This has become a useless no-op on all platforms.

Otherwise use decoder_out which is the currently used decoder output.

The decoder device is created with a NULL window. This can work with some
decoder devices but in some case none will be created and we just get NULL.

We may have to release resources associated to the decoder, like the decoder
device.

We already need to clean some ES formats anyway.

It will be passed to the first filter when it's created.

It is only set when calling filter_chain_Reset() and no filter has been added yet.

We should not be using the filter chain unless filter_chain_Reset() has been called.

Typically used when a CPU to GPU converter is used. It cannot deduce the decoder
device or the video context from the input. Since we are pushing the video
context we don't want to get it from the output/environment either.

When the callback is not provided, a NULL decoder device is returned.

The chain can only have a different output than requestedoutput if a converter
was appended after the filter_chain_Reset() call.

Either because Reset() set the same input value or the we're using the same
value set by the previous AppendXXX().

It's already done in the decoder helper.

If new callbacks are added we want to know (warning or error) if some owners
don't handle the new callback.

Full rework of the SDI buffering to handle stream
interleaving, truncation, decoder output pacing,
threaded frame conversion & filtering.
Now also does proper preroll.

Remaining frame drops are due to munmap()

double captions bug

The core can do it by default.

Decklink vout backport, so
this is the currently the only support