[LAD] announcing envy24control, mudita (*) edition.

[Tim E. Real]

On July 26, 2010 09:07:35 pm you wrote:
> On Mon, Jul 26, 2010 at 12:29 AM, Tim E. Real <termtech at rogers.com> wrote:
> >  trust me it will look great at the expense of only 28 extra pixels for
> >  the minimum height over the original - it's all you can really do, man:
> >
> > Remove the -72 and -108 dB marks on all the monitor sliders
> >  in volume.c: draw_24bit_attenuator_scale_markings().
> >
> > Remove the -42 and -54 dB marks from the DAC sliders in
> >  volume.c: draw_dac_scale_markings().
> >
> > Remove the -18, -30, -42, -54, -60 dB markings from the ADC
> >  sliders in volume.c:draw_adc_scale_markings().
> >
> > Change the level meter heights by changing
> >  gtk_widget_set_usize(drawing, 24, (60 * tall_equal_mixer_ht + 204));
> >  to
> >  gtk_widget_set_usize(drawing, 24, (60 * tall_equal_mixer_ht + 168));
> >
> >  in envy24control.c: create_mixer_frame().
> >
> > Trust me they'll love it !
>
> Thanks! Yes, it does look better. I've applied these suggestions to my
> current sources and they'll be part of an upcoming release that
> addresses the issues...
>
> If you have any suggestions on getting rid of the annoying
> detent/hysteresis as you move the sliders past the markings, that
> would be great.
>
> -- Niels.
> http://nielsmayer.com

Sure, NP I'll take a look.

Good news! 
As I mentioned in a follow-up to another thread, I found a bunch 
 more dB related functions.

I tested replacing the code in volume.c:dac_volume_to_db() and
 adc_volume_to_db() with a call to ALSA's snd_ctl_convert_to_dB().

The results are identical! And better for ADC compatibility, right?

Now comes the fun part. Coming up with an Al-Gore-rithm to
 handle the scale markings using these ALSA functions - 
 how many 6dB steps we have, which ones to display, the range, etc.

It occurred to me 'dynamic markings': If the user expands 
 more, or less, vertically, then more, or less, markings should appear.
But it's still tricky - we want it to look the way I suggested above,
 with more markings concentrated around the 0dB point, whenever
 we are forced to compromise by removing marks.
Ultimately, at minimum height it must look like that - it's a compromise,
 until a (better) solution comes up.

Also I see there is something called TLV based dB conversion where 
 I think it's a table or can be an arbitrary function. Must study more.

------------
For now here's the code, throw it inside 
 dac_volume_to_db() and 
 adc_volume_to_db() but *replace* DAC_VOLUME_NAME with ADC_VOLUME_NAME,
 and I believe mixer_volume_to_db() and wherever else required. 
Do you need a patch?

Hope this is OK, not sure how many members to set in struct snd_ctl_elem_id.
List members, does it look OK?

char* dac_volume_to_db(int ival) {
 if (ival != 0) {
 
  // Use ALSA functions to get dB values.
  long db_gain = 0;               
  float fval;
  snd_ctl_elem_id_t *elem_id; 
  snd_ctl_elem_id_alloca(&elem_id); 
  snd_ctl_elem_id_clear(elem_id);
  snd_ctl_elem_id_set_interface(elem_id, SND_CTL_ELEM_IFACE_MIXER);
  snd_ctl_elem_id_set_name(elem_id, DAC_VOLUME_NAME);
  snd_ctl_convert_to_dB(ctl, elem_id, ival, &db_gain); 
  fval = ((float)db_gain / 100.0);
  sprintf(temp_label, "%+2.1f", fval);
  return (temp_label) ;

 }
 else
 {
    // Rest of function ('off' part). ]
   ...

Tim.