Ross Patterson

> No code changed needed, just yaml Really?! Awesome. I haven't said it yet, but I've been meaning to, I love the way you've implemented ytdl-sub at the lower level....

> It may still be that throttling _is_ based on video duration. I've moved onto a new subscription where the videos are often longer, significant numbers of them are 90+...

> No code changed needed, just yaml. I'm having a hard time getting this working, I keep getting: ``` [ytdl-sub] Validation error in _youtube_base.throttle_protection.sleep_per_download_s.max: should be of type float. ```...

> I've moved onto a new subscription where the videos are often longer, significant numbers of them are 90+ minutes, and some are even 100+ minutes. While downloading this subscription,...

> These do not support overrides at this time, only floats: So does that mean that this isn't the case afterall: > > > No code changed needed, just yaml....

With #1288, #1315, #1316, we're almost ready to call this feature implemented. I'm testing the following preset snippet for different values of `throttle_protection_duration_factor:`: ``` overrides: # What proportion of entry...

Because the "evidence" trying to tune download delays is so anecdotal and not rigorous, this comment is to capture some details of my testing so far in the hopes it...

I prefer breadth of content over maximum quality in my library, so I've been using the `Max 1080p` preset in the hopes of getting AV1/HEVC/VP9 videos. I could swear that...

> So now I went back to fixed ~5 minute delays to see if I can intentionally trigger the throttle and reestablish a "baseline". I got through 21 downloads with...

> I'm definitely unable to reproduce the resolution quality throttling ATM. While I'd love to provide more testing anecdotes, I'd love more to back-fill my archive faster and I don't...