[FFmpeg-user] Anyone having success capturing hours of 4k video, reliably and with low loss, using ffmpeg?

[Jim DeLaHunt]

On 2018-07-23 10:25, Rafael Lima wrote:
>>   They set up a 4k camera on top of a building (have electricity, but
>> limited internet),
>
> 4K on limited internet? is just in my mind that those two words doesn't fit
> together?
>
>
> What do you mean by " capture 6-12 hours of 4k 29.92fps video from that
> camera"? Is the camera streaming the video somehow and you just need to
> store it?

Thank you for your reply, Rafael.

I'm sorry if my original message wasn't clear. I understand why "4K" and 
"limited internet" might not fit together. The missing part is "enough 
terabytes of reasonably fast SSD storage on the server to which the 
camera is attached to hold the video".

So yes, I "just" need to store it. And I need to store every frame of 4K 
at 29.92fps. And I need the capture system to not run out of memory, or 
crash, during the 12-hour session.  And if an individual part of the 
capture system could crash more often than, say, during 1% of the 
12-hour sessions, then I need some kind of redundancy to allow another 
system to capture if the primary system has failed.

On 2018-07-23 10:25, Rafael Lima wrote:
> ...If it is your only limitation are the [bandwith] and storage as
> ffmpeg doesn't need to process nothing and it could be done with any stream
> saver program

This sounds reassuring, as if this isn't such a hard task after all. But 
I don't hear you saying that you know of someone who has done it.

On 2018-07-23 00:42, Roger Pack wrote:
> It "should" work assuming your transcoding/disk can keep up with realtime
> ...

Thank you, Roger. I am hearing you say "should work", not "did work, for 
a real-life case I know".

On 2018-07-23 05:42, Another Sillyname wrote:
> Yes, but honestly I think you [Roger Pack] are over simplifying it......
>
> For example, are you intending to capturing scenes with high activity
> that requires a much higher bandwidth and transcoding capability? ...
> The capability exists and as you've stated the existing solutions can
> be crazy expensive, but if you're going to homebrew a solution you
> need to do some work/testing to make sure what YOU want to capture can
> be done within YOUR budget.

Thank you, "Another Sillyname". This is what I'm concerned with: that it 
might work, might not, and the only way to gather data is to try my own 
experiments.

My goal with the question was to limit the range of possibilities, by 
having someone come forward with concrete experiences. If I hear, "we 
tried it, it was crazy hard", maybe I should tell my boss to pay for the 
crazy expensive commercial solutions. If I hear, "we tried it, it was 
pretty easy", maybe I encourage my boss to let me homebrew a solution.

In either case, we will have to test it.

Thank you all for your replies,
        —Jim DeLaHunt, Vancouver, Canada

-- 

-- 
     --Jim DeLaHunt, jdlh at jdlh.com     http://blog.jdlh.com/ (http://jdlh.com/)
       multilingual websites consultant

       355-1027 Davie St, Vancouver BC V6E 4L2, Canada
          Canada mobile +1-604-376-8953