A few potential tiny issues

[BleuBleu]

Hi!

Im writing my own little SF synth (much simpler) than yours and I am comparing my output to Polyphone. I think I have found a couple of tiny issues, but I may be super wrong. Apologies in advance if I am.

1. Im fairly sure the little recorder tool swaps the L/R channels. The ordering in a stereo wave file is L/R/L/R.
   

   polyphone/sources/sound_engine/elements/recorder.cpp

   Line 147 in 7c364c9

   _dataWav[2 * i + 1] = dataL[i];

2. Im not 100% confident that Polyphone loads the sm24 data from the correct offset in the file. I was testing with this file: https://www.musical-artifacts.com/artifacts/5536 and according to my calculations, the sm24 raw data should start at offset 72577618. Polyphone seems to be off by a couple hundred bytes? I may be wrong but I think this calculation fails to take into account previous headers and stuff.
   

   polyphone/sources/core/input/sf/inputparsersf2.cpp

   Line 181 in 7c364c9

   value.dwValue = SHDR._start.value + sdtaPart._startSm24Offset;

3. Certain envelope calculations are done twice. With large buffers this is not a big deal, but if you run the envelopes 1-sample at a time (which never happens in reality) you will see that the decay and release of the modulation envelope applies the slope twice. Meaning a 2sec decay will take 1sec. Again, this doesnt really matter with 512-bytes audio buffers, the error will be tiny, but it is still a small error.
   

   polyphone/sources/sound_engine/elements/enveloppevol.cpp

   Line 264 in 7c364c9

   lastValue = _precValue + coef;

Thanks so much for developing this tool. It is the benchmark every other soundfont tool should be compared against.

-Mat