Use After Effects and Vegas together, or use Premiere instead of Vegas?

[poisondeathray]

Hold on, actually there is some "weirdness" in the clip, it doesn't quite match...

Can you post a longer clip? There's not enough frames in that sample. It can be from the vob, mpeg, or m2v, but preferably with some horizontal panning motion

[Morten]

A longer clip is uploaded in the folder

[poisondeathray]

Thx,..It's ok...false alarm...it's a normal clip film clip (23.976p)

Just be aware some editors can screw up the IVTC. Vegas doesn't accept .m2v files, but for example After Effects does, and it screws it up (even manually trying the different modes)

I wasn't able to test in vegas, because of the .m2v, so I'm not sure how it will handle the IVTC. If you want to upload a .mpg file I can test for you

To be sure, you can use avisynth IVTC methods, which are more flexible and usually more accurate, and it works on your sample. If you want to learn this method I can show you. This is the method I usually use (I usually avoid NLE's for deinteracing, pulldown removal etc..., because they suck)

[Morten]

I tried the IVTC method in Vegas, and now it seems that the video quality is unchanged. Thank you so much for all your help. I really appreciate it

What would you recommend to render it as, when my project is all done? And do I have to change project settings for that? Because I tried to render in different formats, and sometimes I gets a letterbox on this clip. For example if I render in wmv or avi with a NTSC DV template. If I render lossless AVI I have no letterbox. (Which I would prefer not to have, since all my other clips haven't letterbox)

[poisondeathray]

Really? Are you sure it works ok? Step thru each frame in vegas....you will notice some screw ups. Did you test on that clip or a longer video?

I made some blank audio, and multiplexed it to mpeg just to test your clip in vegas, and it screws up the IVTC as well (just like Premiere and AE), so here are instructions on one way to process it so it works:

There is a lot of info to digest, but if you want to learn more, you can read the included documentation for avisynth and the various filters. Different videos require different approaches so don't necessarily do this for every source

This guide is helpful too for figuring out if the source is interlaced, progressive, or 3:2 pulled-down
http://neuron2.net/faq.html#analysis

1) Install avisynth


2) Download the TIVTC package, and put the TIVTC.dll in your avisynth/plugins folder
http://avisynth.org/mediawiki/TFM


3) Run your MPEG2 video with DGIndex, leave the video=>field operation to honor pulldown flags, file=>save project


4) Open in notepad a text file in the same directory as your video and the .d2v generated by DGIndex, and write the following, changing the filenames to match. Rename the .txt extension to .avs, the name of the file doesn't matter. It could be "myvideo.avs" for example

MPEG2Source("myvideo.d2v")
TFM().TDecimate()

There are several different methods, but this is just one method of IVTCing and seems to work fine on your sample

To preview that .avs script you can open it in avsp, or vdub.


5) To encode that to a lossless intermediate, You can use vdub. Set video to "fast recompress", and then video=>compression to either lagarith or huffyuv. File=>save as avi. The audio should already be demuxed using DGIndex, and you can import that directly into your vegas or AE


Cheers

[poisondeathray]

What would you recommend to render it as, when my project is all done? And do I have to change project settings for that? Because I tried to render in different formats, and sometimes I gets a letterbox on this clip. For example if I render in wmv or avi with a NTSC DV template. If I render lossless AVI I have no letterbox. (Which I would prefer not to have, since all my other clips haven't letterbox)

When you "render as" there is a checkbox on the bottom of the dialog box stretch video to fill output video size (do not letterbox), you want to check that

What you render as depends on your goal. The source is not 1:1 pixels, it's 4:3. To get it right, (for a pc display or uploading to websites for example), most people would resize and use 1:1 pixels. For example 640x480 1:1 pixels is commonly used for 4:3 aspect ratio. If you were using WMV, you would probably do this. The only time you wouldn't use square pixels is probably for DVD

EDIT: I retested on Vegas9 (instead of 8 ) and it's better, but there's definite blending going on. Are you using a different clip?

[Morten]

Really? Are you sure it works ok? Step thru each frame in vegas....you will notice some screw ups. Did you test on that clip or a longer video?

Yeah you are right. Some frames look wierd. I didn't noticed it before.
Thanks for the guide It seems a bit complicated though?

How familiar are you with the Adobe Production Premium Suite? 'Cause I was thinking that this dynamic link function could do the job, or will it mess the frames up too?

I retested on Vegas9 (instead of 8 ) and it's better, but there's definite blending going on. Are you using a different clip?

I use the same clip, but you are right; there is some blending.

[poisondeathray]

Really? Are you sure it works ok? Step thru each frame in vegas....you will notice some screw ups. Did you test on that clip or a longer video?

Yeah you are right. Some frames look wierd. I didn't noticed it before.
Thanks for the guide It seems a bit complicated though?

How familiar are you with the Adobe Production Premium Suite? 'Cause I was thinking that this dynamic link function could do the job, or will it mess the frames up too?

I have the Master Collection, the dynamic link is nice, but like I said in an earlier post in this thread it's not essential

I tested that clip on AE / Premiere and manually went through all the pulldown removal modes, and none of them got it right. It might be a hidden setting or switch, but I sure cant find it... The only one that worked properly is avisynth . I HATED avisynth when first learning, but now, it's the first thing I go to. It's that useful to learn. Really!

It might just be specific this video (other DVD usually work better using the regular vegas/premiere IVTC method, it's just that avisynth is much, much more powerful for IVTCing and deinterlacing , and way more configurable with a zillion more options than Adobe or Sony products)

Many people think they'll fiddle with after effects, and presto, get awesome results, but the truth is 90% of the work goes into preparation , and planning of the assets (including IVTCing, generating photoshopped assets, etc...)

[Morten]

I have the Master Collection, the dynamic link is nice, but like I said in an earlier post in this thread it's not essential

Maybe there is something I misunderstood, but I thought that if you imported the source material into Premiere, it would export it to AE in a way, that doesn't mess with the quality at all? So because of that I would not have any problems?

I guess I will take a look on AVIsynth tomorrow, since it is 3 am here. Forgot the time completely Again, thank you so much for your help so far. You have been great.

[poisondeathray]

I have the Master Collection, the dynamic link is nice, but like I said in an earlier post in this thread it's not essential

Maybe there is something I misunderstood, but I thought that if you imported the source material into Premiere, it would export it to AE in a way, that doesn't mess with the quality at all? So because of that I would not have any problems?

I guess I will take a look on AVIsynth tomorrow, since it is 3 am here. Forgot the time completely Again, thank you so much for your help so far. You have been great.

As I said earlier, you can import MPEG2 sources without difficulty into Premiere and AE CS4, it must be a limitation of the trial, or something peculiar with your system setup. The dynamic link just makes it possible so don't need an intermediate render to re-import into the 2nd application (streamlines your workflow). eg. you might do compositions of various sections in AE and link them to Premiere in your main video.

The problem is both programs have difficulty with your clip, and do not apply IVTC correctly (you get strange blending and artifacts). AE has 10-12 manual modes to remove the pulldown, and I went through each one, and none of them got it correct.

I forgot to mention another big benefit of premiere/AE over using vegas is the avisynth premiere avs import plugin. So you can open your .avs scripts directly in premiere or AE without that big intermediate file (all that vdub or huffyuv steps). The ability to integrate avisynth with premiere is one of the big reasons I moved away from vegas. But vegas has a useful feature that Premiere doesn't: debugmode frameserver. It allows you to access other encoders easily without the lossless intermediate step. It doesn't work for Premiere CS4.

[Morten]

The problem is both programs have difficulty with your clip, and do not apply IVTC correctly (you get strange blending and artifacts). AE has 10-12 manual modes to remove the pulldown, and I went through each one, and none of them got it correct.

Let me se if I got this right. My dvd is on my hdd in MPEG2-format. And it is this source file that is a problem to begin with? So should I use AVIsynth on the MPEG2-file that my clip comes from, before I import it into a NLE? Or could it maybe work if I imported my dvd in an AVI-format instead of MPEG2 to begin with? 'Cause I find it very strange that there are so many problems with the original source. Is this why MPEG2 is not recomended as an edit format?

[poisondeathray]

Let me se if I got this right. My dvd is on my hdd in MPEG2-format. And it is this source file that is a problem to begin with? So should I use AVIsynth on the MPEG2-file that my clip comes from, before I import it into a NLE? Or could it maybe work if I imported my dvd in an AVI-format instead of MPEG2 to begin with? 'Cause I find it very strange that there are so many problems with the original source. Is this why MPEG2 is not recomended as an edit format?

I think there is something peculiar about that specific clip. For example, I just tested a couple of other retail DVD clips and they worked fine with AE's and Vegas IVTC on the first try (no manual fiddling). I'm not an expert on the DVD and IVTCing and fields area... If manono or jagabo (other videohelp members) had a look at it they could maybe explain the reason why. It might be a phase swap or maybe have to do with the way you cut the segment. If you trim a few frames forward it might work better. This is why I avoid NLE's for IVTCing, they are not as flexible and are not as likely to get good results. And I completely avoid NLE's for deinterlacing - they are HORRIBLE.

Even if you imported that specific in lossless AVI - without avisynth processing - the field order and phase remains the same and you will have the exact same problem. This might be related to the the importing as MPEG2 vs. AVI discussion as in the area that you cut might be "orphaned" from the preceding I-frame (I'm just guessing), whereas if you converted to all I-frame lossless AVI (remember you can have long GOP lossy AVI as well, so not all AVI is necessarily "good" for editing). But that is not the case. When you look at the clip in a stream analyzer, the GOP sequence is intact and preceded by an I-frame



What's weird is that the manual IVTC modes in AE failed. Every combination of phase, field order, whole frame (w) and split frame (s) failed. When you have 3:2 pulldown content, the three whole video frames contain two fields from the same film frame. The remaining two split-field frames contain a video frame from two different film frames. The two split-field frames are always adjacent to each other. The phase of 3:2 pulldown refers to the point at which the two split-field frames fall within the first five frames of the footage.



My recommendation is to use avisynth before you will definitely get better results, especially on interlaced footage. It's crucial to get clean assets when using AE (no deinterlace or bad IVTC remenants), otherwise edges get ratty and effects get distorted. Compositing doesn't look as real because your edges are not clean.

[Morten]

Hi again.
I have been busy with other things, and I am now back working with this project. I am looking at AVISYNTH, but I then got an idea. I would try to save my clip as an .aaf-file in Vegas, and then open in after effects, to see whether this would work or not since it would be a more simple workflow. But I still have a problem with importing mpeg-files into AE. When I try to open my aaf-file it says:

"After Effects error: File "movie.mpg" cannot be opened, it may be in an unsupported format.

(86::1)"

This is also the message I get when I try to import mpeg-files into AE.
Can anyone confirm if this is because of a trial limitation?

EDIT: The .aaf was not better than importing an avi. I now have the full version of AE and I can now import mpeg-files But I have some problems with the script thing in AVIsynth. I wrote the script (with the "myvideo" changed of course) but I can't open it in vdub. When I try to open the avs-file it says

Avisynth open failure:
Script error: syntax error
(folder\myvideo.avs, line 1 column, 7)

My script is:
MPEG2Source("raw_title_7_chapter_1.d2v")
TFM().TDecimate()

What is wrong?

[Morten]

Maybe you can tell my what I am doing wrong with Avisynt, poisondeathray? You sure have been helpful all the time

[poisondeathray]

So it was a limitation of the trial then?

Is the .d2v and .avs file in the same same folder as your video? If not, you have to specify the full path

Make sure your spelling, directory pathnames, etc..is all correct

Do you have DGDecode.dll from the DGIndex package in your avisynth/plugins directory? If so, this will autoload it. If not, you have to load it manually in your script (there are instructions in the quickstart manual)

LoadPlugin("...\DGDecode.dll")
MPEG2Source("myvob.d2v")
TFM().TDecimate()

[Morten]

Yeah the trial is limited. After I typed my serial number, it updated the program, and now it works perfectly. So that's great.

I tried to copy DGDecode.dll into the plugin folder, and now it works fine. As long as I am working in 24fps, there is no blending, no jagged lines or anything even if I render the final project in 29,97fps. The video rendered from after effects had another aspect ratio than the original one from Vegas, which must be some setting in AE I guess, that gives a kind of widescreen look(and therefore black bars at top and bottom). This is easily fixed by unchecking "Maintain aspect ratio" in Vegas, which makes the aspect ratio equal to the one in the original clip. So everything is working now

I still find it wierd that it screws my clips because of the fps-number? Is this something typical when using DVD footage in 29,97fps?

Thank you very much for helping. Avisynth sure is a nice program. You are using Avisynth for all your movies, right? Would you recommend me to use it on all movies too, even if no blending, jagged lines etc etc can be seen? 'Cause as far as I understand, when I IVTC with Avisynth, I will always be sure to have a more clean picture when I work with telecined footage/DVD footage?

[poisondeathray]

I'm glad you got it working

Yes avisynth is very powerful, and you're just scratching the surface. It can do many many more things. I suggest you start learning the basics, you won't regret it I promise!

If you recall, I tested other retail DVD's in both Vegas and AE, and the built in IVTC worked fine for them, so it must be your specific sample had a mismatched field or phase order - which requires an adaptive method to "fix" it. The others assume the same linear pattern, so the first half might be ok, and the 2nd half gets screwed up or vice-versa. So this is a unique case, not the norm.

However, if you have truly interlaced content, DO NOT use vegas or AE to deinterlace, they are horrible. They drop 1/2 the fields and leave jaggies everywhere. In "blend" mode for Vegas, everything is a blurry mess.

In general, a source from video is usually 29.97 , a source from film is usually 23.976. Animated content, CGI stuff can be either. In order to make it work on NTSC systems for DVD, film has to have pulldown 23.976=>29.97. In order to recover the original content you just reverse the process or inverse telecine (IVTC), or pulldown removal.

Don't render the file as 29.97, if your content is 23.976. Unless you do it properly (ie. pulldown), the mismatch will cause the editor with either dupe frames or drop frames or blend frames (depending on your project settings and how you do it)

Cheers

[Morten]

Well unfortunately, it seems that all the dvd's I use, have to go through Avisynth before I use them. I used some clips from different dvd's, and I had blending problems with all of them. But I guess it isn't suprising since it is the same tv-show and the same dvd distributor. So I guess I will have to take all my imported files in vegas, and replace them with versions that have gone throgh Avisynth.
I am especially afraid that I will get some problems with masking if I don't use Avisynth on them

Again I would like to thank you very much for all your helping. It is a big relief to know how I can make it work again. You have been so helpful. I will now go back to play with AE

Have a nice day