I'm a programmer. I have a PhD in computer science, know my way around image processing (4 grad courses) and am basically a capable developer. I've programmed microcontrollers, FPGAs, clusters of 10,000 computers, a super computer here and there, my cell phone etc. I have a full time job with a great company that I love. There are other's here who are physicists, enthusiasts, modellers, calibration specialists, etc.

So, why haven't any of the hardware manufacturers out there decided to leverage our collective intelligence to create a box that we can all program? What are the technical challenges behind providing a plug in API and an SDK for a company like ABT or Lumagen?

Wouldn't one of the many frustrated RS1 owners have implemented a decent CMS by now? Open sourced algorithms to do image processing for our video setups would be a benefit to everyone.


Seriously, ABT, Lumagen: Give us a way to help us help you help us.

Hmm, isn't that what ffdshow is all about?

Yes, dScaler and ffdshow are definitely there exactly for that reason.

Many of the technologies which have driven the market forward have come from companies whose main purpose is to create IP and leverage off it.
Faroudja's DCDi, ABTs PReP, Lumagen's scaling, Algolith's Noise Reduction, etc, etc.
None of these companies would allow open source access to their chips and even some of the smaller chips like the HDMI chips, etc will all be tied up in NDAs.

I think Crystalio I was based on linux, so it's definitely not rocket science, but I think for the time being you're going to have to rely on hugely multi-purpose processors and GPUs in PCs to get what you want.

Also, open source systems can often be over complicated and poorly documented - I think that would definitely cause problems for many. Our HTs are compicated enough already without having to look up every menu option on the internet.

There are a couple of bugs in my VP I think I could have fixed by now if I had the code (and the development environment, and the documentation, and the test hardware, and enough time to delve into it all and learn how it's put together...)

It is very frustrating, though. If I were set up with all that, I seriously think I could fix my biggest gripe with my VP in about one easy afternoon. I know it isn't practical for me to actually do it, but believing I could just makes it more annoying that the company doesn't have its paid developers do it.

Hmm, isn't that what ffdshow is all about?

Can I process HDCP content with it? This is the problem...

Two reasons I can think of off the bat:

1) Most of these comsumer scalers are built around an OEM scaling chip. Faroudja, Gennum, Silicon Optix, to name a few. The real meat of the scaling algorithms are burried in the silicon. Yes there are registures and some flexibility left to the product integrator but the quality of the image processing to a large extent is built into the chip you choose.

2) Support: How do you offer support and product upgrades when everyone has tweaked your product into something different?

Note that many of these chips and chip sets do have evaluation boards available. You could purchase one and go from there with your level of expereince. I have considered this as well but I finr Dscaler meets my needs for SD to HD upconversion. I don't use a scaler on HD sources.