all audio changes to float32 /s16l and ugly_resampler ???

[netrage]

I have tried to play various formats in VLC , DTS, Dolby etc. but interestingly the output always changes to float32 stereo or so it seems from debug.
I was a software developer long time back, hence I do understand a bit but not everything, can someone tell me if some configuration can be changed
to get "original" output like dobly , dts etc. without "ugly_resampler", "f32l->s16l" .. what am I missing here ?


------
dts info: DTS channels:5 samplerate:44100 bitrate:1411200
main debug: recycling audio output
main debug: looking for audio output module: 2 candidates
aout_directx debug: Opening DirectSound Audio Output
aout_directx debug: found device: Primary Sound Driver
aout_directx debug: found device: Speakers (Realtek High Definition Audio)
aout_directx debug: found device: Realtek Digital Output (Realtek High Definition Audio)

------

I have tried aout_directx , wave etc. but the output is always
-----
main debug: output 's16l' 44100 Hz 3F2R frame=1 samples/10 bytes
main debug: mixer 'f32l' 44100 Hz 3F2R frame=1 samples/20 bytes

------

ugly_resampler is always triggered, filters are always f32l, and in the end the quality I get is just downplayed ?

main debug: Stream buffering done (301 ms in 7 ms)
aout_directx debug: device supports 5 channels
aout_directx debug: device supports 4 channels
aout_directx debug: device supports 2 channels
aout_directx debug: device supports 1 channel
aout_directx debug: SpeakerConfig is forced to 5.1
aout_directx debug: couldn't use hardware sound buffer
aout_directx debug: channel reordering needed
aout_directx debug: creating DirectSoundThread
aout_directx debug: DirectSoundThread ready
main debug: using audio output module "aout_directx"
main debug: TIMER module_need() : 93.945 ms - Total 93.945 ms / 1 intvls (Avg 93.945 ms)
main debug: output 's16l' 44100 Hz 3F2R frame=1 samples/10 bytes
main debug: mixer 'f32l' 44100 Hz 3F2R frame=1 samples/20 bytes
main debug: filter(s) 'f32l'->'s16l' 44100 Hz->44100 Hz 3F2R->3F2R
main debug: looking for audio filter module: 12 candidates
audio_format debug: f32l->s16l, bits per sample: 32->16
main debug: using audio filter module "audio_format"
main debug: TIMER module_need() : 0.200 ms - Total 0.200 ms / 1 intvls (Avg 0.200 ms)
main debug: conversion pipeline completed
main debug: looking for audio mixer module: 2 candidates
main debug: using audio mixer module "float32_mixer"
main debug: TIMER module_need() : 0.117 ms - Total 0.117 ms / 1 intvls (Avg 0.117 ms)
main debug: input 'dts ' 44100 Hz 3F2R frame=1024 samples/4096 bytes
main debug: filter(s) 'dts '->'f32l' 44100 Hz->44100 Hz 3F2R->3F2R
main debug: looking for audio filter module: 12 candidates
main debug: using audio filter module "dtstofloat32"
main debug: TIMER module_need() : 0.193 ms - Total 0.193 ms / 1 intvls (Avg 0.193 ms)
main debug: conversion pipeline completed
main debug: filter(s) 'f32l'->'f32l' 48510 Hz->44100 Hz 3F2R->3F2R
main debug: looking for audio filter module: 12 candidates
main debug: using audio filter module "ugly_resampler"
main debug: TIMER module_need() : 0.121 ms - Total 0.121 ms / 1 intvls (Avg 0.121 ms)
main debug: conversion pipeline completed

-----------------

In short the net result of any type of output is always everything_to_float32 ?
I m willing to try out any changes if people can recommend things ?

Thx in advance.

[Rémi Denis-Courmont]

Unless you use S/PDIF pass-through, then audio is converted to float32 to apply audio filters.

[netrage]

Hello Remi,

Thank you for your reply, indeed that is my point, how can I disable this conversion ? The point is if
all formats DTS , Dolby etc. get scaled down.. suppose I dont want to apply any filters at all and play
"Pure audio" as source then what is the option ?

Rgds

[Rémi Denis-Courmont]

You need to change the source code.

[netrage]

Considering that changing the source code is not an option, hence , I will leave it here, do you know any other player that can play "pure audio" without conversion, though my guess is not.

Thank you anyway for clarifying this fact. Appreciate your time and assistance.

[Rémi Denis-Courmont]

No I don't know. And anyway, most contemporary sound cards can only render S16N, FL32 or both. Furthermore, FL32 will not incur any loss of quality/precision unless the input is 64-bits, which is unlikely.