jstrait / beats Goto Github PK

Hello how add wav not exactly one by one
for example I need put one in this same time this same after 1seconds

2 wave files playing in this same time but in different starting point

Howdy!

Nice library! I found one minor problem, files read and generated end with noise :-P

The reason of that is the direct usage of File.read and also File.open which doesn't respect binary content on Windows (either MRI or JRuby).

The fix for this is simple.

Replace File.read(filename) with File.open(filename, "rb") { |f| f.read }

Replace File.open(filename, "w") with File.open(filename, "wb")

Found 3 occurrences of that, so it doesn't worth the pull request ;-)

After that, beats start to beat music instead of noise!

Cheers and great work!

Just an idea: add support for accent using caps:

x...x...X...X...

I'd really like if I could pipe a YAML string to the beats executable, and have it treat that YAML as the INPUT_YAML_FILE if none is specified.

For example, in addition to being able to do this:

beats --path="./" tmp/word.beats.txt

It'd be great to be able to do something like this:

cat tmp/word.beats.txt | beats --path="./"

Hi, first off all, great Gem 👍

I'm having an issue with the beats output audio. It's seems that the audio from the kit is slowed down. Do you have a clue what could cause the issue?

You need to add wavefile as a dependency to the gemspec so that it is installed by adding:

s.add_dependency 'wavefile', '~> 0.3.0'

Excellent work on the library by the way!

Are there any plans to support rhythms that don't fit nicely into 16th-note patterns?

Right now I see that the audio engine is calculating the length of a sample by calculating the length of a 16th-note for the current tempo.

My thought was that (and I am willing to work on this & provide a patch) it could instead calculate the length for a quarter note, and then calculate the length of each sample depending on how the quarter note has been subdivided.

For example, we could accept syntax like this:

x...x...(xxx)x.x.

Which would correspond to:

16 rest rest rest 16th rest rest rest 1/4-triplet 1/4-triplet 1/4-triplet 16th rest 16th rest

If a quarter note section is wrapped in (), we calculate the length of each sample based on how many notes are within the parens.

From here, we could further subdivide the 1/4-note by nesting parens. Each nested paren would represent half of it's enclosing rhythm. So:

((xxx)(xxx))

would indicate 1/8-note triplets.

Just some ideas I've been kicking around.

The tests use StringIO so that the wave file is written in memory, instead of being saved to disk. It does this by swapping out the normal File object used by WaveFile::Writer with a StringIO instance.

However, the Writer constructor writes the wave file header to disk. Since replacing File with StringIO happens after the constructor is called, an unnecessary does_not_matter file gets written to disk, polluting the repo's home folder.