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

[kennyb123]

@Zaphod BeeblebroxIt appears that the metadata will be empty if there is an extended characters in the path to the file.
 

[hanshopf]

2 hours ago, austinpop said:

Garbage in, garbage out.

Thanks for chiming in, but I cannot really see the reasoning behind: if everything is fine with metadata before running PGGB, then everything is pointing at PGGB for handling things not properly, if problems with metadata occur afterwards. Even more so since I seem not to be the only person noticing the problem. 
For 1000$ I should able to use it without awkward workarounds, and I believe the bug can be mentioned without doing the integrity of the amazing software any harm. 
 

[austinpop]

11 hours ago, hanshopf said:

 if everything is fine with metadata before running PGGB

 

Sorry, I was not dismissing your experience. It is often the case that even if an album is not sufficiently well-tagged, it can still display fine in a music application, because the app could be defaulting to other attributes like file names as a fall back.

 

However, once you have long tracks that PGGB needs to break up, missing or incorrect track number tags can lead to tracks being displayed out of order. So, all I am saying is that PGGB does best IF all the original tracks have correct track and disc number tags.

 

I had asked you this question: do you rely on the tags shipped with the content, or do you clean your tags?

 

Could you please clarify?

 

[Progisus]

I always run my files through a metadata editor before and after processing. Part of the fun for me.

[davide256]

Now that I finally have streamer software that can handle wav pack, PGGB makes sense to use, the 50% reduction in file size and transmission bandwith for 7xx/24 is workable vs

uncompressed wav.

 

Wondering what off the shelf  streaming servers support wav pack tags and USB output at 7xx PCM? Time to move on/up from NUC products

[Zaphod Beeblebrox]

4 minutes ago, davide256 said:

Now that I finally have streamer software that can handle wav pack, PGGB makes sense to use, the 50% reduction in file size and transmission bandwith for 7xx/24 is workable vs

uncompressed wav.

 

Wondering what off the shelf  streaming servers support wav pack tags and USB output at 7xx PCM? Time to move on/up from NUC products

What software are you using now?

[davide256]

9 hours ago, Zaphod Beeblebrox said:

What software are you using now?

Audirvana

[hanshopf]

On 12/17/2023 at 8:32 AM, austinpop said:

So, all I am saying is that PGGB does best IF all the original tracks have correct track and disc number tags.

 

Hello, I checked the tags for tracks and disc numbers after four of five operas failed to show in correct order after PGGB. The tags are correct in all cases. I therefore still assume it must be a bug in PGGB (Mac version).

 

And by the way: in non of these sets were overlong tracks which PGGB had to separate. So this was not even an issue here.

 

On 12/17/2023 at 8:32 AM, austinpop said:

do you rely on the tags shipped with the content, or do you clean your tags?

 

If shipped content shows correctly I do not change the tags. 

[Zaphod Beeblebrox]

14 minutes ago, hanshopf said:

Hello, I checked the tags for tracks and disc numbers after four of five operas failed to show in correct order after PGGB. The tags are correct in all cases. I therefore still assume it must be a bug in PGGB (Mac version).

I will need more information to replicate the issue, perhaps a sample album where you are having this issue. Is the issue with the track numbers only or is all of the meta data missing?

[seeteeyou]

Is this applicable to other owners of Taiko Audio SGM Extreme with a different DAC by any chance?

 

https://www.whatsbestforum.com/threads/introducing-olympus-olympus-i-o-a-new-perspective-on-modern-music-playback.37939/page-23#post-932646

Quote

I currently use a Chord DAVE, and having 'come up' through that route, a lot of PGGB wav files. Undoubtedly to my ears the 16x sample rate, very large, pre-processed PGGB files sounded better than lower sample rates. I don't use DSD.

Recently however, and especially with the advent of NSM, I had a nagging perception that Red Book files sounded just as good if not better. I did a fairly quick comparison the other day and felt that the difference was now edging in favour of Red Book.

 

In other words, could Chord products be considered a "bottleneck" of some sort?

[austinpop]

@hanshopf

 

As my experience does not match yours, I can't be of further help.

 

As ZB has suggested, it's best to look at a concrete example, so please upload a sample album for him to look at.

[Zaphod Beeblebrox]

33 minutes ago, austinpop said:

@hanshopf

 

As my experience does not match yours, I can't be of further help.

 

As ZB has suggested, it's best to look at a concrete example, so please upload a sample album for him to look at.

He did share a sample with me and with the help of @kennyb123, I was able to confirm the issue on Mac. This is a known issue that has persisted since the beginning, the ID3Tag library I use to transfer the tags does not work when extended characters are present in the track file name. This affects Macs only, no issues on Windows (which will explain why you have not had this issue).

 

A workaround (not ideal) is to use bulk file renaming tool such as this. I plan to give it another shot to fix the issue, but do not have an ETA and cannot promise a fix.

[kennyb123]

32 minutes ago, Zaphod Beeblebrox said:

are present in the track file name

Anywhere in the path to the file.  Filename or folder name.

[taipan254]

Random question: Has anyone here experimented with applying PGGB to vinyl rips from a solid source like PBTHAL? 

 

I recently applied PGGB to a few rips I've had for awhile. On the one hand, there is a charm to these PGGB-ed rips. They sound cohesive and well-put-together. On the other hand, the resolution I get from PGGB-ed CDs and Hi-Res is not even comparable to the musical resolution in these rips. Also, the vinyl rips have different EQ applied relative to what is present on the studio-provided CDs or Hi-Res content. We're also probably hearing the entire system used in the rips (from power, turntable, cartridge through the pre-amp output that feeds the D/A converter). This is opposed to whatever feeds the D/A converter at the mastering studio. Just a fascinating comparison to listen to given how PGGB provides a near-perfect reconstruction of both digitally captured vinyl rips and digital provided directly from studios. 

 

Random rant over!

[elementze]

On 12/2/2023 at 1:21 PM, davide256 said:

So far SRC-DX is the only solution I've found that allows 384k PCM output via BNC or coax.

 

I'm wondering if any alternate endpoint solution exists where I could stream over UPNP my PGGB files to

coax/BNC outputs at 384k or higher rate? Every device I've seen is throttled to 192/24

The chord 2Yu supports 384kHz over coax.

[kennyb123]

2 minutes ago, elementze said:

The chord 2Yu supports 384kHz over coax.

I believe most Chord DACs can do up to that rate over a single coax.  The TT that I owned at one time could do this, as that’s how I used my MScaler while I waited for my TT2 to arrive.

[Progisus]

 

 

 

This works for PGGB files to 768k. Front end is an Allo Usbridge fed by roon/hqplayer.

[Atriya]

@Zaphod Beeblebrox: Is it better to use noise shaping or "dither only" when the final result will undergo further sample rate conversion (in my case 384Khz->192Khz->195.312Khz) - but not other kinds of DSP - in the DAC?

[Zaphod Beeblebrox]

36 minutes ago, Atriya said:

@Zaphod Beeblebrox: Is it better to use noise shaping or "dither only" when the final result will undergo further sample rate conversion (in my case 384Khz->192Khz->195.312Khz) - but not other kinds of DSP - in the DAC?

For 8fS output rate, yes I will recommend Noise shaping even if further conversions may happen later.

[Atriya]

1 minute ago, Zaphod Beeblebrox said:

For 8fS output rate, yes I will recommend Noise shaping even if further conversions may happen later.

 

Okay! But not for 4fS?

[Zaphod Beeblebrox]

Just now, Atriya said:

 

Okay! But not for 4fS?

For 4fS, even for normal playback, the benefits may be debatable (system dependent)

[taipan254]

Happy New Year to everyone!

 

Quick Question for @Zaphod Beeblebrox and the rest of the crew here. Can someone help me reconcile the below results from RASA?

 

Details: I got some new audio gear, which has led to some (re)experimentation on my end. I use PGGB-IT! 64, but upsampled some tracks in 128 bit and 256 bit (all below 2 mins) to see if I could hear the differences. The first, a Nirvana 16 / 44 track, produced results that I expected - as precision increased, noise-floor went down, with big gains from 128 bit to 256 bit (see below). I could hear noticable improvement in the reproduction of the song as precision increased.

 

The second track is a Rafael Blechaz 24/96 piano track. Below are the results - with this track, the noise floor seems higher on 128 bit precision vs. 64 bit precision. Further, 256 bit precision doesn't appear to offer much in incremental gains from a noisefloor perspective. Am I reading this right?

 

 

Also, I'm having a tougher time differentiating the impact of increased precision on the classical piano track vs. the Nirvana track. 

 

Any help, guidance, commentary, or explanation is much appreciated!

 

 

[Zaphod Beeblebrox]

12 hours ago, taipan254 said:

Happy New Year to everyone!

 

Quick Question for @Zaphod Beeblebrox and the rest of the crew here. Can someone help me reconcile the below results from RASA?

 

The second track is a Rafael Blechaz 24/96 piano track. Below are the results - with this track, the noise floor seems higher on 128 bit precision vs. 64 bit precision. Further, 256 bit precision doesn't appear to offer much in incremental gains from a noise floor perspective. Am I reading this right?

 

 

 

Your reading of the plots is not correct, but it is easy to misinterpret. What you are seeing is the out of band noise and not the true noise floor in audible range. PGGB aims to maximize accuracy in the audible range (i.e. up to 20kHz), this is where the noise floor remains the lowest. Because you upsampled a 96kHz track (which can have audio content till 48kHz), you do not see the true noise floor of the noise shaper in the audible range (as it is masked by the content beyond 20kHz).

 

To see the true noise floor, you will need to enable aggressive HF filtering for 2fS or higher content.

 

Below is how the plot will look for different HF filter options. You will see that if you do not change the Noise shaper or bit depth, the noise profile remains the same, and what changes is how much of the original content is retained and then there is a precipitous drop to whatever the out of band noise profile is at that point.

 

 

To better illustrate this, assume your Nirvana track was 24/96kHz, then if you had used the same settings, there would have been audio content up until 48kHz, the graph would have looked something like below (very close to your Piano track)

 

Edit: To be clear, I am not suggesting enabling the aggressive HF filter all the time, it is more for illustrative purpose. though depending on the mastering, some tracks benefit for HF filtering.

 

12 hours ago, taipan254 said:

Also, I'm having a tougher time differentiating the impact of increased precision on the classical piano track vs. the Nirvana track. 

Assuming the graphs did not bias your opinion, it is hard to say (I would have expected the opposite). PM me if you wish a trial extension so you can evaluate with full tracks.

 

 

[NoNaim]

Some may have seen on Head Fi that I've been able to use the Auralic VEGA G2.2 to play 705/768 32bit PGGB files via the ethernet connection using the LDS app.

 

I currently have a 10m run of fibre running from the network cupboard (Ubiquity UDM Pro that's fibre connected to a Ubiquity 16 switch +2 fibre) to my music room with a Startech FMC & Finisar 10gb sfp and a 1m WireWorld cat8 ethernet cable to the DAC. Works without fault when changing from 705 to 768 and back again. Very simple and clean setup with terrific sound quality.

 

The VEGA's are supposedly only 384/32 DACs and Roon limits them to this but using UPnP server with the LDS app seems to give you the extra scale.

 

Not sure if this is possible with the VEGA G2 or G2.1

[Markus8]

On 12/1/2023 at 3:43 PM, GoldenOne said:

PGGB first reconstructs the PCM data and outputs high rate PCM.

Isn’t HQPlayer capable of doing that too?