What about sound formats?

[daniel-birket]

@lawrenceper , the SDL2 audio mixer library (which pygame also uses) is a multi-channel stereo player capable of many sound formats. It can play this formats:

• FLAC
• MOD
• MP3
• OGG
• MIDI
• OPUS
• WAV
• AIFF
• VOC

I read that MIDI synthesis is CPU intensive relative to a simpler format like WAV. Obviously MIDI is a lot smaller, but you would be at the mercy of the playback library for sound quality compare to what you could tweak with audio tools. I don't know much about audio formats beyond WAV is simple but bulky and MP3 is compressed. Size doesn't seem to be a huge factor these days but I still like efficiency. Note that the library seems able to handle multiple formats at once.

It also has one long-playing 'stream' channel intended for long audio, like music, not short samples. It is also able to loop samples, so if you have 5 seconds of swamp frogs croaking, you can loop them as needed. There is also volume control, including fade-in-fade-out. I imagine sequences like:

• fade-up scene ambiance (forest sounds and stream), looping
• play scene description "You are standing at the end of a road before a small brick building. Around you is a forest. A small stream flows out of the building and down a gully."
• player types "go forest"
• play action audio (footsteps on leaves), limited loop
• fade-out forest with stream and halt and fade-in forest and light wind, looping
• play scene description "You are in open forest, with a deep valley to one side."

Scene descriptions might be synthesized from text.

Anyway, I'd like your thoughts on what format would be good to use store channels of audio for playback during the game, including:

• scene ambience
• character speech
• scene description/ action narration
• object and action foley sounds

[lawrenceper]

Hi,

I think that WAV files would be better for audio loops than compressed audio like MP3. MP3s are great especially for speech because of the smaller size. I have messed around with MP3 files in Audacity, and there seems to be an audio gap when looping an MP3 file vs a WAV file.

To truly use open source tech, we could use OGG and FLAC file formats. OGG is a file format that is compressed (lossy) like MP3, and FLAC is compressed but is still lossless. Think of FLAC like a ZIP file with a WAV file inside it.

I could do some experimenting in Audacity with both file formats to see if there is a noticeable gap between the loop. To my ears, the OGG format did not sound that grate, but we could use a higher bitrate when exporting in Audacity. From what I remember, the FLAC encoder has options for compression as well.

A FLAC file might be better suited for ambient sound than OGG, especially if we're using 5 second clipps. The OGG format will work better with speech )E.G. acting/dialogue).

Hope this helps.