First time deinterlacing footage HELP

[ultimateX021]

Hi everybody.

After all the time spent trying to do homework, mindless internet surfing, and mixes of priorities I am reading the guide to all technical things about amvs. I am at the stage of removing interlacing in footage from a DVD. Using the sample footage, I followed the steps closely determining if it was purely interlaced or a hybrid source. From analysis, it is pure interlaced. However, I'm skeptical whether or not it is good quality if the footage runs slow on VirtualDub and AvsP. The commands I used were TDeint(mode,type) mixing in values as a test of the results. Another command I used in a copy file for comparison was the AssumeFPS plugin. I notice both show lack of interlaced horizontal lines, but AssumeFPS makes the footage run at its normal speed (Or somewhat fast at least).

Would the speed of the footage reveal the best quality as well as the lack of interlaced footage? Or do I need to look out for other factors?

[Zarxrax]

Its unlikely that its full interlaced. What anime is it?

[ultimateX021]

Orphen, an anime about a spellcaster seeking this dragon called the Bloody August. It was developed in 2001. Not as old looking compared to anime made in the 80s or 90s. So, how is it unlikely that the footage is purely interlaced?

[Zarxrax]

Pure interlaced it just a very rare thing. I think I've only ever seen 1-2, ever. But being that your dvd is so old, it may very well be pure interlaced. Is it PAL or NTSC format?

An easy way to check is to apply the following script:
TFM(pp=0)

If its interlaced, you should still see some interlaced frames after doing that.

[ultimateX021]

Pure interlaced it just a very rare thing. I think I've only ever seen 1-2, ever. But being that your dvd is so old, it may very well be pure interlaced. Is it PAL or NTSC format?

An easy way to check is to apply the following script:
TFM(pp=0)

If its interlaced, you should still see some interlaced frames after doing that.

NTSC format, brought to you by ADV films, lol. That is basically what it says on the back of my dvd cover.

So I'll see what will happen if I apply that command line you just gave me. I'll be back to report any slight changes or things I notice once I use it.

Thanks

[ultimateX021]

I used the TFM function at the end of my script as follows:

mpeg2source(source of orphen clip)
KillAudio()
Crop(8,0,-8,-0)
Spline36Resize(640,480)
Tweak(hue=0.0, sat=1.0, bright=11, cont=1.3, coring=true, sse=false))
TDeint()
AssumeFPS("ntsc_film")
TFM(pp=0)

I watched most of the clip on AvsPmod and VirtualDubMod. I saw no interlacing at all compared to before I applied the TFM command. I tried it without AssumeFPS and still saw interlacing. With no interlacing after applying TFM at the end, does that mean there is no interlacing at all?

[ultimateX021]

Whoops.

I meant "does this mean it is not purely interlaced?"

[mirkosp]

That's not what you should have done.
First thing first, mpeg2source doesn't load audio, so killaudio is useless there. Also, orphen isn't an old source (like, it's not from the 70s or early 80s), so using tweak is strongly discouraged.
For starters, do this

Code: Select all

mpeg2source(source of orphen clip)
TFM(pp=0)

Go around the source, see if there are obvious interlacing leftovers (not just small combs).
If there are, your source is interlaced, if not it's telecined. It is also possible that your source is hybrid, which means that it mixes telecined and interlaced footage.
Hybrid in late 90s/early 00s tends to happen with CG (CG was generally done interlaced while anime was telecined), so that would be something to look out for.
Either way, I really doubt you have an interlaced or even hybrid source (to the best of my memory, orphen didn't have any massive CG usage), so at this point, if that tfm line pretty much fixed it, you should then change your script to this:

Code: Select all

mpeg2source(source of orphen clip)
tfm(mode=1,pp=5,slow=2,micmatching=2,clip2=tdeint(mode=2,type=3))
tdecimate()
Crop(8,0,-8,-0)
Spline36Resize(640,480)

And you'll have your progressive 23.976 clip. Do not use assumefps here, it is not needed.

However, if your source is interlaced, then:

Code: Select all

mpeg2source(source of orphen clip)
tdeint(mode=2,type=3)
Crop(8,0,-8,-0)
Spline36Resize(640,480)

Will give you a 29.97 progressive stream. Again, unless you have to edit your vid at 23.976, you won't need assumefps.

If you are so out of luck that it's hybrid, then make two scripts, one with the tfm and tdecimate as if it was telecined, and the other with the tdeint as if it was full interlaced.
Then, you use assumefps on one of the two scripts, either to bring the ivtc'd one to 29.97 or to bring the deinterlaced one to 23.976.
Finally, while you edit, you'd pick scenes from one or the other script depending on whether it was originally interlaced or telecined.

This is about as easy as it can be put down, I think...

[ultimateX021]

Thank you very much Mirkosp.

I used the TFM command as you showed me. I still saw tons of interlaced parts in my footage, so I assumed it was purely interlaced. However, to be sure I wasn't making a "fatal" mistake, I both scripts assuming it either had mixed in telecined footage in one script and assuming it was plain interlacing in the other. I ran them both several times in AvsP just to see if they looked different. They both looked quite similar in most parts of the footage: interlaced cleared and barely any sign of other problems although I almost thought it had blended scenes, but through playback, the transitions from one scene to the next look smooth to me. Another concern was finding a few interlaced portions in the hands of Orphen when he's about to do a magic attack. This appeared from the script with the TFM and tdecimate commands; not the one with tdeint. (Oh, and I understand the importance of using TSM to restore progressive frames together in case if they weren't interlaced and just was telecined, so I thought that was helpful using TSM in my script! Thanks! ).

To be sure whether or not I may be working with hybrid, I put the AssumeFPS command following all the steps outlined in your message, so i don't place commands in the wrong order or setup the script in a non-functional way. This is how both scripts looked with only one the interlaced one having AssumeFPS:

Code: Select all

mpeg2source("C:\Users\Owner\Desktop\Anime Footage\DGIndex Project Files\OrphenPF.d2v")
Tdeint(mode=2, type=3)
Crop(8,0,-8,-0)
Spline36Resize(640,480)
AssumeFPS("ntsc_film")

Code: Select all

mpeg2source("C:\Users\Owner\Desktop\Anime Footage\DGIndex Project Files\OrphenPF.d2v")
tfm(mode=1,pp=5,slow=2,micmatching=2,clip2=tdeint(mode=2,type=3))
tdecimate()
Crop(8,0,-8,-0)
Spline36Resize(640,480)

I would like to point out that after running both scripts in VirtualDub, both scripts ran at a slow pace. The second script with TFM and Tdecimate ran super slow compared to the first script. If a single script runs super slow or just slow entirely, what does that mean? Would this have an adverse effect on video editing? (Which in this case, slow down the need to synch when it's running "super" slow).

Speaking of slow editing, I split footage in separate AVI clips to save some space in both scripts. I wanted to test them out in Sony Vegas Movie Studio 10, so I determine how my clip will run while editing. I adjusted the settings according to the Technical Guide (I ensured all settings were set to match with the aspect ratio, FPS, Resolution, frame type, etc...). I ran them separately one at a time. The clips altered by the first clip had a "stop and play" phase during preview. Almost jerking and jittering for a bit. I recalled someone on here gave some solutions about dealing with this problem, so I played the clip over and over at certain points until the clip ran at better speed. The same thing with the second footage clips from the script assumed to be telecined. I used the same replay method and it ran at natural speed as well. All and all, both footage types were clear, but kept having them jerks and "stop and play" phases.

So to ensure I am doing this in the correct way possible, what do you think of the scripts and does it sound like something else needs to be fixed? Cause I literally followed all as much as what was said here.

[mirkosp]

The slow editing speed and seeking through avsp just mean that your pc has a hard time with the footage, which, if it's a rather old pc, is actually possible. It does not however mean the footage will not be synced or anything, it just means that editing with the avs files directly won't be smooth.
However, if you did make avi clips to edit with, they should be working smoothly, especially if you used utvideo as a codec for your avi clips.
If vegas can't keep up with those clips, then my guess is that either your pc can't or can just barely keep up with vegas (thus making it even harder to properly playback clips in real time), or the hdd with your footage has a VERY low rpm (or there is some other related issue that undermines a smooth decoding).