2 - I've tried running a test tone out of my laptop in to the HDP2 and the harmonic peaks are only slightly improved.

I do agree that it's probably worth testing at other levels than -60dB.

I've done a first pass on the 8 chips A'af sent me to select good pair from. These don't seem to be quite as good as the best of the previous four, so I might have struck gold with the good pair from the originals. All the chips in this batch have a "39" code on the bottom plus a 2 alpha code. The previous chips have a "48" code.

 

[irgendjemand]

All the chips in this batch have a "39" code on the bottom plus a 2 alpha code. The previous chips have a "48" code.

Paul, I had once the same problem. I remember that I was very fond of the "48". Unfortunately, I can’t tell anymore if I indeed ended with a pair of 48 or with a mix with 39. I am almost sure however that I ended with a pair of “48” having different code numbers. Something like AD and EF (or very similar). Can't take them off at the moment

There are however PCM63P-Y, BB Japan, 0323 E9 027 (see also on Herb post #215) .

Greetings!

 

[PA0SU]

quote
PA0SU, I'm interested in how you characterize the difference between the Tent DAC and your modded 624? I'd expect that they would be very close, the 624 perhaps slightly less focused and slightly less precise in imaging compared with the Tent DAC?
unquote

(sometimes I cannot qoute in the normal way....)

It is just the other way around: the modified 624 is best, immediately followed by the TentDAC. At loud pieces of music the low level details are slitely better.
The '624-U' sounds bad, smears all together when the music is complicated, etc. Even with low level music in a reverberating environment (church) the depth and the reverberation is much worse....
It is still my meaning that the 'small' differences in the measurements we are doing now, is of less importance than a low jitter clock etc. (I stated it before).

I swapped the P's in the TentDAC with the K's in the modified 624. NOTHING changed in the sound, and, and,.... The 624 remains superior to the TentDAC. They use the same PCB! but in the modified 624 I treated the reclocking before the DAC's better......

 

[PA0SU]

Paul, I had once the same problem. I remember that I was very fond of the "48". Unfortunately, I can’t tell anymore if I indeed ended with a pair of 48 or with a mix with 39. I am almost sure however that I ended with a pair of “48” having different code numbers. Something like AD and EF (or very similar). Can't take them off at the moment

Yes, gentlemen, can you explain the meaning of the codes?
A code like:

9441 K4 030

I understand that the date of fabrication is in the year 1994 in week 41. Right?
But the other figures......

At the moment I possess 14 PCM63's of all different kinds. I CAN NOT find any correlation between the types and/or date of fabrication and the quality. My far best chip is a Y with code: 0403 G4 045 from A'af. The second best is a P (!!!!) with code: 9441 K4 035.
The rest is poor, if measured at different levels as I told before.

 

[PA0SU]

Re: Scaled to match....

Yes, it's a bit of a hack job. I've shifted the scale so that the peak of the fundamental is at the same point on both graphs, then resized the Tent-DAC pic so it matches the CD624U to within a pixel or so.

Thank you very much for the effort !!!

By the way, I swapped my best two chips into the modified 624.... NO DIFFERENCE....

Sunday, may 25, I demonstrate the differences in sound using noisy and low jitter clocks (together with Guido Tent) for 'the ESL club' in The Hague......

 

[irgendjemand]

Herb,

Good to read your mails here. Meanwhile I also read your page, includes about the ESL. Many thanks for the advice to read the page!

I am not an engineer, so I base my knowledge on a combination of what I read, see and mostly – what I hear. Readings your page, there are several topics to be discussed, for example - about the using of a loudspeaker membrane (in the HF) with having no cabinets (?!). There is also the Cables issue & the ESL, and and...

But for now (and considering that this is the Digital forum and a PCM thread), I wonder what does it mean - are you swapping PCMs and not getting a sound differences? How come? This is very strange, isn’t it?

If we talk about the Chip's quality, I do remember that once Bernhard wrote something similar to what you are saying, that a Chip can be bad or good, like your 1x “Y” and 1x “P”. I am sure that this is correct information, but my own experience always showed that a good “K” was always better then a good “P”, etc. As for the “Y” as THE leading Chip, my opinion has never been of a secret…

My own experience shows, that every step/measure which I took in my HiFi system – did influenced the sound, in one way or another: Sometime the sound differences were not such major, like - when I was swapping between 2 pre-selected pairs of “K” (which I selected before by myself). But, can one swapp different selection level s of chips and HEAR NO DIFFERENCE ????

Rightly or wrongly, as no DAC's PCB can be free of some kind of fault (I assume), even the swapping/changing THE SAME PCMs SIMPLY BETWEEN "R" and "L" - will be heard! By the way, this swapping can also be of an importance - one might get a bit better sound stage simply by doing this!

If I do understand you right, you mean that by “re-clocking” the DAC, you arrived a level where the PCMs having no any notable "character". I have no idea how this can be!!!

Another question: When you say "my best PCMs" do you mean - by measuring or by listening?

I would have love to come to Den-Haag on Sunday! Unfortunately I can't but I wish you and Guido Tent a lot of fun there!!!

 

[spencer]

Irgendjernand,

I share your view on the performance of PCM63. That is there is very small difference in sonic performance within a grade but there is quite a obvious difference among grade to grade. I have Taiwan PK, Japan PK, Japan PK2, Korea PK2 and Japan KY and each grade give me quite a difference in the sonic character.

To rank them, Taiwan PK is worst, Japan PK in the middle but K2 and KY are best but the choice will come to sonic perference. K2 is more extended in high and low freq but KY is more full in harmonics.

The differences are either more extend freq spectrum, content of harmonics, sound stage, resolution & most important noise level etc... By measurement, we are dealing with pure sine wave in steady state but in real application - music, all the signal are dynamic, and changing every second. Thus the dynamic performance is also very important to correlete to the utimate sonic performance.

I believe with the same platform, same design circuit, it is also able to differentiate the sonic charasteric of different grade of chips. Unless the design of the circuit of other part dominate the performance and then change a DAC chip will not tell the differences. Just like a resistor in the signal path will always has very very small effect to the final sonic perfromance.

I also support that a better clock with low jitter is able to improve the resoution of the DAC, less smear in the sound etc... and thus I prefer very much on the DIR9001 instead of the old chip CS8414 or 8412.

I jsut hook up my new toys CD-PRO2 transport last night and it is very obvious that the background is quieter than my DVD player as transport, the bass freq is more extend and also the mid and high sound more tasty and real.

Cheers,
Spencer

 

[spzzzzkt]

Herb,

The codes are stamped into the underside of the chip. The date code on the top is identical on each chip, but the combination of 2 x numbers plus 2 x letters seems to be unique.

All but one of the current batch are *A 39 coded, and one is *E 39. The *E 39 coded chip seems a little better than the *A code chips. I've only got a couple of the old chips out at the moment, these are ** 48 coded. I've retested the worst measuring of the first batch, and the current crop are all better in terms of worst harmonic peak by at least 1.5dB. However, if you take into account the first 5 harmonics this chip sits in the middle of the "pack".

I've got the "best" of the new batch in the D1V3 and there seems to be a small improvement in overall clarity. It's difficult to A/B so it's hard to pinpoint the differences. I'll probably have to try listening to one chip at a time - SE and mono - to see if I can identify best chips.

cheers
Paul

 

[irgendjemand]

Spencer,

I agree & share the same experience. The only thing I was not able to try yet is to fit a DIR 9001 instead of my CS8414, unfortunately.

Newly I am using a wonderfully modified Marantz CD-10 as a CD-Transport. Sound became even more realistic and completely alive. With this and the "Y", it all gives a feeling of harmony: The sound stage is big and wide (but not overdone), the orchestra’s instruments are full, voices are “in the room”, dynamics are excellent, there is no any background noise (unless recorded on the Master-Tape…), sound is clean, transparent but also full resonating. All this was there before, but not to SUCH an extent. Superb!

I was just listening to a new Master CD (Studio recording, to be released soon): I don’t think that Monitoring have ever been so “easy” and natural thing for me to do before.

Indeed, an excellent CD-Transport is absolutely essential to all our talks here: Among others I had before Meridian 207 & 500 for a try, as well as Mephisto II (!!) and a modified “Parasound CDP (with the Belt Drive)”. None of them did deliver what it was supposed to.

And we did not say one word yet about the Digital Cable…



Paul,

along side your struggle with finding the best "Y"s (good luck & it seemed that you are coming very close now!), I listened to your “quieter concert hall” from Post 223: It is a wonderful recording of the nature!! I shell not forget to ask you for a full resolution/CD-sound version (instead of mp3). It will be interesting to listen to it on the HiFi, also as a reminder - what does it mean to have a silence in the nature. Thanks a lot for this recording.

Greetings!

 

[irgendjemand]

p.s. "Y"

I'll probably have to try listening to one chip at a time - SE and mono - to see if I can identify best chips.

Paul,
I would suggest you (if you allow me) to listen to Stereo and to exchange permanently the PCMs, also between R and L (as I just wrote to Herb in one of the last posts above), and this - until you are happy. It should come to a point where all parameters will "sit down" and harmonize together: You should expect it to be not only more transparent (and less veiled) but also to have the other characters which Spencer (and I) describe above.

Even though I am not using the D1V3, Spencer and I seemingly share the same experience concerning the PCMs. You are using the D1V3 - and it should be a complete improvement, at least as comparing to the K.

As you are already using one pair of "Y", may be there is no much more improvement to expect... but my “guess” is that once you will “get it”, the all effect will be indeed strong.

Best Regrads!