a stress test of youtube at 1440p

Started by 70MM13, January 17, 2021, 06:45:10 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

70MM13

With the talk about youtube performance, I thought I would give it a serious test using TNT at 1440p:

I shot this at 2560*1440 (14 bit using TNT) and graded it in resolve to put all it's got into the darks.

Watch it in a dark room on a good screen at 1440p!  Look closely at all the textures and detail of my black shirt under the coat.

As always, no noise reduction of any kind.  TNT at its best!

Encoded using NVENC h.264 @ 30MBPS

https://www.youtube.com/watch?v=T3oikPHyUNY

mineralof


70MM13

It's a setup for the 5d3 made by timbytheriver and myself that facilitates optimising the gains on the camera to improve dynamic range and reduce noise in the shadows, especially chroma noise.

A1ex discovered the registers many years ago but development stopped quickly afterward.  It has since become taboo in this community, but if you search the forum you can find information if you are interested.

The long and the short of it:
If you carefully calibrate your camera, you can gain close to 1 stop of usable dynamic range, which is important to some of us.  It is not supported and there is no cooperation from anyone here.  Quite the opposite.

Everything you need to know is searchable.  I made a couple of tutorial videos some time ago.

It is only useful to anyone who is wanting to push their camera right to the limit.  For "typical" users it is probably not useful.  A possible barometer of applicability would be if you use a hardware tool to calibrate your monitor, and if you can see the details of my shirt in the video above.

Luther

Great video, as always. Two things I've noticed:
- The highlights are over saturated (particularly on your face). Using a cuve rolloff would help it a bit.
- 30MB/s bitrate is too low for that resolution... youtube itself recomends 50MB/s. Personally I just use "crf" (constant rate factor) of value 19. It always gives near perceptually lossless output, without being as heavy as dnxhr/prores.