Add Docs for AudioEncoder #717
Conversation
facebook-github-bot
added
the
CLA Signed
docs/source/conf.py
Outdated
There was a problem hiding this comment.
Changed in this file, as well as the files renaming, are meant to separate our "tutorials" page into 2 separate sections: one for decoding, one for encoding.
| @@ -16,8 +26,11 @@ def __init__(self, samples: Tensor, *, sample_rate: int): | |||
| raise ValueError( | |||
| f"Expected samples to be a Tensor, got {type(samples) = }." | |||
| ) | |||
| if samples.ndim == 1: | |||
| # make it 2D and assume 1 channel | |||
| samples = samples[None, :] | |||
There was a problem hiding this comment.
Drive-by, I think this makes sense, i.e. if the input tensor is 1D we assume it's 1 channel instead of raising an error
There was a problem hiding this comment.
Is that the most idomatic way to do that? There's also unsqueeze, view or reshape. Minor, but I'm surprised at how this works.
| uses the version of FFmpeg you already have installed. FFmpeg is a mature | ||
| library with broad coverage available on most systems. It is, however, not | ||
| easy to use. TorchCodec abstracts FFmpeg's complexity to ensure it is used | ||
| * Relying on [FFmpeg](https://www.ffmpeg.org/) to do the decoding / encoding. |
There was a problem hiding this comment.
Nit: I prefer "x and y" as opposed to "x / y" in prose.
| order = [ | ||
| "audio_encoding.py", | ||
| ] | ||
|
|
There was a problem hiding this comment.
Can we add a comment explaining that we have two top-level galleries, and for that reason, we need to figure out which gallery we're using (decoding versus encoding)? I was real confused until I concluded that must be what's going on.
|
Moving the examples into decoding and encoding subdirectories makes sense, but I'm curious: will that change the resulting tutorial URL? We already have some pointers to these tutorials floating around blogs and social media. |
|
|
||
| # %% | ||
| # We first instantiate an :class:`~torchcodec.encoders.AudioEncoder`. We pass it | ||
| # the samples to be encoded. The samples must a 2D tensors of shape |
There was a problem hiding this comment.
Nit: "The samples must be a 2D tensors of shape"
This PR adds docstrings and a tutorial for the
AudioEncoder, plus some minor changes listed in the comments below