Re: [fluid-dev] Patch for fast midi file rendering

[Pedro Lopez-Cabanillas]

Hi,

I've not yet said so in appreciation of all the work you have done, and it 
must be shout: Thank you very much! :-)

About the patch: I want to commit it after the 1.0.9 release, because it has 
several ramifications that need to be developed further, and it has enough 
importance  to deserve a complete cycle of testing, during several weeks or 
so. It may be ready to be released in FS 1.1.0

Things that I would change before the commit (if you don't beat me)

* Setting "player.timing-source". Bad name, and bad place. The same setting 
would be used from the player and from the sequencer. But for me, it belongs 
to the synth and the type is boolean. Either you trigger the timer callback 
from the synth or you don't. That was my first proposal: synth.slave-timer = 
yes/no.

* Define a new "slave-timer" member of _fluid_synth_t, and initialize it 
according to the corresponding setting in new_fluid_synth(). 

* fluid_sample_timer_process() is called unconditionally from 
fluid_synth_one_block(). This would impact the usage of FS when there are no 
timers needed (for RT-only clients) in a very sensitive place. Here you need 
to use the setting value. An "if" is much less expensive than a "call".

* fluid_renderer* is another bad name. It is not rendering anything, in the 
sense of translation, like the verb used in "rendering SVG into bitmap 
graphics". The proposed functions are only audio file helpers, because you 
don't include the MIDI player into the scope. 

* new_fluid_file_renderer() has a parameter "period_size", that is available 
as a setting. Instead, I would prefer to read synth->setting inside this 
function. You may have access here to "fluid_synth.h" :-)

* The raw audio format is not very useful by itself, but can be converted to 
anything else, for instance to mp3 with lame. It would be nice to send the 
output to stdout if the file name is not specified, so the output of FS can  
be piped to lame.

* "fluid_seq.c" should be converted to this timer, as discussed recently.

I've only started testing the results, sorry. I have very little free time. 
But the preliminary results look very promising.

Regards,
Pedro