dynamic link creation model

[dragoncoder047]

How is this library modeled to create links between different nodes? There seem to be multiple different base classes for things (AudioStream, AudioOutput, AudioSource, StreamCopy, etc) and none inherit from any common superclass.

The Javascript web audio API, as well as PJRC's Teensy Audio library, designed everything as audio nodes that have input and output "ports", that are then connected together. A sine wave generator might have 1 output (mono) and 0 inputs, for example. I find this kind of organization easy to think about, but unfortunately I'm trying to do some audio work on an ESP32, using this library.

I'm not asking for any rewrite here. I'm just asking is there any base model/abstraction/idea about how audio is routed between nodes? Mostly because I want to dynamically re-route audio at runtime, and knowing when to pass in an input stream as a parameter to a constructor and then using that stream as an output, versus using a StreamCopy to pump data between two streams is a little confusing.

[pschatzmann]

The starting point was the Arduino class model! At the root you have Print and as subclass of Print you have Stream. Therefore AudioStream inherits from Stream and AudioOutput inherits from Print!
The common class of all is therefore the Arduino Print.

Arduino Print is for writing data and Stream supports writing and reading! I tried to be consistent in the classes, so when you can read and write it is a Stream and if it is output only it is ending with Output and inherits from Print and not Stream.

This has the big advantage that we can also use any Arduino subclasses which inherits from Print of Stream: in our framwork as well: e.g. File, Serial etc...

[pschatzmann]

I corrected the explanation for the AudioSource!

[pschatzmann]

StreamCopy is used in an Arduino loop() to push the data from a source to a destination. You could do reads or writes yourself as explained in the Introduction Wiki, but this saves you from some pitfalls and is much less code to write

[dragoncoder047]

I corrected the explanation for the AudioSource!

Well, then where do the codecs come in? You can't just play an MP3 as raw data, you'll just get noise.

Also I noticed according to the documentation, it seems odd that if you have a Stream and want to add an effect to it, you have to pipe it through a GeneratorFromStream because AudioEffects will only take a AudioGenerator as input. (Also, you've marked AudioEffects as obsolete. What's the replacement?)

[pschatzmann]

Just read the chapter about encoding and decoding
The Wiki is quite complete!

[dragoncoder047]

Thank you. I didn't know the wiki existed, and the doxygen documentation is a little cryptic.