Anyone having success capturing hours of 4k video, reliably and with low loss, using ffmpeg?
I'm working with a video production company that is thinking about using
ffmpeg for a field production. They set up a 4k camera on top of a
building (have electricity, but limited internet), and they need to
capture 6-12 hours of 4k 29.92fps video from that camera reliably and
with very low quality loss. We know we can use ffmpeg to excerpt and
zoom in on parts of the main video afterwards, if the recording system
doesn't crash in the middle of the shoot, and if the main video retains
Vendor A seems to be saying, buy our expensive system, and it will
capture the video and put it on a file server which your computer can
access. But my boss is saying, can't ffmpeg just capture and write out
the video, and save us the expense of Vendor A?
We tried a simple experiment. We set up a 4K camera in the office. A
straightforward ffmpeg invocation did capture a couple of minutes of
video. But there were nasty artifacts, such as 25% of the pixels being
corrupted in the first second or so, and occasional speckles of
corruption. This tells me that reliable capture by ffmpeg is easier than
"impossible", but harder than "works without even trying".
Is anyone on this list capturing hours of 4k video, reliably and with
low quality loss? Or, has anyone tried it and discovered that it's
really, really hard to do it well? Is ffmpeg up to the task of running
for 12 hours of capturing 4k video without crashing? What bottlenecks or
weak points in our capture card or computer do we need to watch out for?
I looked through the archives of /ffmpeg-user/. The most relevant thread
I found was this: "4K 60Hz Directshow Video Capture" (Feb 2018). The
conclusion I drew from this thread was circumstantial evidence that my
task might be reasonable for short periods, and no evidence about longer
periods like 12 hours. a) they are trying to capture 60Hz, and we only
want 30Hz. b) it really matters that the capture card exports something
which is accepted by the ffmpeg code you use. c) disk I/O is a primary
bottleneck, CPU speed is a secondary bottleneck. d) the right GPU card
can do the encoding, relieving the CPU speed bottleneck.
The drawback of /ffmpeg-user/ is that generally people post here about
problems, not about successes. If there are people quietly having
success with ffmpeg for 12 hours continuously or ffmpeg for 4K video at
29.92 fps, I'd love to hear it.
Re: Anyone having success capturing hours of 4k video, reliably and with low loss, using ffmpeg?
On 2018-06-28 14:21, Carl Eugen Hoyos wrote:
> 2018-06-28 21:58 GMT+02:00, Jim DeLaHunt <[hidden email]>:
>> We tried a simple experiment. We set up a 4K camera in the
>> office. A straightforward ffmpeg invocation did capture a
>> couple of minutes of video. But there were nasty artifacts
> Command line and complete, uncut console outupt missing.
> In this case, you should also tell us which operating system
> you have, which cpu and which graphics-card (if applicable).
You are missing the question.
I am not asking for help to diagnose that simple experiment. I am asking
if others have had experience with the top-level goal: capturing hours
of 4k video, reliably and with low loss, using ffmpeg.
I can use whatever capture card, operating system, CPU, and graphics
card are necessary to build a system that works, and costs less than
what Vendor A would charge us for their closed system.