[nas] [PATCH] changed method of setting the input gain

[Jon Trulson]

On Tue, 25 Jul 2006, Erik Auerswald wrote:

> Hi,
>
> On Mon, Jul 24, 2006 at 06:53:44PM -0600, Jon Trulson wrote:
>> On Mon, 24 Jul 2006, Paul Fox wrote:
>>> i wrote:
>>>> erik wrote:
[...]
>>         But, aside from RECSRC and devmask, nas doesn't have to 'read'
>>         the mixer (for example, the output gain).  Why would it?
>
> Well, it _does_ already read them (and ignores them later on...). It
> would want to to read them to use these values on startup instead of
> it's own default values.

   Yes, in that case, we would use them :)

>
>>         If some external app has set output gain to '50' and a nas app
>>         sends data, it will use whatever the mixer is already setup at
>>         (well it will :).  Right now it will simply write the
>>         'default' gains in nasd.conf.
>
> This is kind of a complicated matter... The NAS server will write it's
> own gains (default or previously set) whenever opening a device or when
> there is an open device and some app changes the settings. An external
> app can alter the gain settings while NAS is playing without the NAS
> server even noticing this.
>

         True.

>>         Don't do that.  Again, your problem is not that NAS isn't
>>         'reading' the mixer gains, it's that NAS is writing the
>>         default gains to the device, without your express written
>>         approval.
>
> IMHO the NAS server should have some internal settings that are used
> whenever playing a sound. To achieve this these settings have to be
> written to the mixer whenever opening the device, because some other app
> using it might have changed the settings. Then it is possible to control
> NAS playback even with a released device and other (local) apps using
> the same audio device when NAS is not playing/recording.
>

         I agree... I think Paul does too... What we are trying to
         determine here is how the mixer shou;d be initted - 1. left
         alone 2. set to the last NAS setting (last*) or 3. set to the
         configured default (gain).

> The MixerInit option should define whether or not the NAS default values
> are used on NAS server startup.
>

         Well, we can stretch it to also mean 'How' the devices are
         setup too.

> A RestoreMixer option could be used to restore the previous values of
> the audio device whenever it is closed by NAS. I don't think this is
> needed, because the current mixer settings should be viewed as
> "undefined" by every sound app and therefore set if some specific values
> are needed (hence the default behaviour of NAS to set the default
> gains).
>
> A NoMixer option could be used to not change the mixer settings at all.
> This could be useful when an external amplifier should be used to
> control the volume. The internal amplifiers of most sound hardware are
> of low quality, the Gravis Ultrasound even had a special output port
> with fixed amplification because of this. I prefer to set my sound
> hardware to some working values (without distorted sound) and use
> external amplifiers to control the volume.
>
> The effect of this option can be had with setting the mixer device to
> something not a mixer (or not existing), but the NAS server still thinks
> it can change the settings, NAS mixer applications will report changed
> values but the mixer will not change.
>

         Yeah I thought of that too, but it seems hacky - tell someone
         they should specify some non-existant file for the mixer device...

> If I understand Paul correctly he wants to be able to control the NAS
> sound via NAS mixer applications but the NAS server to use the current
> mixer settings whenever opening the device (losing the changes made by
> NAS applications before). IMHO this should be a seperate Option (or a
> seperate value for MixerInit ("startup", "reopen", "never" instead of
> "yes", "no"). With these values "startup" is the same as "yes" and
> "reopen" the same as "no".
>

         Yes seperate string or integer options as to how the device is
         used.  We just need to define these options :)

> Erik
>

-- 
Jon Trulson
mailto:jon at radscan.com http://radscan.com/~jon
#include <std/disclaimer.h>
"No Kill I" -Horta