[Progisus]

I don’t know how that would be possible.

Kenny, As far as I know the only changes to squeeze made by Antipodes is to the material skin. To confirm this I:
1) turned off squeeze server on k30
2) installed lms 8.4 on an iMac
3) loaded squeeze player on k30
4) loaded squeeze player on pi4 using ropieee
5) played 16fs 24b pggb tracks
6) k30 player - success, pi4/ropieee player - success
7) no C-3PO plugin installed

@taipan254 - this may help you.

BTW - sound from imac/pi4 was still exceptional

 

[kennyb123]

Kenny, As far as I know the only changes to squeeze made by Antipodes is to the material skin.

Not true, as I understand it. They advanced the version of Squeezelite with AMS 5. Previously it was an old version that was still compatible with C-3PO.

Not sure from which repo you are pulling Squeezelite, but it could be that the enhancements that enabled 16FS playback without C-3PO (and also up to DSD512 payback) were also merged into the version you pulled down. That’s how things are meant to go with open source.

 

[Progisus]

post deleted by author

 

[kennyb123]

post deleted by author

 

[kennyb123]

To confirm this I:
1) turned off squeeze server on k30
2) installed lms 8.4 on an iMac
3) loaded squeeze player on k30
4) loaded squeeze player on pi4 using ropieee
5) played 16fs 24b pggb tracks
6) k30 player - success, pi4/ropieee player - success
7) no C-3PO plugin installed

I failed to give credit where credit was due. Excellent finding!!!

 

[Progisus]

For squeezelite player the ropieee verision:

ropieeeusbridge [RoPieeeXL]

• Model: ropieeeusbridge [RoPieeeXL]
• Type: squeezelite
• Firmware: v1.9.9-1432

The antipodes version is v1.9.9-1392.

 

[Rayon]

I hated HQPlayer with all my guts and was relived to rid myself of all that tinkering and hiccups once I got the MS.
Hustle aside, MS sounded far superior to HQPlayer in my subjective opinion.
I‘ve been listening to the M-scaled M2 for a few days now and not missing DAVE for a second, seriously considering selling it.
By comparison, I could only listen to MS/TT2 for a few minutes before going back to MS/DAVE In panic mode.

Anyhow, back to the HQPlayer.
Can you set it up to pass PGGB untouched and upsample anything below 705/768 without constantly changing settings?

As a May + Mojo 2 owner who lately found PGGB and has been wondering about Dave, these posts warm my heart so much. These comments are also supported by the fact that I've been amazed by how good PGGB'd Mojo 2 is in comparison to May. I don't feel like I'm missing out with Mojo 2 at all. Said that, Mojo 2 has never made me smile the way May occasionally does. To me PGGB'd Mojo 2 is great for musical enjoyment, but for exploration of what can be done with audio, May is still the king to me. It's also somehow more effortless for longer listening. What's also interesting: with Mojo 2 PGGB is hands down better vs. HQPlayer. With May, I honestly am not sure. It breaks my heart that PGGB doesn't support linux and HQP Embedded does not support Windows (natively) as currently it's a bit of a hassle to have both and when run in virtual machine, I can't run same sattings with HQP vs. bare metal.

 

[chelseafc]

As a May + Mojo 2 owner who lately found PGGB and has been wondering about Dave, these posts warm my heart so much. These comments are also supported by the fact that I've been amazed by how good PGGB'd Mojo 2 is in comparison to May. I don't feel like I'm missing out with Mojo 2 at all. Said that, Mojo 2 has never made me smile the way May occasionally does. To me PGGB'd Mojo 2 is great for musical enjoyment, but for exploration of what can be done with audio, May is still the king to me. It's also somehow more effortless for longer listening. What's also interesting: with Mojo 2 PGGB is hands down better vs. HQPlayer. With May, I honestly am not sure. It breaks my heart that PGGB doesn't support linux and HQP Embedded does not support Windows (natively) as currently it's a bit of a hassle to have both and when run in virtual machine, I can't run same sattings with HQP vs. bare metal.

Are the PGGB benefits with May as profound as with Mojo 2?
Are you using USB to feed the data into May?
One of many reasons I chose to move away from DAVE is its sensitivity to noise (from the mains or other sources).
Is May similar in that regard?

 

[Rayon]

Are the PGGB benefits with May as profound as with Mojo 2?
Are you using USB to feed the data into May?
One of many reasons I chose to move away from DAVE is its sensitivity to noise (from the mains or other sources).
Is May similar in that regard?

Yes, I'm using USB, feeding it from Intel NUC based NAA and HQPlayer. I'm using HQPlayer's NS5 to noise shape those 7xx/32 files into 23 bits and add crossfeed. I'm also occasionally further upsampling it into 32fs with HQP.

For me May is challenging. Some mixed thoughts:

• 16fs is very clearly inferior to 32fs so I'd need to have both 16fs for Mojo and 32fs for May if I'd like to use PGGB "proper way" with May as well
• I really like crossfeed when using May with Sus/TC and PGGB doesn't offer that -> I anyway need to further process it in HQP
• It's not clear if I generally prefer the sound of PGGB compared to HQP with May
  • With Mojo 2 the preference is very clear
• Roon doesn't support 32fs
  • This is a major pain point for me personally

With Mojo 2 it's a no-brainer (ignoring that the price of PGGB is very high in comparison to Mojo 2). With May, I'd make sure that the preference is there and think through your full setup before pulling the trigger. With May HQP is as no-brainer as PGGB with Mojo 2. For Mojo 2 I wouldn't personally recommend HQP.

Also I'm using some short USB cable that came with iFi iNano. I haven't noticed any USB related problems with different cables. Mojo 2 on the other hand is very picky. I never believed in USB cables before Mojo 2, but that is like a different dac between good and bad cables (assuming laptop source and PGGB).

 

[Rayon]

Just want to elaborate still on May + PGGB. I've again been trying different filters in HQP to take 16fs to 32fs and noticed that closed-form-M sounded very natural. I've now been enjoying a lot PGGB 16fs 32bit (optimal noise shaping) -> HQP closed-form-M LNS15/NS5 23bit 32fs. I realized that closed-form-M keeps the original samples untouched, which in this case is the data created by PGGB and interpolates just the in-between samples (PGGB to my understanding shares the same philosophy). This is quite amazing combination with multiple benefits:

• HQP doesn't mess the hard work done by PGGB, letting it's beaty shine through
  • Also as HQP only interpolates every second sample in the final output, the result should be quite close to what PGGB would produce to 32fs directly
• I still get the (obvious) benefit of 32fs with May
  • The benefit from interpolation that HQP does is clear
• I get crossfeed from HQP which is super nice both with Sus and TC
• I can share the same PGGB files between May and Mojo 2
• I can still listen to these files from Roon
• Not a big thing, but still nice: 16fs are smaller than 32fs
• The sound is clearly better than using just HQP to take it all the way from original (using closed-form-M)
  • However other HQP filters still offer different sound signature. When using those, it's better to use non-PGGB'd content IMO, excluding very few exceptions (if any).

 

[LucyWu]

I'm not sure if you've tried selecting no processing in HQP and playing the PGGB files directly (HQP is very good just as a player for these upsampled files as it's gapless playback is excellent). Sounds superb, but I also like the Closed-Form filters.

Also - especially is you're using DSD, but valid with PCM upsampling in HQP - PGGB can save files at 64bits (I know more massive file sizes). HQP can process these for 32 bit output and it means noise shaping only gets done once. 32 bit PGGB files have been fully noise shaped, if you then upsample again you're noise shaping twice which the developer of PGGB doesn't recommend.

I run all my local files (upsampled in PGGB at 16fs (have tried 32fs 64bit files too!!!)) through HQP to output as DSD512. The results are worth the effort.

 

[Rayon]

I'm not sure if you've tried selecting no processing in HQP and playing the PGGB files directly (HQP is very good just as a player for these upsampled files as it's gapless playback is excellent). Sounds superb, but I also like the Closed-Form filters.

Also - especially is you're using DSD, but valid with PCM upsampling in HQP - PGGB can save files at 64bits (I know more massive file sizes). HQP can process these for 32 bit output and it means noise shaping only gets done once. 32 bit PGGB files have been fully noise shaped, if you then upsample again you're noise shaping twice which the developer of PGGB doesn't recommend.

I run all my local files (upsampled in PGGB at 16fs (have tried 32fs 64bit files too!!!)) through HQP to output as DSD512. The results are worth the effort.

Yes, I've listened to PGGB directly as well. I know that my setup has some "non-optimalities", but so far this has been the best middle ground for me personally. I really like Roon a lot and with HQP Embedded the client UI was quite a downgrade. Then I also have Mojo 2 so if I would like to use 64bit and/or 32fs for May, it means that I would need to have two separate PGGB upsamples. To my ears 32bit noise shaped 16fs still sounds amazing with HQP. Note that the noise shaping is done in the LSB of 32bit and in the end I'm noise shaping to 23bit for May. There are 9 bits in between, so I'm not sure if PGGB's noise shaping "harms" anything vs. using just dither (assuming now that 64bit is not an option due to Mojo 2). I'd guess that any effects of PGGB's noise shaping vs. dither will anyway disappear when I cut out those bits and PGGB's noise shaping work is merely wasted when using May. Especially if I'm using closed-form, which should only interpolate the in-between samples, those rounding errors shouldn't "accumulate" enough to be audible. The case may be a bit different with DSD.

I did try 64bit as well, but luckily I don't remember how good it sounded and as I'm currently happily tapping my toe and listening to this setup in which I'm not doing too big compromises in convenience, I dare not to look into that rabbit hole again

EDIT: I re-tried 64bit as well. Actually the difference is quite big, especially when processing further to DSD.

 

[LucyWu]

I did try 64bit as well, but luckily I don't remember how good it sounded and as I'm currently happily tapping my toe and listening to this setup in which I'm not doing too big compromises in convenience, I dare not to look into that rabbit hole again

I dragged myself back from the precipice of 64bit/32fs too. Life's too short even with a fast computer to process and a 20tb hard drive.

 

[Rayon]

PGGB is such another rabbit hole... I'm getting more intoxicated over time with May as well. I tried HQP filters today directly with original content and I'm starting to see my preference shifting into PGGB direction and what the PGGB dev mean when he talks about the importance of time domain. PGGB kind of "flows" more naturally. PGGB has detail, but it's not popping up and feels natural and music is kind of holistic. Also tried those noise shaped 32 bit 16fs files as a source to DSD1024 with closed-form-16M and AMSDM 512+fs. Deadmau5 - XYZ is pure ridiculousness with TC.

 

[Progisus]

My goodness this sounds great with PGGB 256 and Squeeze: