4:4:4 bitstreams with intra prediction now pass

[frankplow]

With ffvvc/FFmpeg#90, the following bitstreams now pass:

• STILL444_B_ERICSSON_1
• 10b444P12_A_Sony_2
• 10b444P12_B_Sony_2
• 10b444P12_C_Sony_2
• 10b444P12_D_Sony_2
• 10b444P12_E_Sony_2
• 10b444P16_A_Sony_2
• 10b444P16_B_Sony_2
• 10b444P16_C_Sony_2
• 10b444P16_D_Sony_2
• 10b444P16_E_Sony_2
• 12b444vvc1_A_Sony_2
• 12b444vvc1_B_Sony_2
• 12b444vvc1_C_Sony_2
• 12b444vvc1_D_Sony_2
• 12b444vvc1_E_Sony_2

Some take longer than 300s to decode on my machine and likely on the GitHub runners:

• 10b444P12_E_Sony_2
• 12b444vvc1_C_Sony_2
• 10b444P16_C_Sony_2
• 12b444vvc1_E_Sony_2

@nuomi2021 How would you like to deal with this? I could add a timeout flag to the ffmpeg.py, or we could not use these tests – I don't think they test any features not already covered by smaller bitstreams.

[nuomi2021]

@frankplow , every conformance clip has some special features.
better include them all.
Are they all big clips? if the fps is 10, 300 * 10 = 3000 frames. Should be enough for most of clips.

thank you

[frankplow]

@nuomi2021 They're all only 50 frames but large (4K, HDR). I think we add them for now and if it proves to be a problem we can look at solutions.

[nuomi2021]

sure, it may mean we have performance issues for these clips. we can only get 50/300, 1/6 fps for every frame.
Go ahead to lose the timeout condition and add them.