[Achel]

Hi, I search an explanation about smthg I really don't Understand ... First of all here is my configuration :
 
Computer under Linux Unbuntu using ALSA (for bitperfect) => DAC HRT Music Streaming V1 16/24bits on USB => Audacious using the Alsa Drivers for bitperfect with any plugin activated (even crystaliser) => Little Dot I+ with Burson V5i OpA and Voskhod Gold plate tube from 1963 => Grado Sr80 / Sennheiser Momentum 1 Over Ear.
 
- When I configure Audacious on 24bits, I heard differences between mp3 320kbps 44.1khz and 48khz ... and prefer 48khz, the sound is more sharp. On 44.1khz songs the sound seems damped and smthg stranger : right channel seems more damped than the left one on the two headphones ... 
 
- When I configure Audacious on 16bits, there is no differences between mp3 320kbps 44.1khz and 48khz and the sound is more realistic on both of them ... more soundstage, better spatialisation, sharper, precise ...
 
Is it my ears (35years) ? Is it the DAC wich saturate on 24bits ? Or is it the Little Dot amp ? Does mp3 codec at 320kbps 44.1khz is degraded in 24bits or downsampled/heavily upsampled ? I thought putting the player in 24bits I'll have more accurate sound on hight rate mp3 ... why does it sound less accurate and sometime damped ? That's the same on FLAC soundtrack ... 
 
Thks for all wich gives explanations, I really want to understand this phenomenon (technical explanation are really welcome even on the os tweaking).
 
Many thks

 

[Currawong]

MP3s are 16 bit, 44.1kHz, so any other output settings are not "bit perfect". If you set the computer to output anything other than that when playing them back, then the computer is re-sampling (from 44.1 kHz) or dithering (from 16 bit). That "sharpness" you hear is either a: distortion from bad re-sampling, or b: a change in filter in the DAC (less likely, but not impossible). That's what I reckon anyway. 
 
To be sure, you'd have to check the docs for Audacious, or see if there is any information how they handle audio data if the output is different from the file.

 

[Achel]

IMHO I think that audacious resample/upsample (?) when I ask to put it in 24bits ... I was thinking that with the use of Alsa drivers choice and choice of the output hardware without any modification the player choose the spec of the files BUT it seems that it doesn't adapt this parameter on the fly (24 bits is not "the maximum" of wath the DAC can support but the stack value of the player) ...
 
I'have tried Deadbeef player wich does it perfectly and adapt on the fly ... same sound as in 16bits on audacious. 

 

[RRod]

To verify what format alsa is sending out to the DAC, you can use an incantation like:
cat /proc/asound/card0/pcm0c/sub0/hw_params
 
You can delve into /proc/asound for more information. The 'pcm' file will help you find the right card/stream, but there is probably already a symlink to the card name.

 

[Achel]

I have watched the sources, it seems that when you ask 24bits and the souce is in 16bits, it return to the intern mixer an try to upscale at the place of adapting to the source ... 

 

[RRod]

  I have watched the sources, it seems that when you ask 24bits and the souce is in 16bits, it return to the intern mixer an try to upscale at the place of adapting to the source ... 

 
Some DACs don't support 16-bit word lengths. My EMU-0404USB is like this; it only runs at 24-bit so ALSA pads on the extra eight 0s. I would still call this pretty close to bit-perfect, since a truncation back to 16-bits would recover the original data.

 

[Achel]

Thks RRod, my HRT is 16/24bits ready without problems, In fact, if I play a 24bits (with 24bits activated on Audacious) it plays perfectly, but if there is a 16bits song after, it don't resolve it in 16bits but try to upscale it in 24bits (even do I have selected the divers without any conversion) ... so I moved on Deadbeaf Player which switch the résolution for each track perfectly with ALSA drivers.  Hope audacious will modify their sources to be more flexible on this point cause I liste music on it for many years now and have taken it as an habit.