5
\$\begingroup\$

I want to take a NTSC/PAL component video and overlay graphics over it.
I am aware of solutions that either use the max7456 overlay chip, or accomplish it with just an AVR but these seem to have limited resolution and are monochromatic. Ideally I would like to have something comparable to the performance of this setup.

Here is what I know I will need:

  • Video decoder (to digitize NTSC/PAL)
  • Some sort of buffer/processor to overlay graphics onto the digitized frame
  • Video Encoder (to re-encode to NTSC/PAL)

The middle part is what I am not sure how to setup. Are there existing chips that will automatically do this overlaying for me somehow, or do I just need a sufficiently powerful MCU and RAM to buffer/manually manipulate a bitmap? Ideally I would like to have something low cost, and preferably a single chip solution.

\$\endgroup\$
5
\$\begingroup\$

Something like an ADI Blackfin could do that, or an FPGA with some fast RAM.

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

I'm not sure about the overlay part, but you can generate an NTSC/PAL video signal on an AVR using AVGA.

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

I have been working on my project Super OSD which supports 192x128 pixel graphics and is open source. I'm planning to migrate to a high resolution display - up to 512x384 pixels bilevel (black and white pixels) or 320x240 with 2-bit greyscale, on a PIC32. A dsPIC33F with SPI and external muxes can easily do it (it's even possible without muxes or SPI, but you get less processing time that way), and it's available in DIP.

\$\endgroup\$
0
\$\begingroup\$

what about a BOB4 - http://www.decadenet.com/

not really cheap but I used them in my last job to good effect

\$\endgroup\$
0
\$\begingroup\$

For OSD solution, an FPGA-based approach could be an alternative.

  • OSD can be achieved with an alpha-blending implementation.

The alpha-blending is an algorithm for mixing 2 images into one. The good point is that this gives you the possibility to configure the level of transparency of individual picture elements.

In few words: being x and y the inputs and z the output video signal. An alpha-blender circuit can mix them implementing the equation:

z = x.(alpha) + y (1-alpha)

alpha is the coefficient or level of blending. Then you can define "x" as the NTSC video and "y" as the overlay.

  • Additionally, an FPGA could match your needs in case you may consider additional logic as the video encoder/decoder you mentioned. Everything (enc/dec and their required memory interfaces and the OSD) could fit on a single FPGA (if big enough).
\$\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.