You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When the encoder is fed with discontinuous CTS (let's say, some incoming buffers are lost) it outputs wrong DTS.
I've prepared a PR that adds a simple test showing what's happening: #478 In this example, the ticks per seconds parameter is configured to "auto" (cfg.m_TicksPerSecond = -1) and the framerate to 30. The result is as follow:
So... we discussed it. As I understand, this is an issue when the client code has to skip frames for some reason. Correct me I am wrong, but this can really only heppen in live scenarios, right?
VVenC is an offline encoder and it seems an overkill to add all this logic to handle input discontinuities. We would probably need to add a scene cut or similar also, which adds a ton of logic.
We will eventually try and improve it, so at least the output timings are not incosnsitent, but as its outside of the scope of the main use case of VVenC, this is a lower priority for now.
When the encoder is fed with discontinuous CTS (let's say, some incoming buffers are lost) it outputs wrong DTS.
I've prepared a PR that adds a simple test showing what's happening: #478 In this example, the ticks per seconds parameter is configured to "auto" (
cfg.m_TicksPerSecond = -1
) and the framerate to 30. The result is as follow:Incoming CTS:
Outgoing DTS:
The text was updated successfully, but these errors were encountered: