Questions about TPRivtc.dll

[por_que2k]

I LOVE tprivtc, it lets me do the exact ivtc I want, and ir runs real quick in CCE, since it's not doing any actually filtering.

I have a couple questions though:

Can it be updated for AVISynth 2.5?

Is it posible to read in not only the frame selections, but also the frame-by-frame deinterlace settings? Sometimes ivtc isnt perfect, and you need to deinterlace specific frames. tprivtc doesnt seem to support this... can it?

I'd really like to have a fully functional tprivtc, then I wont need to filter through tmpg at all in any of my encodes :)

[Kurosu]

Here is TPRIVTC for Avisynth 2.5, supporting probably all colorspaces (although only YV12 and YUY2 were quickly tested).

It uses the IVTC information from Tsunami MPEG Encoder Project Files (some pretentious 'reverse engineering' was needed there), and is able to perform the following deinterlacing processing:
- interpolate odd field
- interpolate even field
- double (ie blend fields)
- nothing

A help file named 'TPRIVTC.txt' is provided in the package and explains how to use the filter.

[EDIT]
Maybe I should post a new thread, or it may pass unnoticed, in spite of the paper cliper <-> attachment)

[Zarxrax]

Thanks you so much, I have been waiting for one of these for days :D

[por_que2k]

hhmm, did you attach the file to your post? I cant find any link...

[BlackMetal]

Nice work, Kurosu!
I've waited for an Avisynth 2.5 version for 100 years at least.:D
You da man. Gimme five!
Now all we need is a link.:)

[Kurosu]

1) This was quickly tested, I think it does the job, maybe you won't find so.
2) I've looked at the post, it does have an attachment. I want to restrict the plugin to that thread, so no weblink untill it's approved. All will come in time, just be patient, thanks for understanding.

[The People's Elbow]

I'll have a look at it, either ^^
Good Job, if it can handle the interlace settings of tmpg it's even an improvement.

[The People's Elbow]

Kurosu, the plug-in passes the ivtced frames nicely and without a problem, but the deinterlace settings I set in tmpg seem to be ignored... any ideas?

regards, elbow!

[Kurosu]

To make it straightforward:
1) "No deinterlacing",
=> The frame is returned as is

2) "Even field",
3) "Even field (adaptation)",
4) "Even field (animation adaptation)",
5) "Even field (animation adaptation 2)",
=> The even field is interpolated (3, 4 and 5 are just processed as 2, I have no idea of what those particular settings do)

6) "Odd field",
7) "Odd field (adaptation)",
8) "Odd field (animation adaptation)",
9) "Odd field (animation adaptation 2)",
=> The odd field is interpolated (7, 8 and 9 are understood as 6)

10) "Double",
11) "Double adaptation",
12) "Double field",
13) "Double (field,adaptation)",
=> The fields are merged (11,12 and 13 are understood as 10)

14) "Odd/even",
15) "Odd/even (adaptation)",
16) "Odd/even (field,adaptation)",
17) "Odd/even (field,adaptation 2)"
=> The even field is interpolated (I have no idea what that 'odd/even' means; 15,16 and 17 are understood as 14)

I don't really want to dig into what all deinterlace methods do, nor how they work, so I'll definetely leave it as such, unless someone explains me simply what they do, and if it's easy to implement. I didn't mail TMPEG's author, and I don't plan to do so, unless I have to because of copyright stuff. The only addition I can see (not bugfixes) is to invoke for instance another deinterlacer.

[BlackMetal]

Kurosu, it seems the filter doesn't work correctly.

I IVTC'd The Abyss S.E. in TMPGenc correctly, but in avisynth2.51 and with your tprivtc filter, sometimes a frame is replaced by another frame from some other place in the movie.
I tested with avisynth2.0.8 and Daxab's tprivtc filter, and there it works perfectly. :)
If you don't find the error in your code, maybe you could look in daxab's source how he did it? He offered to send it to me once, but I didn't need it, cause you made a tprivtc filter.
But now I've messaged him again, and asked for it, so I hope I'll get it soon, it seems he's not online that often.

[Kurosu]

As we still can't add attachments, I simply link to an external link the new source code + dll. You can find it here (http://kurosu.inforezo.org/avs/TPRIVTC.zip).

The replacement occured on each MOD65536 frame. So with any shorter encode it was ok.

[BlackMetal]

Kurosu, thanks, now it works!:)
Thanks for the fast help and the fast reply.

Just another little thing, what is a MOD65536 frame? Is it every 65536th frame?:confused:

[Kurosu]

On some occasions, a frame index must be increased by 256 (2^8) to point at the right frame. Otherwise, it would do: frame 510 then 248 then 516, while the expected value was 514.

I thought the same was happening with 65536 (2^16), and automatically added this vale. This has proven wrong.

Notice: it only affects the input frame (with DoubleWeave() before TPRIVTC use). The output frame can be any depending on your frame rate . The best mean to check is to use DebugView and see what frame is actually used: instead of using 65538 (for instance), it was using frame 131074.

Anyway, my definition wasn't correct, but I was too lazy to make a proper one.

[orulz]

I can't find a link to this AVIsynth 2.5 TPRIVTC filter anywhere... either the attachment is gone, or whatever. Could someone please repost it?

[neuron2]

Look closely at the 11th post in this thread.

[orulz]

I don't know how many times I looked over the entire forum missing that link.

Thanks.

[Dreassica]

double outputs a green frame with tprivtc.dll

[Kurosu]

Ah, I thought I had updated my webpage...
Well, it is now. Please go have a look :)

[Fahruz]

Ooooh, up date of webpeji is the extremely nicest! :D

[Kurosu]

This is a warn of you, but I am no good webnushi and often broken the links it is. So, let's mailu with me when it cannot access the page, and quick I fix it bely.


I don't really care about my webpage, so you may often find a broken link or some other quirk. Whenever this happens, just mail me, and I'll fix that ASAP.