spatial layer bandwidth distribution and screen change detect using libvpx

[Fuji Nafiul]

I have 2 questions :

1) I capture screen data and after converting from rgb to yuv i use vp9 encoder to encode. Now is the any way that I can detect static image using libvpx ?

Meaning libvpx will return true/false whether the input image data is the same as previous or not ?

2) I enabled 3 spatial layer and 1 temporal layer for my screen sharing scenario using vp9 and svc. In my setup my upper layers are fully dependable on all subsequent lower layers to become fully decodable.

If I shift to only 1 spatial layer will it reduce my bandwidth and encode cpu time significantly anyway ?

Thanks in advance for any kind of help..!

[Marco Paniconi]

Hi,

On Tuesday, November 14, 2023 at 9:22:34 PM UTC-8 Fuji Nafiul wrote:

I have 2 questions :

1) I capture screen data and after converting from rgb to yuv i use vp9 encoder to encode. Now is the any way that I can detect static image using libvpx ?

Meaning libvpx will return true/false whether the input image data is the same as previous or not ?

No libvpx does not return such a flag to the user. 

 

2) I enabled 3 spatial layer and 1 temporal layer for my screen sharing scenario using vp9 and svc. In my setup my upper layers are fully dependable on all subsequent lower layers to become fully decodable.

If I shift to only 1 spatial layer will it reduce my bandwidth and encode cpu time significantly anyway ?

Yeah shifting to 1 spatial layer for screen would reduce you encode_time and save bandwidth (over 3 spatial layers), without much quality loss.