Original music

[dethrace-labs]

• Adds support for music as in the GOG release. That is, MUSIC/Track0[1-8].ogg files. If they exist, the original CD audio functions are redirected to play the ogg files.
• Adds stb_vorbis to enable vorbis support in miniaudio
• Switches miniaudio to use the default header-only implementation (to make it work with vorbis)
• Fixes bug in sound options where the dials are both initially set to the music volume

Fixes #278

[madebr]

dethrace's IRandomBetween(0, 7) returns variable in the range [0, 7] (inclusive 0 and 7).
However, gRandom_CDA_tunes has length 7, with maximum index 6.

The MUSIC folder of my gog copy has 8 tracks (including a 4s silence track) so I think changing gRandom_CDA_tunes to int[8] makes sense.
In the Splat Pack, there are 3 tracks and 5 dummy (4s of silence) ones.

[madebr]

The 7th item on my system is currently 0, which causes a stack overflow in S3GetDescriptorByID(0) -> S3GetDescriptorByID(0) -> S3GetDescriptorByID(0) -> ...

[dethrace-labs]

Appreciate your second pair of eyes on this one! I originally made the change to [8], based on there being 8 tracks in SOUND.TXT, and seeing 9607 in IDA at the end of the array. When I saw the last track was silence, I figured it was deliberately left off and switched back to [7].

I didn't think about the IRandomBetween(0, 7) though! :)

I wonder if the original code was declared as 7 but initialized with 8 items like int gRandom_CDA_tunes[7] = { 9600, 9601, 9602, 9603, 9604, 9605, 9606, 9607 }; maybe

[madebr]

I'm not sure assert is the right approach in S3GetDescriptorByID because it is removed when building a Release build.
But LGTM!

[dethrace-labs]

I'm not sure assert is the right approach in S3GetDescriptorByID because it is removed when building a Release build. But LGTM!

That was my sort of my intention. The check is not present in the original code, but the assert could be helpful if we are working on the audio code in the future and break something