29 May
2023
29 May
'23
8:45 p.m.
On Mon May 29 20:36:49 2023 +0000, Rémi Bernon wrote:
it's not clear if we will be using full encoding pipeline, or will
need to implement individual encoder MFTs It may be possible to squeeze a full encoding pipeline into the muxer component, but I think the decoding side experience has shown that it is not future proof and that we should have individual components, and compose the pipelines on the PE side.
To be clear I don't think that any positions on the decoding pipeline should inform those related to the encoding pipeline, and I don't even think it makes sense to use them as an argument.
None of the reasons to expose a generic decoding pipeline apply to encoding, I think.