Does anyone know how telecidehints.dll works ?

[osgZach]

Just wondering where it takes its hints from and how..

I would like to include it as an option in the rebuild of my Find30FPS utility as an additional option to speed up decombing (As I intend the default to be a TGMC call with a few edimode options) .

However its mostly used with YATTA and as such resources and actual source files are rare to find - let alone documentation.

It seems it doesn't matter whether you use TFM or Telecide or whatever, but I'm wondering of course, where it takes its hints from and how - because I can call it in a script that doesn't even invoke TFM or Telecide and it will decomb frames, if I recall correctly.

I'm unaware of any invocation flags that can be set, (i.e specifying a MI value or cthresh, etc) so it's a little mysterious to say the least.

[osgZach]

Interesting.. I used to use Fieldhint until recently, so that would explain where it was pulling the hints from I guess?
After some investigating of some minor after-image/ghosting type artifacts, however, I began commenting out fieldhints (and noticed it changed the frame count too).. So I guess if I wanted to include it as an option w/out fieldhints, I would have to invoke TFM in the final AVS with appropriate flags. Does that sounds right?

Just so we're clear, the intent is to have the user do a TFM stats run, and then process the stats file (much like YATTA does) to get the usual progressive matches and identify VFR sections, bypassing YATTA completely and removing a few extra tedious steps (when batch encoding the tedium adds up). Since the goal is a quick-"mostly automatic" method in cases where YATTA is unnecessary for further manual processing, or the desire to use it is not there. I'm hoping people will actually find it useful (its still worth it for personal use anyway) as most seem to be scared of YATTA, so maybe this way they will get the most basic benefits of it without actually using it.

I know most would say to just use TFM/TDecimates VFR Hybrid mode, but that produces a video and timecodes file that for some reason, my WDTV Live (and I suspect other hardware devices) will not playback properly.. Coincidentally this stems from the thread that lead to my first meeting with you and YATTA. So I'm hoping to take what you taught me and automate it as much as possible (the Find30FPS python script was a good test case and it worked well, now its about refining and adding options) :p

Although I suppose I should worry about perfecting my algorithms/methods for calculating frame number changes for the Trims and timecodes file before worrying about "extras" like this....

[osgZach]

That's certainly good to know. Although currently I do not plan on implementing Tdecimate as a callable filter (seems counter-intuitive in a program having "options" I know).. Maybe through a custom insert option in the future though.

Are there any problems for this with decomb's Decimate? That's pretty much what I exclusively use.

Truth be told, I'm not hoping to optimize the process, so much as automate it so anyone can do it reliably, without multiple steps, hopping programs, etc.. Under ideal circumstances, the user will not choose TFM anyway and it will only be used to dump a stats file for section ID and frame matches. I'm including it so as not to alienate people from trying it though, but as far as I'm concerned the default option will always be to call TGMC and process every frame, enduring the wait time (And we are assuming the typical user will share this bias) in exchange for no combing, or comb detection threshold guesswork. I'm also making a gamble on the future of Avisynth and users converting to MT / x64 setups for speed.

Bonus is, no worrying about overrides or other generated files, except for timecodes obviously. I think it kind of makes sense in as far as what my goals are. I'm just trying to mainstream a very small subset of YATTA's capabilities, not usurp it as the next best IVTC workhorse.