colinbdclark at gmail.com
Sun Mar 6 17:21:20 UTC 2016
Thanks for introducing yourself, and welcome to the Fluid community. I'm glad to hear that you're interested in the SoundFont Google Summer of Code project.
In terms of your questions, I think the key issue isn't so much that MIDI.js doesn't exploit SoundFont's richness, it's that they don't actually really support SoundFonts at all. MIDI.js implements an "extractor" written in Ruby (https://github.com/mudcube/MIDI.js/blob/master/generator/ruby/soundfont_builder.rb <https://github.com/mudcube/MIDI.js/tree/master/generator/nodejs>) that extracts WAV files for each MIDI note number by "playing" the SoundFont with a fixed 0.5-second duration. This has two fundamental drawbacks:
1. Because it packages a sound file for each note, it leaves a significantly greater amount of network and storage than a SoundFont (which was the whole point of SoundFonts in the first place)
2. It limits sounds to no more than 0.5 seconds in duration, which is fine for percussive sounds or short attack instruments such as a pianos or guitars, but it means that you can't usefully play back longer-sustaining sounds such as violins.
What we need is a genuine SoundFont parser and playback engine, which is capable of understanding the SF2 format and playing back its contents in an efficient and flexible way. Ultimately, we want to be able to use this engine in order to add SoundFont support to our various accessible sonification projects, providing an easy way to play back realistic instruments sounds.
Does this answer your question?
> On Mar 3, 2016, at 2:39 PM, Saksham Saxena <saksham.saxena.1994 at ieee.org> wrote:
> About the idea, I understand that MIDI.js et al don't always exploit SF2 richness, and how our un-opinionated sf2-parser will save the day with Flocking as synthesizer. However, I'd like to know more about the idea, and the previous experiences/issues with MIDI.js and/or any other library which allegedly isn't doing justice with .sf2 files.
> Correct me if I understood anything wrong. Hoping for an enriching discussion!
> Saksham Saxena
> Member, IEEE Delhi Section Student Network
> Member, IEEE DTU Student Branch
> Student, Delhi Technological University
> LinkedIn <https://www.linkedin.com/in/saxenasaksham> | GitHub <https://github.com/sakshamsaxena>_______________________________________________________
> fluid-talk mailing list - fluid-talk at lists.idrc.ocad.ca
> To unsubscribe, change settings or access archives,
> see http://lists.idrc.ocad.ca/mailman/listinfo/fluid-talk
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the fluid-talk