Audio Pipelines

Everything is io.Reader and io.Writer

Diago follows GO std lib and providing interface for Reader/Writer when it comes reading and writing media. This made easier usage of RTP framework and optimizations, by providing end user standard library io.Reader io.Writer to pass his media.

In other words creating reader or writer pipelines allows to build interceptors, encoders, decoders without introducing overhead of contention and buffer reuse.

media and audio package also provide all this helpers for making audio pipelines straightforward. In realtime media streaming, audio is sampled at constant rate.
Some of io operations you encounter have helpers media package like media.Copy , media.ReadAll, which are better suited instead std io

Creating Echo

Everything mostly begins by getting actual AudioReader and AudioWriter from Dialog Session.

ar, _ := dialog.AudioReader()
aw, _ := dialog.AudioWriter()

Now you can use this helpers

n, err := media.Copy(ar, aw) // It creates buffer of media.RTPBufSize 

For less memory alloc, better option is media.CopyWithBuf where controlling buffer reuse can be made.

More audio processing

For more audio processing audio (ex transcoding, wav streaming) package provides you with prebuilt helpers for pipelining More you can find here https://pkg.go.dev/github.com/emiago/diago/audio