Separate coded picture detail from modelled grain
AOMedia’s AV1 tool description explains a model-based process that adds synthesised noise to a decoded, denoised video signal. That is different from transmitting every original grain sample as ordinary picture detail.
The implication is limited: the feature can preserve an intended grain appearance efficiently, but it does not prove a particular video uses it. Nor does grain reconstruction restore unknown lettering, faces or scene detail discarded elsewhere. A texture model and missing semantic image information are different things.
Check the actual reference and processing path
For a developer comparison, retain the exact file or authorised reference, decoder version, reported codec and any available grain-feature information. Keep the player’s post-processing and television noise-reduction settings in the record.
Compare the same passage at the same size and playback state. A paused frame and moving playback ask different perceptual questions. A photograph from another screen also changes the evidence, so avoid declaring one decoder broken merely because two loosely matched images have different visible texture.
Do not equate AV1 support with every separate grain system
AOMedia also publishes AFGS1 as a standalone specification. Its description says supporting devices generate grain according to transmitted model parameters and add it to decoded video, with signalling usable alongside other modern codecs.
That relationship does not mean every file with a grainy appearance is AV1, or that an AV1 label certifies support for every separate metadata/container implementation. Preserve the particular feature and signalling evidence rather than using film grain synthesis as an interchangeable codec name.
Describe defects more precisely than grain is bad
If investigating an actual presentation issue, note whether the observation is intended fine texture, coarse blocks, banding, frozen frames or a repeatable failure at one time. Those descriptions guide different investigations. Keep the original reference unchanged while testing any proposed processing adjustment.
No decoder or display was tested here, and no title is claimed to use a particular grain model. The article explains the primary format concept and a controlled comparison task. It does not recommend removing a filmmaker’s chosen texture or promise that smoothing can improve every source.
Sources and scope
Scope notes distinguish published documentation, supplied information and editorial methods. They do not report a hands-on service test.
- AOMedia: AV1 and libaom tool description — Primary2021 technical document and indexed film-grain section: modelled noise added to decoded denoised signal; explanatory scope, not tested title.
- AOMedia: standalone AFGS1 — Primary full body: standalone model parameters/signalling and synthesis on supporting devices, not universal AV1-container equivalence.
