Update FFmpeg for Rockchip

[MarcA711]

Changes:

• update FFmpeg
• update docs
• remove go2rtc presets for older FFmpeg version

ToDo/ known issues:

• reimplement hardware object detection; I want to use yolonas as model, but I am not sure if there will be new license issues, does anybody have experience with this?
• include hardware acceleration presets for go2rtc (see Update Rockchip presets AlexxIT/go2rtc#979)
• there is a hardware limitation that can lead to crashes (see 4G limit on an 8G board rockchip-linux/mpp#560)

[netlify]

✅ Deploy Preview for frigate-docs ready!

Name	Link
🔨 Latest commit	15a6cc4
🔍 Latest deploy log	https://app.netlify.com/sites/frigate-docs/deploys/66411c8d4f25b10008278fb0
😎 Deploy Preview	https://deploy-preview-11351--frigate-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[NickM-27]

@blakeblackshear may be able to provide more info on the first question

[blakeblackshear]

I believe yolo-nas should be fine. I don't think they interpret the AGPL license quite the same way. I will need to train from scratch to use it in Frigate+, but I think it can be used here. You may want to ask in the YOLO NAS repo directly though before doing it.

[MarcA711]

Thank you very much for your help. I did some additional research and asked here: Deci-AI/super-gradients#1993

[MarcA711]

Maybe it was a bit confusing what I wrote. This PR is ready to merge. The things that I mentioned as "ToDo/ known issues" was meant to address in separate PRs. However, if this PR was not yet merged because you are busy, don't feel rushed.

[MarcA711]

Hey @blakeblackshear,
unfortunately, I did not get a conclusive answer if I can use yolo-nas with pre-trained weights (see Deci-AI/super-gradients#1993). Do you think it is okay to use the pre-trained weigths as long as they don't object?