Decode at the on-screen size: canvas CSS width times devicePixelRatio. Best
for fixed-size surfaces fed by a much larger source (frame sampler, inline
players). Falls back to native while the box is unmeasured.
The box it reads is the one bindCanvas measures, so this strategy resolves to
native under presentation "frames", where the engine holds no canvas and
nothing ever measures one. A frames-mode consumer passes displayBoxResolution
instead.
Decode at the on-screen size: canvas CSS width times devicePixelRatio. Best for fixed-size surfaces fed by a much larger source (frame sampler, inline players). Falls back to native while the box is unmeasured.
The box it reads is the one bindCanvas measures, so this strategy resolves to native under presentation "frames", where the engine holds no canvas and nothing ever measures one. A frames-mode consumer passes displayBoxResolution instead.