HQPlayer Impressions and Settings Rolling Thread

[borkenarrou]

Anyone got HQPlyaer to work with Apple Music (Lossless) via NAA with Apple Silicon Macs.

 

[BShaw]

Anyone got HQPlyaer to work with Apple Music (Lossless) via NAA with Apple Silicon Macs.

Pretty certain it's Qobuz and Tidal only.

 

[dericchan1]

Anyone got HQPlyaer to work with Apple Music (Lossless) via NAA with Apple Silicon Macs.

https://audiophilestyle.com/forums/topic/70792-apple-music-stream-to-hqplayer/#comment-1305966

You just need an iPad and a rpi4

 

[borkenarrou]

https://audiophilestyle.com/forums/topic/70792-apple-music-stream-to-hqplayer/#comment-1305966

You just need an iPad and a rpi4

This is great!! will experiment with it today, my only question why ipad has to be connected to rpi4 usb-C (power input), can't we connect to the regular usb-A ports, hopefully someone on the forum will clarify on it.

 

[dericchan1]

This is great!! will experiment with it today, my only question why ipad has to be connected to rpi4 usb-C (power input), can't we connect to the regular usb-A ports, hopefully someone on the forum will clarify on it.

Because only the usb-c port of the rpi4 is otg. You need that for the rpi4 to be used as an input NAA that will support sample rate switch. Essentially you can use the rpi4, same setup for Spotify, Amazon hd, deezer……

 

[borkenarrou]

Because only the usb-c port of the rpi4 is otg. You need that for the rpi4 to be used as an input NAA that will support sample rate switch. Essentially you can use the rpi4, same setup for Spotify, Amazon hd, deezer……

Ah, thanks for explaining it

 

[Rayon]

I've found PCM with 24 bits and Gauss dither to be remarkably punchy with the Sinc-Mga filter. I've always wondered why NOS sounded a lot better than polynomial-1 at 32x rates with LNS15 but it comes much closer with the 24-bit Gauss settings.

Btw if you like punchiness, try 4fs with different filters. I've found 4fs @ 24bit with gauss1 dither to be the best punching setting while still keeping sound smooth. At these rates I mostly use PGGB, and with that there is a clear step down in punchiness when I move from 4fs -> 8fs. To me it sounds like the further we go in sample rate, the more bits we need. I've noticed the same thing with Mojo 2. 24 sounds very "tall" at 1fs and 2fs, but starts to lose it's height quickly. 8fs already really benefits from 32bit.

 

[plumpudding2]

Btw if you like punchiness, try 4fs with different filters. I've found 4fs @ 24bit with gauss1 dither to be the best punching setting while still keeping sound smooth. At these rates I mostly use PGGB, and with that there is a clear step down in punchiness when I move from 4fs -> 8fs. To me it sounds like the further we go in sample rate, the more bits we need. I've noticed the same thing with Mojo 2. 24 sounds very "tall" at 1fs and 2fs, but starts to lose it's height quickly. 8fs already really benefits from 32bit.

Punch is the main reason for me to listen to PCM over DSD so I'll definitely give this a try!

I've also concluded that using LNS15 20 bits moves the PCM sound more towards DSD, while Gauss 24 bits moves it more towards NOS and makes PCM more a flavor choice instead of simply inferior DSD.

Also your plots of the 16-bit sine wave having only 3 levels is due to it being undithered. I'd be interested if the dithered version looks like a proper sine or whether you get something wonky looking that averages to being a proper sine.

 

[kopczas]

Is it possible to get to DSD 256 (PCM to DSD) on Ryzen 5600g with 16 gigs of RAM?

 

[Gibs]

It’s too late here, really must call it, but from recollection the 7EC-Super modulator and for filters gauss-sinc-long (or Sinc mx) as a balanced baseline to begin with (Jussi’s default for a reason).. then try XLA (or the longer non gauss sincs like Sinc-mga or Sinc-L) for a more expansive if less dynamic presentation. So many options that others can advise on better than me. Have a play.

Oh yeah make sure for PCM you have bits set to 15 or 16, have a listen to see what sounds better.

Hello there,
I'm bringing an ooold post of yours to the surface here, because i was hoping you could answer a question about my r26.
So, i've bought hi res 24bit music, and i'm wondering if converting 24bit to 16bit would cause loss? Because apparently the r26 can't process further than it's dynamic range of 115db right?
I'm not knowlegeable about this subject at all! May i get some help please?
Also, in general, the difference between 16bit, 24 and 32bit is just dynamic range right? So am i just better off buying 16bit and letting hqplayer do the processing?

 

[Rayon]

Also your plots of the 16-bit sine wave having only 3 levels is due to it being undithered. I'd be interested if the dithered version looks like a proper sine or whether you get something wonky looking that averages to being a proper sine.

These would likely give some indication:

Assuming that NOS dac plays those as is.

It can't make the dac play levels that are not available in the resistor ladder, but instead it will re-distribute the quantization noise.

 

[plumpudding2]

These would likely give some indication:


Assuming that NOS dac plays those as is.

It can't make the dac play levels that are not available in the resistor ladder, but instead it will re-distribute the quantization noise.

I can imagine the human ear prefers the top signal to the bottom one, human hearing is very impulse/waveform attack based and I would imagine all these little extra dips and peaks from the second signal only confuse the ear, despite it looking much cleaner in the frequency domain.

Really hope proper scientific research may be done on these things, I learned a lot about human hearing from Jussi who did a lot of his own research and testing, but it would be good for the hobby if these became widely accepted, peer reviewed conclusions (or disproven of course!)

 

[Rayon]

I can imagine the human ear prefers the top signal to the bottom one, human hearing is very impulse/waveform attack based and I would imagine all these little extra dips and peaks from the second signal only confuse the ear, despite it looking much cleaner in the frequency domain.

Really hope proper scientific research may be done on these things, I learned a lot about human hearing from Jussi who did a lot of his own research and testing, but it would be good for the hobby if these became widely accepted, peer reviewed conclusions (or disproven of course!)

The undithered signal sounds quite bad when volume is very close to the minimum volume the dac can play. I at least prefer dithering over truncation based on some tests I've heard. However, what 24bit achieves over 20bit is that it pushes the dithering downwards into quieter bits. This not only enables us to get some information from those lower bits as well, but it also enables those linear bits to produce us music instead of dither noise.