OrgMaker audio quality.

Hai guise,
I'm on a quest of making a music tool of sorts for Organya format, and was wondering if you guys can do some recordings of OrgMaker/CaveStory audio output for a certain tune, so I can verify something.
Tune in question.

I want to see how it sounds across different audio hardware and different sound settings. In particular, the detuned lead in this song.

This is Realtek ALC662 onboard audio with full acceleration and best samplerate conversion, in winXP.
And this is OrgPlayer, for reference of how it's suppose to sound with perfect tuning.

 

Here are some audio recordings I made:
rnhart.net/orgmaker/disco-zax-soundmax.zip (5.5 megabytes)
(I might delete this file after a week or so.)

I have also been unable to predict OrgMaker's unusual behavior when using the Freq values. It sounds like OrgPlayer was more predictable for you? I assumed OrgMaker and OrgPlayer used the same code base and behaved the same way. I haven't thoroughly tested OrgPlayer yet.

When I wrote my OrgMaker Notes, I attempted to test and predict the exact frequencies that would be produced by OrgMaker for any Freq value, but didn't find a predictable pattern. I couldn't tell if the strage behavior was caused by OrgMaker, caused by DirectSound, only happens with certain hardware acceleration settings, or only happens with certain hardware. I gave up and just documented what freqency values OrgMaker passes to the setFreqency function instead. I guess I should have also written that what OrgMaker passes to the setFreqency function might not match the audio you actually hear, but it was hard for me to tell if it was just happening to me.

It's been a while since I've investigated it, but your post has inspired me to start looking at it again. Here's a current hypothesis I'm testing: When a new tone begins playing, if there is an already playing tone whose sample freqency is close to the new tone's sample freqency, the already playing tone's sample freqency is changed to the new tone's sample frequency.

Also, I found this again:

What exactly does the sample rate conversion quality setting do? This is what I've found so far: Policy for Sample Rate Conversion of Audio Streams says "The slider shown in the Advanced Audio Properties dialog assigns the settings Good through Best to linear interpolation, multipoint interpolation, and high-end multipoint interpolation, respectively." Types of Sample Rate Conversion describes those interpolations slightly more, and also mentions Add/Drop (also known as truncation or nearest neighbor) is used if system constraints prevent a higher quailty conversion.

Does this mean when system resources are low, DirectSound will use a "Simple" (nearest neighbor) sample rate conversion, no matter what the setting is set to? There's so much about digital audio conversion I don't understand, but could this be contributing to OrgMaker's strange behavior?

 

Here are some audio recordings I made:
rnhart.net/orgmaker/disco-zax-soundmax.zip (5.5 megabytes)

No acceleration seems to produce best tuning(at the cost of many dropped notes). I did a few more recordings(OrgMaker maximised):
ESS1969 fullAcc bestConv
ESS1969 basicAcc bestConv
ESS1969 noAcc bestConv
ALC662 fullAcc bestConv
ALC662 basicAcc bestConv
ALC662 noAcc bestConv

It sounds like OrgPlayer was more predictable for you? I assumed OrgMaker and OrgPlayer used the same code base and behaved the same way. I haven't thoroughly tested OrgPlayer yet.

Derp, should have linked. OrgPlayer is different code altogether and does everything "in-house", with very predictable results.


I could either attempt to emulate some of this shittyness or maybe just include a 'your mileage may vary' warning in the tool .