Jump to content
IGNORED

A toast to PGGB, a heady brew of math and magic


Recommended Posts

I'm considering acquiring a new (for me, could be second hand) DAC and it could be a T+A DSD8 of which I would have no use of pre, headphones, PCM via chips capabilities or a HOLO with the idea of keeping PCM PCM through use of R2R (and use it DSD only with my DSD files). If that makes sense. But I wonder if "Pure" PCM exists any more than "pure" DSD since, at recording stage, all ADC I know of are Sigma Delta Modulation based. This PGGB thing, the very fact that I don't doubt that well respected AS contributors actually get a SQ improvement from their DAVE DAC, makes me wonder if the the dip dig in the encoded material performed by PGGB is not simply proving there's no absolutely right path for reproduction, be it PCM or SDM, for DSD (besides offers which musical content never appealed to me) is never PCM stage free nor PCM free from SDM origin.

 

Just an humble question by someone who wonders if it makes any sense to have PCM capability built in his next DAC (while I find myself listening more and more to vinyl and thus maybe in the mindset to think that PGGB is another proof that digital, too, has its deep inherent unsurmountable flaws). If yes I guess PGGB would be to be added to the Holo solution. The fact that the May is claimed to benefit from PGGB rules out the hypothesis that the DAVE would benefit for being a SDM DAC outputting PCM only, I guess

 

 

 

 

HQ Player 4 Mac Mini M1

Link to comment
1 hour ago, Zaphod Beeblebrox said:

A DAC need not be pure PCM to take advantage of PGGB, Instead, the DAC would need to be able to accept higher rate PCM (preferably 16FS or more), which both Holo and DAVE satisfy. 

 

The idea here is to provide the DAC with a highest rate remastered version of the original music track that the DAC would accept (for example, this could be 705.6kHz or 1411.2kHz etc) and if these remastered track is as close to a true recording made at those rates, then the DAC will likely benefit  from these remastered tracks as a bulk of the heavy lifting has already been done.  If the DAC does not do any further processing (like Holo in NOS mode) or if the DAC is able to take advantage of the higher rate PCM input (such as Chord DACs) it is likely to benefit even further.

 

There are DACs that accept up to DXD rates and they may internally resample to much higher rates. These DACs would benifit only to the extent to which they can take advantage of DXD rates.

 

You raise a good question regarding sigma delta modulators being used in the ADC conversion process, and Ideally one would think it would be better to have the signal direct from the sigma delta modulator  and not convert to PCM. But the problem is most modern ADCs use multi-bit sigma delta converters which is not the same as single bit DSD format and cannot be directly written  in DSD format. Going from multibit to single bit DSD without high quality modulators (such as the ones used in HQP) can significantly impact  the noise floor and it is often times better/easier to just record in PCM format.  I am not suggesting one format is better than the other, it boils down to implementation and the real bottleneck/culprit is the single bit DSD format.

are there DACs mirroring the ADCs by being multibits delta sigma ?

HQ Player 4 Mac Mini M1

Link to comment
13 minutes ago, Zaphod Beeblebrox said:

There are many muti-bit  sigma-delta ADCs and there are many multi-bit delta-sigma DACs (for example Chord DAVE uses 5 bit delta-sigma), but there is no standardized way of using this intermediate format and bypassing a lot of processing, as the implementations differ  (unless you convert to DSD or PCM) . 

"standardized way of using this intermediate format and bypassing a lot of processing" : what we need ? can we imagine a brand successful in the pro recording market promoting a distribution format (the one concept I liked in MQA was the idea of correcting the specific-ADC-used-for-production's footprint) of that intermediate and selling DACs capable of reading it?

HQ Player 4 Mac Mini M1

Link to comment

I don't have enough RAM to generate PGGB files, I have a SD DAC I feed SDM via HQP per @Miska 's recommendation and logic behind HQP and thus PCM is not at its best with my DAC, PCM 8 fs capable though,  but I'll be happy to test if a proponent is willing to send/process me a file

HQ Player 4 Mac Mini M1

Link to comment
2 minutes ago, austinpop said:

 

Pardon my ignorance... what is an "SD DAC?"

 

If you're finding it works best with DSD from HQPlayer, and that it tops out at 8FS, I'm not sure how much it would benefit from PGGB.

Sigma Delta. I'm not sure either it would benefit but some of the DACs listed as benefiting shouldn't either then

HQ Player 4 Mac Mini M1

Link to comment
  • 2 weeks later...
2 hours ago, hols said:

This is a follow up of my earlier post in this thread.

Since I have some spare time this weekend I revisited the issue of which is the sweet spot for my R2R DACs Holo May and Terminator plus again. As it is still difficult to gather some friends around I have to do it all by myself. I have given some thoughts on how I should improve from my previous test and I come to the following points. First I must have been over aggressive last time to compare so many bits from 19 all the way to 64 bits because this cannot give a really good AB testing and also with so many comparisons one tends to go into fatigue and affects one's judgement especially when we are doing critical listening on the many aspects of music. So this time I would only compare 20bit with 24 bit. Second I reviewed the criteria as how to judge what is best SQ or the sweet spot. Last time I could have been too obsessed with details and neglected other aspects like the pureness of tone, sense of being involved with music. Probable reasons for my emphasis on details is because it seems to be the most objective thing that one can detect in the critical comparison whereas musicality seems more subjective feeling so that was regarded as a less important feature. That might have led to result that 24bit is preferred without realizing that the resulting music at 24bit is over-analytical and not musical enough.

 

With these in mind I started my assessment again. First piece is a Mozart sonata for violin and piano. The observation that 24 bit has more detail seems to be still true on first listening. There is more low frequency making the piano seems more grand and the violin having more low frequency harmonics. But on more careful listening these low frequency sound is not as good quality as the main body of the violin or piano. The violin tone in 20bit is also more pure and coherent than in the 24bit. There is a tinge of high frequency harshness in 24bit though not serious. 

The second piece is Beethoven symphony 7. In 24 bit the double bass comes up with better presence though again there is some loss of coherence compared with 20bit. The 20bit sound and image is again more pure and musical.  The remaining pieces basically follow the same pattern. So most probably there may be more sound(but not detail) in 24 bit but some of them are not as high quality as in 20 bit. And these might be related to some low level distortions being added to the resultant sound. And as a whole now I would prefer 20bit instead of 24 bit. So the result this time is more in line with prediction by Jussi and ZB. And it is no surprise too that this applies to both the Holo May and Terminator plus.

 

I must say it is not easy at all to find out the sweet spot of the DAC because it is very difficult to determine which one is more accurate. So this is some good experience that helps you understand your DAC more.

I congratulate you for your humbleness ( I'm humble enough to consider myself part of the system and to admit that some days no gear no nothing can make me consider my system or even the acoustics of a live event perfectly palatable). But at the end of the day, 20 or 24 bits, is it worth having a (or 2) 32 fs capable DAC and mining a 333 Mo CD into a 10 Go file with a 64 Gb machine or are differences within margin of error of appreciation?

HQ Player 4 Mac Mini M1

Link to comment
11 hours ago, Zaphod Beeblebrox said:

There is only one way to find the answer to that question ...

 

It is hard to predict if PGGB is worth the while for someone. Using PGGB does not mean one has to abandon streaming. If it brings enough of an improvement,  a hybrid approach can be used where a subset of albums can be remastered.

I don't have the machine to create the files and my DAC is 8FS limited Sigma Delta BB 1795 based that I currently use exclusively SDM fed by educated choice/preference, but I reissue my call to try a file if someone is willing to send me one. In exemple I think I own about every mastering of Kind of Blue, starting with the first CD, including original SACD, Mofi's, long format Mastersound Gold etc etc, so I guess at least one PGGB proponent has a copy I have too for comparison.

 

BTW, I thought of an alternative to KOB : G Gould 1981 Goldberg as another CD "everybody" has. But it's famous for heavy editing, splicing : I greatly appreciate and want to congratulate and thank you you for the FAQ page but before I can audition, before I can even audition on a more suitable DAC than mine, I'm under impression that only unedited live recordings are safe bets to get full benefits from PGGB's capability of a complete look forward and backward in time at any moment of the reconstruction process and that the remastering tool it is should be better handed by mastering engineers with better knowledge of the history of the recording than our guesses.

HQ Player 4 Mac Mini M1

Link to comment
1 hour ago, Zaphod Beeblebrox said:

For the best results, the main requirement is the track itself being one continuous recording to make most use of the long reconstruction filters and in general better recordings will benifit more. Knowledge of provenance and how the track was recorded is helpful if one wishes to combine multiple tracks.

From your own FAQ : "What if a song is composed of many 30 second recordings that all got digitized and mixed together? Any tap length more than 30 seconds would be using bad information. "

 

Not to mention an artist (ie Bjork) recording her voice in her bedroom and then completing the track in 3 different studios, even a classical movement is more often than the other way round made of splices of different takes. Seems to me that without knowledge there's a risk of doing serious harm in many cases with modern digital productions. 

 

Of course, analog tapes digitized in one effort escape that risk.

 

A question : I would be thankful to know how you define short length in the following (reference to HQP filters would help me digging ) "For the above reason, for short lengths, minimum phase have a significant advantage over linear phase filters. " 

HQ Player 4 Mac Mini M1

Link to comment
6 hours ago, dmance said:

I ran out of space in Dropbox for my posted PGGB'd music videos link. Use this Google Drive link here.

Thanks,

Dan

 

 

Will report tomorrow hopefully on pure audio tracks compared to HQP on my system but yet, through my laptop speakers, there is something good in the opening drumming of the Hozier video I did not get when I found and played the regular youtube 

HQ Player 4 Mac Mini M1

Link to comment
2 hours ago, mancolh said:

Went back and tried HQPD but stil much prefer MPD. I tried Audirvana from a nas but the files stored on the EX with direct playback (no player involved once started) peovide the best results. One issue... when using MPD and then paused or stopped, there is a high pitched squeal heard until restarted. So far this is just with MPD.

what is MPD?

 

HQ Player 4 Mac Mini M1

Link to comment

I have been listening to 4 PGGB processed to 8fs tracks and their sources this morning, this evening.

 

this morning I considered files as remastered and OK to apply my usual process : HQP output everything as SDM and applied the same filters and 7EC modulator to A & B in A/B

After reading the afternoon remarks I set HQP as bit perfect but that raises several questions when comparing to HQP processes:

level matching?

is it OK then to search for the best matching HQP filter per track for the unprocessed?

doesn't it boil down to compare PCM vs SDM performances of my DAC?

I returned to outputing everything SDM with 128/EXT2/7EC settings in HQP. This was after I listened to a Keith Jarrett Standard 2 track sourced from Japanese SACD from recent German remastering, first PGGBed via "bit perfect" HQP, then via 128/EXT2/7EC HQP and did not find much difference

Actually I did not then hear much difference with the dsf file either.

This is not to dismiss PGGB. Actually if it was a flavor in HQP, I would have picked it and generally preferred the PGGBed files, but not so much so this evening that I consider buying a 64 Gb RAM machine a few weeks after investing in a Mac M1.

I spent less time comparing tonight but I was more indulging, less sensitive to harshness or my ears/brain were in a better mood/shape : even Lana del Rey's Born to Die sounded good loud and so did Neil Young's Hurricane (24/176.4 source).

This morning I was much more sensitive to harshness and ended with strange filter choices such as Sinc L for the 24/44 Lana and mqa mp for Hurricane. This morning I found the PGGBed files (but for the Lana, more carved, detailed but overboard in so doing) more palatable while being unsure of where accuracy lies, made me think of people claiming SDM is too soft and the other party claiming that it is PCM that is edgy. While everything was output SDM HQP processed, the soft tracks were the PGGBed while the ones I might have feared to be dubbed edgy if I was leading a demo were the sources.

Like for the Hozier video there's something in the drumming in the Neil Young with added realness cues when PGGBed and Barenboim's piano from 24/96 source had deep colored left of the keyboard notes that were more charming PGGBed.

That was consistent morning and evening but perceived more faintly this evening.

The most surprising was that I was expecting more details (echoes, recording venue acoustic signature, trails, reverb, etc) while it was the other way round, the presence region being less prominent with PGGBed files.

That test is one more tick in the Pros to get a new DAC with good PCM capability (Holo Spring 3 probably) and I will reevaluate in that context.

that's against what ZB wrote this afternoon but I think that even with a DAC performing better as SDM it's worth keeping one's settings and listen to a PGGB feed

 

HQ Player 4 Mac Mini M1

Link to comment
1 hour ago, Zaphod Beeblebrox said:

That is one way to look at it, the best PCM you can provide vs the best SDM you can provide.

 

With 8FS PCM input, the differences are going to be less pronounced in comparison to 16FS, so your findings are not entirely surprising. 

 

Using PCM files from PGGB and processing then using PCM filters in HQP and outputting at the same rate has  no positives to it.

 

What you did is little different, I am not entirely sure how HQP handles PCM to SDM conversion, if it uses 8FS PCM and directly convert it to SDM without downsampling it, then I can see how that can be beneficial. 

 

If you have a M1 Mac with 16GB RAM, then you can use it to process PCM files, DSDs will be harder. This way you may be able to try few more tracks. Also if you are going to use HQP in SDM mode, I will recommend either output 64bit from PGGB or turn off noise shaping. You will have better results.

now I regret not picking a 16 GB RAM, only have 8

HQ Player 4 Mac Mini M1

Link to comment

@Zaphod Beeblebrox If I remember correctly @Miska recommends to output to at least twice the source rate when applying convolution for so called DRC.

Regarding PGGB what is your recommendation ?

Use PGGB, LPSU etc only, to tailor the sound and no DRC?

PGGB up to say 8fs for a 16fs capable DAC and have HQP do the last mile?

Add a new offline step for applying convolution filters? which SW does that, HQP pro?

Use an analog solution ? (ie my Meyer CP10 parametric eQ does marvels but I might opt for digital convolution for the sake of convenience, switching filters on fly rather than fiddling with controls)

Use a digital external eQ box but is there any 32 fs capable?

HQ Player 4 Mac Mini M1

Link to comment
  • 2 weeks later...
6 hours ago, austinpop said:

 

Hi Chrille, 

 

I find PGGB to improve everything — from the crappiest recording in my collection to the best. In general — and I'll defer to @Zaphod Beeblebrox to weigh in if this is predicted by the math — the improvement is greatest with Redbook albums, and smallest with DXD. Stated another way, the greater the upsampling multiplier, the greater the apparent improvement. Here are a few examples:

  • Redbook: Haydn Creation (Arkiv, mid-80s) (16x multipler)
    - I love this performance, but never really thought of this as a great recording. PGGB has just transformed this album.
  • 24/96: the aforementioned Sibelius BIS recordings by Osmo Vänskä and the Minnesota Orchestra (8x multiplier)
    - These are already-great modern recording, but improved even more with PGGB. Are they "transformed?" Perhaps not, but I would never go back, having heard the PGGB version.
  • 24/352.8: The 2L Arnesen Magnificat album (2x multiplier)
    - This DXD version is considered reference quality already. Yet, PGGB improved it in important ways. 

All that said, a PGGBed lower-res mastering does not leapfrog a PGGBed hi-res mastering, although the delta between the 2 PGGBed versions will certainly be smaller. Anyone who wants to try this for themselves should just go to the 2L testbench and try the experment with some of the samples there.

 

 

 

Check your PM.

 

As regards PGGB vs. MScaler, this is really something you should evaluate for yourself. I have always asked that readers of my reviews and postings not use my findings as a basis for making purchasing decisions. At best, my findings are a guide, a suggestion to explore for yourself to try, and then let your ears be the ultimate judge. What is "transformational" to my ears on my system may be "hm, nice, but not amazing" to your ears on your system.

 

I actually wish I had kept my MScaler for use with streamed Qobuz music. Since you don't stream music, your decision may be simpler. If you find PGGB to your liking and sufficiently superior to using MScaler, then the sale of the latter could easily fund the purchase of a PGGB license, a kickass system for PGGB processing, extra storage for your music server, and still have a nice chunk of change left over.

 

You just have to hear it for yourself and decide for yourself. 

From your system description, you listen exclusively through headphones, is that correct?

 

HQ Player 4 Mac Mini M1

Link to comment
4 hours ago, Zaphod Beeblebrox said:

I respect Rob Watts and can see how the word 'Apodizing' can be misleading and lead to his comments. I have updated the FAQ to clarify this: PGGB - FAQ (remastero.com)

 

But... does "the algorithm (will) create transient timing uncertainty ?

BTW some Hires increment HQP apodizing counter like crazy ie latest Robert Plant. And besides a minority of acoustic jazz and classical DDD so do most 44.1 . So whats worse?

 

HQ Player 4 Mac Mini M1

Link to comment
7 hours ago, Zaphod Beeblebrox said:

Hmm. lets see:

Aliasing introduces transient timing uncertainty, in a visual sense, aliasing is akin to moiré  effect in image/video. This is already present in many CD recordings near the end of audible spectrum (i.e near 20kHz)

the reason being, the filters or A/D conversion process do not use steep enough filters to prevent aliasing fully.

 

There are two choices

  1. Be pedantic about it and retain every bit of information and use that for reconstruction (non-apodizing in the context of PGGB). This would mean retain and reconstruct moiré  effect in a grand scale.
  2. Throw away possible aliased part of the music and reconstruct what is 'good'.

At some point I provided an option (I still can if there is a need for it) to turn off apodizing option but no one (among the betas) cared for it.

Then one may ask, 'What if my CD is well mastered and there is absolutely no chance it has any aliasing'. Will you miss anything because of apodizing? The answer is, unless you are a teenager, it is very unlikely.  This is a very different question than asking will I hear a difference between apodizing and non apodizing, on that yes you will very likely hear a difference.

 

ps: I am not a teenager, I am middle-aged and I cannot hear past 15kHz.

 

(Image source: Aliasing - Wikipedia)

image.thumb.png.aa666f24135b7977cdd9e0958c58e0e8.png

I do not know how HQP computes the apodizing counter, I could only speculate that it looks for the energy content near 22.5kHz. Also with Hires, provenance matters. If it was derived/upsampled from a CD quality recording, then all bets are off.

Thank you, I'm not a teenager either, even doubt anyone my age aligns 15 KHz ; nevertheless I played with high shelves this weekend. Without correction above 350 Hz, my in-room response follows nicely the 2 lowest references at right hand till 10 KHz and is 2 dB lower at 15 KHz and growing. The 1 per L and R Hi shelves aimed at gaining 2 dB @ 10K and 4 above 15 to follow Floyd Toole's idealised Steady State, compensating for distance and absorption while still respecting my loudspeakers. It sure makes the kick drum (or contrabasses in a classical ensemble) shines. Maybe it's thanks to the 1dB gained @ 8K rather than the 4 @15 but I don't readily buy the can't hear above argument

3 TARGETS.jpg

HQ Player 4 Mac Mini M1

Link to comment
8 minutes ago, Zaphod Beeblebrox said:

Yes at 44.1kHz & 48kHz if it is possible, I understand that this is not always possible as that would require using a Mic, preamp and a D/A converter  (which is a significant investment) instead of a simple USB Mic that typically does only 48kHz and the EQ/measurement software would do the resampling at different rates. 

why would you want the initial REW measurements; you redo the filters?

HQ Player 4 Mac Mini M1

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...