Comments on: component updates

By: Acropolis

Acropolis — Sun, 26 Aug 2007 17:30:10 +0000

I’ll improve it later, because it is hard coded the value to treat the data as silence.

By: squeller

squeller — Sun, 26 Aug 2007 14:56:08 +0000

Hmm, I seem to have problems with the gapless part. Middle gapless is always off, and “remove beginning silence” does not matter: Quite often I’ve seen the component eating a bit of the beginning of a track. I.e. not only silence.
I try to find audio snippts and will post them to HA, but not today.

By: squeller

squeller — Tue, 21 Aug 2007 15:07:31 +0000

OK, it seems I need at least 1300. If 1000 or below, I earn a serious gap.

By: Acropolis

Acropolis — Tue, 21 Aug 2007 13:08:44 +0000

yes, because the samples in the output buffer are being force played, it will hold foobar until it is done.
so the output buffer must be less than 1000ms to make it not obvious enough, and I found 1000 is the most suitable value.

By: Squeller

Squeller — Tue, 21 Aug 2007 13:00:12 +0000

Thx. for the hint “dsp after crossfader”…

Output device buffer: 2000ms, local aac files, no other dsp… A quick computer at work. Ah OK, it depends on the output buffer, because this is being played fully?

By: Acropolis

Acropolis — Tue, 21 Aug 2007 12:11:10 +0000

hm…

1. it is easier to implement in %, as I said, I’m lazy~.
also, if I implement it in ms, it will need to update the sliders for every changes to the max length slider, quite serious flickers to the window.

2. it works well here, did you use any other DSPs?
I never see the 2-3 seconds problem. and what files are you playing? because the manual fading really depends on the reading/decoding speed. just a note, did u forget the set the output device buffer to 1000ms?

btw, if you have some particular DSPs after crossfader, when doing manual fading, it may have glitch happen, I’ve been trying a lot of methods but still don’t what causes it and how to fix yet.

3. because it is just something next to me when writing the crossfading routine, and it’s useful if not using DS as the output.

By: Squeller

Squeller — Tue, 21 Aug 2007 11:39:18 +0000

Many thanks, yes, I guess people want a shorter fade on manual track change… Three comments:

1. I’d prefer to have milliseconds for all settings instead percent values
2. Short manual fades don’t work here. The values:

Max length: 5000
Track change: 100%
Manual: 10%
Seek: 5%

When double clicking a playlist entry while playing, it always takes 2 or 3 seconds before something happens, then the new track fades in…

3. Why do you have “fade on seek” implemented? It’s in the fb2k core…

Greetings to Hongkong.