Solved Thoughts/Recommendations for setting up OSS for high quality audio processing

shkhln said:
Latency is completely irrelevant to "quality" and the defaults are fine unless you are playing music professionally. Listening should not require any adjustment unless you hear stuttering (in which case the value should be increased).
Click to expand...
In my case they are fine, and I do not have any problems, I just wanted to know how both options are connected to each other, if they are connected at all.
 
i think it is more sutied for the usb audio cards. i do not know what latency profile do. i never changed it. i set latency to "0" and have musicpd running in realtime. Whether it gives any advantages I have no idea. it just works. in linux setting up realtime is not a straightforward task. or it was such in the past.
 
Just about quality. i was in a dedicated shop and had a brief sound check of Audeze LCD2-Classic ( because it was the model i wanted to buy some years ago when they had discount ), powered by TEAC NT 505. What I want to tell is that I absolutely had no wow effect. Sound is veiled and spacious and it wasn't quality i expected compared to FreeBSD in bitperfect mode and in-ear headphones that 100 times cheaper. Of course you can't compare this 2 completely different headphones. So all in all, other than quality, and tech characteristics and measurement, there is a lot of subjective perception of a sound. Other thing is that i have my hearing slightly impaired due to reason I do not want to disclose, so it can be this too. Blind tests are needed. Of course, if I had an opportunity and affordability, I would choose higher end models just because of build quality and respect to labor required to produce them.
 
I noticed 2 days ago that this sysctl option hw.snd.vpc_0db changes the sound spacing even more.
The standard value is set to 45, but setting it higher or lower is just a user preference.
If you set it higher, you will get a more balanced sound, where the instruments are more seperated in the soundstage, and the instruments in the background get more audible.
If you set it lower, then the instruments get more tight together, and the instruments in the background will be harder to hear.
That is at least my perception while playing with this setting, and hearing music.
I am still unsure whether I should set it to 47 or 49, but these values are my recommended options.
 
I am not sure I completely understand what is exactly hw.snd.vpc_0db
however some years back it was discussed with MPD developer:
github.com

The 'next' command resets volume even after a setvol to the contrary · Issue #1055 · MusicPlayerDaemon/MPD

Bug report The 'next' command sets the volume to 45. This number is not configured anywhere, nor can I seem to alter the behavior of 'next' with respect to volume. Describe the bug So if I use 'set...
github.com github.com
So i have in my /etc/sysctl.conf
hw.snd.vpc_autoreset=0
hw.snd.vpc_0db=100
 
hw.snd.vpc_autoreset
When a channel is closed the channel volume will be reset to 0db.
This means that any changes to the volume will be lost. Enabling
this will preserve the volume, at the cost of possible confusion
when applications tries to re-open the same device.
Click to expand...
I think if you have the option to rise or lower the db value for a channel of a specific device inside an application, and set this option to 1, the db value set inside that application will override the global value of hw.snd.vpc_0db for the channel of that specific device.
If you then open the same device with another application, you can eventually get strange behaviour, because the other application now also gets the overriden hw.snd.vpc_0db value.
If you want to have different hw.snd.vpc_0db values for each audio application, you probably should enable this option.
This is my understanding of what this parameter does.

hw.snd.vpc_0db
Default value for sound volume. Increase to give more room for
attenuation control. Decrease for more amplification, with the
possible cost of sound clipping.
Click to expand...
As far as I can tell it is the volume for the foreground stage, middle stage, and background stage where the instruments play, it is not the same as the audio volume you can set through the mixer.
Clipping in that context eventually means that a low value will completely blend out some instruments.
A to high value eventually degrades the transition between one instrument to another, and gives a very rough unsmooth feeling to the whole sound stage.
hw.snd.vpc_0db=100 is definitely to high, alteast for me, because the sound stage gets very rough.
I would recommend you to try some values between 40 and 50.
 
You must log in or register to reply here.
Back
Top