You misunderstand me. I'm not saying you need to capture twice in all cases. It's just that this 'relative TBC' is useful in conjunction with certain techniques to eliminate noise. If I were to use the multicap technique with a hardware TBC, there would be no difference - I would still need multiple passes. I would only save a slight amount of CPU time. Even so, I would rather register images in software because I'm registering the image itself, not a timing indicator of it.
I certainly agree that doing corrections realtime saves, well, time. In fact, did you know that Avisynth scripts *can* be run in realtime? Even while capturing and saving to a file? Having said that, Avisynth does have limitations, for example it's not easy to adjust settings in realtime.
So yes, there is an advantage to doing things in realtime - hardware is one way to achieve this - and TBC is for now, practically best done in hardware. However, though you haven't seen software which does software TBC effectively, the technique does exist. I'll dig up the paper for it. From what you're saying, anyone who could write that software could find some commercial use for it. It's something I could certainly look into. In fact you're giving me a good idea of the requirements for a capturing program in this area.
|This question has been split from another post or thread. This was done in order to address the topic more in depth, make it easier to find via forum searches, as well as prevent the other post/thread from going too far off-topic. Please continue this part of the conversation here. Thanks for understanding our tech Q&A policies.|