0
\$\begingroup\$

I'm looking to make something I can put between a video/audio source (DVI/DP) and a 240hz monitor that can, with precise timing, replace single frames with a solid color. (the application is around device testing and calibration and could be used with monitors that support other refresh rates)

So, say, every 100th (1/240)s long frames should display a solid red frame, but otherwise pass through the video/audio stream we're receiving during the rest of the time.

Are microprocessors sufficient for this application? An FPGA board feels overboard, but I'm unsure. Do you have any recommendations? What about if I wanted to modify the color of every pixel slightly in an independent way (increasing brightness, say), to help with calibration?

Thanks for pointing me in the right direction, most advice I've seen has been focused on more complicated processing requirements and I'm trying to avoid implementation complexity if possible.

\$\endgroup\$
1
\$\begingroup\$

Are microprocessors sufficient for this application?

No, Video links in general and high-definition digital video links like DP/HDMI/DVI especially are really high-rate applications. You can't even process this in real time on a normal PC CPU, from the sheer volume of data alone. It takes dedicated hardware to generate it in a GPU; DP can carry significantly more bits per second than datacenter-grade network links (10GBase-XX ethernet). (note that I don't say a PC can't generate 240 Frames per second, I just say that the resulting rate in raw video is too much for normal PC-style computing. I do multi-100-Megasample/s signal processing.)

An FPGA board feels overboard, but I'm unsure.

Absolutely not. ASICs and FPGAs are how you deal with video signals.

Do you have any recommendations? What about if I wanted to modify the color of every pixel slightly in an independent way (increasing brightness, say), to help with calibration?

Get an FPGA board with the necessary interfaces.

In fact, there's already projects that process video, and forward it, for conference recording purposes: See mithro's 33c3 talk:

https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8057.html

I'm trying to avoid implementation complexity if possible.

I think you might be on the wrong project than. DVI/DP/HDMI is not a simple thing, and the timing requirements for any processing done on are not easy to meet without significant FPGA design experience.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks, that makes a lot of sense. The conference recording link in particular is a lot of help as I hadn't seen that in my searching. \$\endgroup\$ – vii Mar 13 '17 at 0:27
0
\$\begingroup\$

A better way to calibrate any monitor is to use the video card to generate test patterns with an application.

For Windows, I prefer DPT.exe then use the video card toolbox to adjust gamma, brightness, contrast etc for video and static desktop on each monitor separately.

enter image description here

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.