Go Back    Forum > Digital Video > Video Project Help > Edit Video, Audio

Reply
 
LinkBack Thread Tools
  #1  
07-26-2012, 08:48 PM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
Hello,
I am very new to all this and trying to learn as much as I can. Sorry if this is a stupid question.

I'm trying to rip some footage off a DVD of a film that I was in (I'm an actress) so I can update my demo. I'm using Mpeg Streamclip to do this. I've used the gspot tool to analyze it. Under pics/s it says 23.976 and under frames/s it says 29.97. Should I be telling Mpeg Streamclip to export it at 23.976 fps or 29.97 fps?

I believe that film is always at 24fps, but I read a bunch on your site here about maintaining the original specs of the material when encoding. So I am a bit confused about what do to here.

Any suggestions would be appreciated.

thanks!

Reply With Quote
Someday, 12:01 PM
admin's Avatar
Ads / Sponsors
 
Join Date: ∞
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
07-26-2012, 11:53 PM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Hi,
It depends on the format of your destination demo reel. It's true that film is shot at 24fps, but when it's put on video, it's slowed down imperceptibly to 23.976fps. Another aspect is that, traditionally video has always been displayed at 29.97fps.

So how do you display film at the slower rate? They simply repeat parts of it so that it averages out in speed. This is called 3:2 pulldown.

How it works doesn't really matter; it's just that if you're putting clips together for a DVD, it will all eventually have to be 29.97fps. If you are making some other file, like for the web, you have the option of going straight to 23.976 or 24fps.

So that's my first question. Also list the other programs you are using to put it together, and if all the clips are from film or if there's some parts from normal video.

It's funny, my Dad had a small part in a somewhat famous film, and I went through the same process!

ps I just realized that movie has never been released! I have the only copy now...

Last edited by jmac698; 07-27-2012 at 12:13 AM.
Reply With Quote
  #3  
07-27-2012, 12:20 AM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
HI JMac698,
thanks so much for your reply.

Ultimately where my demo will go is on the web. Everything is done that way nowadays, so I doubt I'll need to burn a DVD of it. HOWEVER, it would be nice to have that option if I were to need to do that for someone in the future.

I am on a PC, and right now am using the free version of Videopad to edit my clips together. If I get too frustrated with its limited capabilities I might try some other editing software geared towards the consumer, but I don't know yet. (Is there one you'd recommend?)

I have a variety of clips, some are from films that I'm ripping off of DVD's. Some are mov and mp4 files that the production sent to me (instead of sending a DVD version), some are from TV shows that I have to rip off the DVD, and some are TV shows that I already have as digital files (I think avi's). So, my clips are all over the map.

Just now I looked at the analysis via gspot of a TV episode I have on DVD and it shows both the pics/s and fps as 29.97. Do I leave it as such when I rip that off the DVD? Or do I change it to 23.976/24 fps?

I've ripped some clips off of one of my film DVD's and set the fps as 23.976, just cause it seemed like the right thing to do.

So, if I plan to go straight to the web it seems like 23.976/24 fps is fine. But what if I want to burn a DVD of it someday? Should I be saving a different version with it at 29.97 fps?

Thanks so very much for your help!
Reply With Quote
  #4  
07-27-2012, 01:18 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Hi,

I've been thinking of how to explain this to you. Someone could write a detailed guide with "click this, click that", or else you can learn some basic concepts. I realize you just want to get this project done, but if you have some understanding, you will be able to problem solve on your own, and communication will be easier.

I should also mention that if this seems overwhelming, well it's not hard in practice. Learning is the price of doing something for free. You could also hire someone to do it all for you, but if you really understood how simple it is, you might feel like it's a waste of money.

23.976fps is best, and will apply to most of your clips. It is suitable for turning into a DVD later.

A few clips could be tricky. There's 3 cases for video, generally:

1 Film with "soft" pulldown. Can be exported safely to 23.976fps. When you do so, the command to repeat parts of the video is simply stripped out, and there is no quality loss.

2 Film with "hard" pulldown. Needs a process called IVTC, but can eventually be exported to 23.976fps. There is a slight quality loss. Mostly applies to older, poorly mastered DVD's.

3 Video. This material is inherently 29.97fps and if you try to change it to 23.976fps, it will look stuttery. Probably applies to older TV shows (Miami Vice), made for TV movies, and videos with CG (Star Trek the Next Generation).

You can't automatically tell between 2/3 with gspot. The best way is to look at a short section frame-by-frame.

Definitions

Soft pulldown - applies to DVD's. Stored as true 23.976fps with a setting which commands the player to automatically repeat parts of the video to stretch out the framerate to 29.97fps.

Hard pulldown - Stored as 29.97fps video. No setting to indicate to slow down on playback. Since the setting is missing, the actual video has to be analyzed to find the original 23.976fps.

video - Stored as 29.97fps video. Can't convert to 23.976.

3:2 pulldown - a method to repeat parts of the film so that the speed averages out to 29.97fps. Can be hard or soft.

combing/interlacing - the presence of horizontal lines during movement in a video. It's like looking through a comb. This only applies to video or hard pulldown film.

IVTC - a process to analyze video to determine if parts of it are repeating. If so, the 23.976fps video is extracted. The result has to be encoded again, which means a slight loss in quality. If the pulldown pattern is regular, this is quite successful. If the film was converted to video and then edited, it will break the pulldown pattern, and the IVTC can slightly mess up.

Deinterlacing - a way to convert video into a format suitable for the web, in terms of removing the combing effect.

Interlaced video - a video which shows the combing effect.

Progressive video - a video which naturally has no combing effect, or from which the combing effect has been removed.

Process general overview:

-Using Gspot as you have been to identify clips with 23.976 pics/s. Use your software to export as 23.976fps.

-If GSpot says 29.97pics/s (it will also say 29.97fps), export as 29.97fps, then determine if it's case 2 or 3. You need software which properly shows a single frame at a time. I can explain this bit later. The process will involve looking through the frames for signs of jagged horizontal lines. The pattern of these lines will tell you what kind of video it is. Briefly, hard pulldown videos will have some frames with no lines, followed by frames with lines. True video will always have lines.

Make any sense so far?

Last edited by jmac698; 07-27-2012 at 01:48 AM.
Reply With Quote
  #5  
07-27-2012, 01:22 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
What I suggest is that you determine whether you have at least one important clip where GSpot says pics/s and frames/s are 29.97fps. Then we can practice the skill of determining if this is video or hard pulldown. In the case of hard pulldown, you can then learn the skill of "IVTC" and successfully export 23.976fps video from it. In the case of video, you must learn the skill of deinterlacing.

Why can't you just leave the 29.97fps clips as they are? It won't show well in web format.

Last edited by jmac698; 07-27-2012 at 01:46 AM.
Reply With Quote
  #6  
07-27-2012, 02:16 PM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
Thanks so much for all that info. And you pegged me correctly that I much prefer to learn and understand the basic concepts.

yes, all of the above does make sense, I think.

question:
how would I know the difference of whether a DVD is soft or hard pulldown?

I think soft pulldown will say pics/s is 23.976, but fps is 29.97, yes?

so how would hard pulldown show up in gpot? would it be the same? I'm making that assumption because you said that hard pulldown does have the original 23.976fps in it, and that it has to be found, so I'm wondering if gspot would indicate that? Or would it show up in gpot as both pics/s & fps as 29.97? (and that's why you're saying it's hard to tell the difference between 2 & 3?)

I do in fact have a clip on a DVD that gpot says both pics/s & fps are 29.97. It's an episode of "ER" from the final season - so it wouldn't be that old, but possibly it was poorly mastered. I don't know.

I'm eager to learn the next step!

thanks!!
Reply With Quote
  #7  
07-27-2012, 09:49 PM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Ok,
If GSpot says 23.976 pic/s and 29.97 frames/s, it's soft pulldown. If it says 29.97 pics/s and 29.97 frames/s, it's hard pulldown or video. But there is a much more precise way of determining this. This is only for DVD/MPG clips!

-Download DGIndex http://www.videohelp.com/tools/DGMPGDec There's no install program, just extract all files and click the DGIndex.

-Open your DVD clip in it. It can be in (decrypted) vob or mpg format. File->Open->select clip. Then click Ok.

-The first frame is displayed. File->Preview. The video is played, and meanwhile analyzed. The only important information is "Coded #" and "Playback #".

Building upon our definitions, I stated that soft pulldown was a command to automatically repeat parts of the video. Here you can see the effect of that command. Coded# means the actual pictures/s. It will read 24 after 1 second. Playback# is what the file commands to be played back. It would say 30 after 1 second. You will also the # Field Repeats. This is the difference between the two. The command is literally to repeat a field*. So you are seeing the raw technical commands embedded in the file. The command is called a pulldown flag.

Another indication is Video Type. For video, it will say NTSC. For soft pulldown, it will say FILM. For badly mastered soft pulldown film, it will say "97% FILM" or something like that.

-Use the arrow keys to step through the video. You will see spots that are combed. An dieal pulldown pattern will be like this, combed, combed, clean, clean, clean, so look for any pattern of 3 good frames followed by 2 combed frames.

-If you see 29.97 frames/pics/s, and coded#=playback#, and a pulldown pattern, you have hard pulldown (also called hard telecine).

-If you see 29.97 frames/pics/s, and coded#=playback#, and no pulldown pattern, but only combed frames, you have video.

*I didn't define what field means. A field is the set of even or odd lines of a video. For example, the odd field is the first horizontal line of the picture, then skipping a line, then the 3rd line. This must seem very odd, but when TV was invented in the 50's, for technical reasons they decided to send the picture one half at a time. The fact that the two fields are interwoven leads to the term "interlacing". We all wish this would go away as it has no use anymore and is generally a pain to deal with!

As I'm sure you can guess, if you have to send two half-pictures to get 29.97frames/s, you must need 59.94 fields/s. That's right! And if a field is repeated to slow down the film speed, it will appear as if the odd lines of the picture are "frozen" in time, while the even fields move on to show their part of the next film frame. Now, two different film frames are displayed interwoven, thus you will see fine "comb" lines where they are different.

Last edited by jmac698; 07-27-2012 at 10:39 PM.
Reply With Quote
  #8  
07-27-2012, 10:46 PM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Quote:
Originally Posted by cabney
...and that it has to be found, so I'm wondering if gspot would indicate that?

Read more: http://www.digitalfaq.com/forum/vide...#ixzz21sxBWSSE
GSpot is a great tool, but it doesn't attempt to analyze the actual images. It only reads information about the file. It's the equivalent to talking about a book from it's library listing without actually reading it.
In fact, AVI is equivalent to a library listing. We call it a container. The container lists what's inside it, but then the video itself is it's own data. The two can even be mismatched, for example, it's possible for a video to go between different speeds, but the container can only list one framerate. How would you handle this? In reality, the video is modified to a single framerate before it can be put in that container.
Reply With Quote
  #9  
07-28-2012, 01:25 AM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
OK, I downloaded the DGIndex.

1) I'm not sure what "decrypted" vob means, but I do know what vob is, so I opened one of the vob files on my "ER" DVD. I analyzed it and it showed same number of playback# and coded#. It also said it was NTSC. The field rpts said 0. So, I take it it was 0 because the playback# and coded # were the same, therefore it didn't need to repeat...Then I scrolled through with the arrows to look at the frames. MOST of them had the combing. Occasionally I came across a frame that seemed to not be combed, but it also seemed to coincide with a section where I knew there was less movement going on. it also seemed like where there was more movement (of the actors or the camera) the combing was more severe. Does this make sense? In any case this one it seems is clearly video.

2) I tried a 2nd one where gpot had already told me it was progressive, but the pics/s & fps it said was 29.97. coded# and playback# are the same. It also indicates NTSC and no field repeats. But going through the frames with the arrows showed no combing. I suppose because it is already progressive. So this one is de-interlaced video, yes?

3) just to see the difference I tried one that I already know is a softpull down cause gpot told me it was 23.976 pics/s and 29.97 fps. This one gives different playback and coded #'s, also shows there are field repeats, frame type is progressive & video type is film. scrolling through with the arrows showed no combing because it's progressive (yes?). See, I;m starting to get it now.

4) gpot showed this next one I tried as 29.97 for both pics/s & fps. DGindex gives me the same # for coded & playback, says NTSC, interlaced, & no field repeats. scrolling with the arrows I see this pattern: combed, clean, combed, clean, clean, combed, clean, combed, clean, clean (so it's 1-1, 1-2, 1-1, 1-2). so it SEEMS like this one is hard pulldown because it has a pattern, but the fields did NOT repeat. So I'm confused about this one.

5) another confusing one: pics/s & fps are 29.97. DGIndex shows NTSC, no field repeats, interlaced. But when I scroll with the arrows there is no clear pattern of combing. I'll go through whole sections with no combing, then whole sections with it, but the # of frames of each are completely inconsistent. Not sure what's up with that one. I suppose it's video and maybe the DVD was poorly made?

6) yet another confusing one (sorry!): pics/s & fps are 29.97. DGindex shows NTSC, no field repeats, interlaced. But when I scroll with the arrows I see NO combing at all! Could it have anything to do with what kind of camera this was shot on? I'm pretty sure this was shot on one of the very high end professional digital cameras, although I don't remember what.

so in terms of exporting:

1) is clearly video at 29.97fps, so should be exported at 29.97 and I need to learn how to de-interlace it, yes?

2) is also video at 29.97fps, so I should export at 29.97, but since it is already progressive, I don't need to worry about de-interlacing, correct?

3) is a soft pulldown, so I should export at 23.976fps and don't need to worry about anything else in terms of interlacing.

4) I'm confused about what this one is and how I should export it.

5) seems like it's interlaced video at 29.97fps, so export as such, and it needs to be de-interlaced?

6) all the elements indicate it's video, but it looks like progressive film cause I see no combing. Not sure how I should export this one.

I was hoping to find one that showed the hard pulldown indicators that you mentioned, but I'm not sure that I did, other than possibly #4. I'll keep looking tomorrow.

so my remaining questions, besides how to deal with the confusing videos above, are at this point:
- when it's clearly video and I'm exporting at 29.97fps, it seems I need to be sure and check the "deinterlace video" box in Mpeg streamclip, correct? There are also choices for "interlaced scaling" and "reinterlace chroma". should I check or uncheck those boxes?
- when exporting video at 29.97fps and I've deinterlaced it, will I then be able to edit it onto other material that is 23.976 fps and be able to save it at 23.976fps? If not, how should I deal with this?
- how would I export videos that are hard pulldown?

the plot thickens!

thanks so much for your patience and all your help!!!
Reply With Quote
  #10  
07-28-2012, 01:57 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
I'm still reading but I just have to say, brilliant! You learn very quickly and I am impressed. You've made the correct deductions and also learned some of the finer points along the way.
You must be an amazing actor/actress, no wonder you've landed so many gigs
I believe there are shared personality traits in successful people, no matter what they're field.
Reply With Quote
  #11  
07-28-2012, 05:03 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Your analysis of the clips was very helpful. Their variety astounds me. Some will need expert attention. Don't worry about further analysis. We need a new approach.

I'm surprised that you didn't see my pattern in #2. I'll have to try myself. There's other pulldown patterns, but I didn't expect you to see them. They come from importing foreign material, old 8mm footage (18fps), variations of film to video, editing film as video, or unknown.

There's going to have to be some thought about workflow here. I can do it with the tools I'm familiar with, but I can't promise it's the easiest way.

Personally I'd ditch streamclip at this point. It works fine but, it's deinterlacing isn't going to handle the complex cases very well.

This is going to require Avisynth. It's a program that's popular in the community, including production professionals, for advanced video work. It's not an editing program at all. You write commands in a text file. Just to give you an idea, it's something like this:
Code:
#Comment: open the episode from the DVD
open("vts_01_1.vob")
#The episode is 29.97fps interlaced video, so deinterlace it
deinterlace
#Let's sharpen is a bit just to make it look nicer
sharpen(amount=2)
The program uses a neat trick that makes this text file work like a video file in windows; you can actually open it in Mediaplayer and play it! The above "script" (list of actions to take) would be suitable for your clip #1. All that would be left is to open it in an editing or encoder program (as if it were a real video), and save it back out as a format suitable for the web.

Your clip #2 would use this line instead of deinterlace:
Code:
#Recover original film frames from 29.97 telecined video
#The open command is set to ignore the repeat commands embedded in the vob, returning true progressive film
open("vts_01_1.vob", ignore_pulldown=true)
Are you planning to post your demo reel to YouTube? About how many clips do you need to put together?

Last edited by jmac698; 07-28-2012 at 05:58 AM.
Reply With Quote
  #12  
07-28-2012, 05:58 PM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
wow, thank you so much for the kind words! it made my day!

Yes, I am planning to put up my clips/demo on You Tube, and also a couple other sites that are specifically for actors that agents use to submit to casting directors. I already know that one accepts .mov, mp4, m4v, avi and I think a few other formats as well.

The Avisynth program seems interesting but also complicated.

OK, organizing my DVD's now to see how many I actually have. Some are definitely soft pulldown, so Mpeg streamclip should be able to handle those ok it seems. I already exported some of those as mpeg-4 using h.264 and they look fine.

Other than that I have some files that are .mov, some that are .mp4, an .avi (xvid), an mpeg-1, and some things I have to get off of VHS, but I'm not even thinking about those yet. A friend of mine has equipment to rip footage off of VHS, but I wanted to deal with these DVD's first.

Will respond again when I figure out how many I actually have that are NOT soft pulldown DVD's.

when I edit these together can I safely edit an .mov onto an .mp4 and onto the .avi and save it all as either an .mov or .mp4? in other words will I come into problems when trying to edit different file types together?

thanks!
Reply With Quote
  #13  
07-29-2012, 01:10 PM
kpmedia's Avatar
kpmedia kpmedia is offline
Site Staff | Web Hosting, Photo
 
Join Date: Feb 2004
Posts: 4,311
Thanked 374 Times in 341 Posts
jmac698 gave you a bunch of excellent advice, based on my quick skim/reading of the page.

We work with a number of filmmakers, studios, etc, and the rule of thumb is this:
  • Always go with the best progressive version available.
  • When you're operating from a DVD as the source, and it was film telecined to 29.97 broadcast, run IVTC to make it 23.975 again.
  • Do not simply de-interlace the 29.97 video when IVTC is possible.
  • When IVTC is not possible, and de-interlace is required, use a complex method, such as QTGMC in Avisynth.
Also do not overcompress the source video submitted to Youtube, as it will get worse. Match the Youtube specs as best as possible, and you may avoid re-encode altogether. Since DVD is the source, don't waste your time doing anything more than 480p source upload. Encode with a quality H.264 encoder, and submit the video in an MP4 wrapper.

Be careful using cheap VCRs on important tapes. Consumer VCRs love to eat tapes, whereas professional equipment is not so misbehaved.

- Did my advice help you? Then become a Premium Member and support this site.
- Please Like Us on Facebook | Follow Us on Twitter

- Need a good web host? Ask me for help! Get the shared, VPS, semi-dedicated, cloud, or reseller you need.
Reply With Quote
  #14  
07-29-2012, 04:21 PM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
You guys on here are officially awesome!

Thanks so much Kpmedia!

Let me step backwards for a second with a question. Does this whole issue of pulling 23.976 fps out of a DVD assume that the original material was shot on FILM, or shot on a format using 23.976fps? Does it matter how it was originally shot in terms of determining what can or can't be pulled out of it?

I know that a couple of the movies in question that I have on DVD that are clearly video (29.97pics/s, 29.97fps, NTSC, interlaced, & seem to be combed) were student projects and shot on DV (I don't think very high end). would this make a difference in terms of how I treat that footage when trying to export off the DVD?

Another question: would it be more efficient or effective to be doing this exporting using editing software? Or is it best to first export using these other methods mentioned (avisynth, Mpeg streamclip, or whatever else), then import the clips into editing software?

(I'm currently using the free version of videopad, but was looking to get something better.)

jmac698,
I've analyzed my DVD's and this is what I have:

a) 3 that are clearly soft pulldown (23.976 pics/s, 29.97 fps, has field repeats, video type is film, & progressive)

b) 1 that seems to be hard pulldown (29.97 pics/s, 29.97 fps, NTSC, interlaced, combing pattern is 1-1, 1-2, 1-1, 1-2 ...)

c) 4 that seem to be clearly video (29.97 pics/s, 29.97 fps, NTSC, interlaced, I'm seeing combing everywhere)

d) 2 that I'm not sure what they are: (29.97 pics/s, 29.97 fps, NTSC, interlaced, but I see NO combing anywhere at all!). I know both of these projects were shot on a higher end digital camera, in case that matters as per my question above.

e) 1 that I don't know how to classify: (29.97 pics/s, 29.97 fps, NTSC, progressive, of course no combing since it's progressive). This was also shot on some kind of digital camera. would this simply be video that is already de-interlaced?

f) 2 that are confusing: (29.97 pics/s, 29.97 fps, NTSC, interlaced, I see combing but I can't figure out the pattern, large areas with combing, then some areas seem to be without)

Which of the above can I export with Mpeg Streamclip (since it's easy), and which should I run IVTC on?
How do I run IVTC?
How would I determine if IVTC is NOT possible?

thanks so much for all the help!!!
Reply With Quote
  #15  
07-29-2012, 05:01 PM
kpmedia's Avatar
kpmedia kpmedia is offline
Site Staff | Web Hosting, Photo
 
Join Date: Feb 2004
Posts: 4,311
Thanked 374 Times in 341 Posts
Quote:
Does this whole issue of pulling 23.976 fps out of a DVD assume that the original material was shot on FILM
Yes.

Quote:
or shot on a format using 23.976fps?
That's more rare. You generally shoot either 24p or 29.97i for NTSC.

Quote:
were student projects and shot on DV
DV is 29.97i. Therefore deinterlace with an advanced method, and leave as 29.97 fps.

Quote:
would it be more efficient or effective to be doing this exporting using editing software?
Doing what? Not sure I follow you.
NLEs are really quite lousy at IVTC and deinterlacing. Use special tools before inputting into a standard editor.

Quote:
I'm currently using the free version of videopad, but was looking to get something better.
Adobe Premiere is excellent. Even the low-cost consumer Elements version has its uses.

It's available cheapest from Amazon: http://www.amazon.com/gp/product/B00...=digitalfaq-20
Or the download version for higher cost direct from the adobe.com store.

- Did my advice help you? Then become a Premium Member and support this site.
- Please Like Us on Facebook | Follow Us on Twitter

- Need a good web host? Ask me for help! Get the shared, VPS, semi-dedicated, cloud, or reseller you need.
Reply With Quote
  #16  
07-29-2012, 06:04 PM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
Thanks so much, that helps to clarify a lot.

Quote:
me: were student projects and shot on DV
kpmedia: DV is 29.97i. Therefore deinterlace with an advanced method, and leave as 29.97 fps.
By "advanced method" do you mean QTGMC in Avisynth? (although I have no idea what that is yet.)

Am I deducing correctly then that if something was shot in 29.97 (such as DV) you would NOT do ITVC since the purpose of IVTC is to pull the footage back to the 24fps or 24p that is was originally shot in?

Since I should export stuff shot on 29.97 at 29.97fps, when I go to edit that together with other footage that is at 24p will that cause me problems? I would think I'd want to save the whole project at 24fps (or 23.976?), so then what happens with the footage (now in the larger project) that was at 29.97? Will it look distorted in any way now that it's saved at 24fps (or 23.976)?

Quote:
me: would it be more efficient or effective to be doing this exporting using editing software?
kpmedia: Doing what? Not sure I follow you.
NLEs are really quite lousy at IVTC and deinterlacing. Use special tools before inputting into a standard editor.
What I meant by "this" was: exporting/ripping footage off DVD's for the purpose of re-editing my demo.
I'm not sure what NLE's are? I assume that means editing software?
So, you're saying it's best to first export using special tools then import the clips into editing software, yes?
What about footage that is soft pulldown, should I also first export those clips in something other than editing software?

thanks for the Adobe recommendation.

And sorry for my beginner level questions!

thanks so much!!
Reply With Quote
  #17  
07-29-2012, 11:52 PM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
- Instead of clicking Reply, click Quote - then erase stuff as needed

-Advanced method means QTGMC, a "script" for use in Avisynth.

-He seems to second my call for the use of Avisynth, because we are both concerned about quality. In this case I don't mean nitpicking details, but badly IVTC/Deinterlaced material will look obviously wrong. Ultimately I don't want this to detract from your presentation of yourself. And just to give an example of how QTGMC is used, if you take one of my "script" examples above, you would add QTGMC(preset="slow") at the end. One line, that's it!

-IVTC (InVerse TeleCine) is a process to return pulled-down/telecined video back to it's original framerate. Usually this means turning 29.97 with repeated fields back to 23.97 progressive with no repeated fields, and no interlacing. There's other possibilities; it can apply to recovering 18fps (silent 8mm film for example).

-Deinterlacing (which may use QTGMC as a tool) is meant to create an artificially progressive video from a truly interlaced video. An interlaced video may be truly interlaced, in which case it consists of two pictures taken at different moments in time, and woven into one frame by odd and even lines, or it maybe be progressive 29.97, marked as interlaced but really progressive, meaning the two fields come from the same point in time. (This fine point is a bit confusing).

Deinterlacing will guess what should come between the missing odd or even lines in the case of truly interlaced material. The result is that the 240 odd lines from one moment in time are expanded to 480 lines, same for the even lines. The input is 29.97 frames/s, each frame was interlaced 480 lines, which can also be called 59.94 fields/s, where each field is 240 lines, and the output is 59.94frames/s, progressive, and 480 lines per frame. Every other frame is usually thrown away to bring it to 29.97 progressive. You're missing finer steps of motion, but it doesn't look so bad, first because 30fps is good enough for smooth motion, second because the 60fps was changed by an even divisor of 2.

-Mixing 24p and 30p together is an issue, yes. It would be great if there were some way to put everything to 24p, but the 30p stuff won't look right. It's always better to go up than down; 24p presented at 30p is probably better than 30p reduced to 24p. There's a few ways to deal with this, but I'd like to hear other suggestions as well.

There are 3 basis for tools to change frame rate
-following the motion of objects in the image and guessing where they would be at intermediate times.
-blending two frames together into a blurry intermediate version.
-repeating frames of a video instead of fields; the added coarseness makes the motion less smooth (only for increasing framerate).

-NLE means editing software.

-I think you should export all your clips in Avisynth. Yes, this "interesting" and complicated way is really the best The output would be a lossless video file with full quality, in one common format no matter what the source, that can be simply imported into any editing program and used without any further complications.

In this workflow, you have the maximum possible quality with only one re-compression/encoding, and each clip pre-processing to it's fullest extent.

Last edited by jmac698; 07-30-2012 at 12:43 AM.
Reply With Quote
  #18  
07-30-2012, 12:55 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
*I disagree with KPMEDIA about available frame rates, there are 23.976, 24, and various others possible. He's probably correct that 24p is more common in practice.

Sony CineAlt F35 Camera Manual, p73, http://pro.sony.com/bbsc/as.....sion=original

ps admin
I really need more editing time, I'll start something then go finish reading an article, and find a correction, so it's cramping my style
Reply With Quote
  #19  
07-30-2012, 01:03 AM
cabney cabney is offline
Free Member
 
Join Date: Jul 2012
Location: Los Angeles
Posts: 11
Thanked 0 Times in 0 Posts
Thanks for all of that explanation.

Quote:
Originally Posted by jmac698 View Post
-Mixing 24p and 30p together is an issue, yes. It would be great if there were some way to put everything to 24p, but the 30p stuff won't look right. It's always better to go up than down; 24p presented at 30p is probably better than 30p reduced to 24p. There's a few ways to deal with this, but I'd like to hear other suggestions as well.

There are 3 basis for tools to change frame rate
-following the motion of objects in the image and guessing where they would be at intermediate times.
-blending two frames together into a blurry intermediate version.
-repeating frames of a video instead of fields; the added coarseness makes the motion less smooth (only for increasing framerate).
I think I should clarify what I meant about this. I wouldn't actually be editing clips together with 2 different frame rates into one clip/scene. I'd have a scene from film "A" that is 24p, then a quick fade out/in, a scene from film "B" that is 30p, then a quick fade out/in, then to a scene from film "C" that is 24p, fade out/in, etc. So 2 frame rates would never have to "blend" into one scene.

Given that scenario would I then be able to save the whole project at 24p? Or would the 30p scenes (saved as 24p) still look weird?

Quote:
Originally Posted by jmac698 View Post
-I think you should export all your clips in Avisynth. Yes, this "interesting" and complicated way is really the best The output would be a lossless video file with full quality, in one common format no matter what the source, that can be simply imported into any editing program and used without any further complications.
In this workflow, you have the maximum possible quality with only one re-compression/encoding, and each clip pre-processing to it's fullest extent.
I like the idea of maximum possible quality and only one re-compression/encoding, but the idea of writing scripts to tell it what to do is daunting. I'm not sure that I would know what to tell it. is it explained somewhere what to write for any given case?

thanks
Reply With Quote
  #20  
07-30-2012, 01:27 AM
jmac698 jmac698 is offline
Free Member
 
Join Date: Dec 2010
Posts: 387
Thanked 73 Times in 56 Posts
Thanks for all the thanks There's a way I can collect "brownie points", just click the thanks button at the bottom of my replies

I understand your confusion about mixing 24p and 30p. However, what I said still applies, because they are still stored in one YouTube file, which must have a single frame rate, so you have to decide the final framerate to use and adjust everything. If you were willing to prepare your presentation as two separate YouTube entries, then yes you could avoid some work here. It's possible to put YouTube videos in a playlist for part1 and part2, let's say part1 contained only 24p clips, at the end it would play the next part in the playlist, with all your 30p clips.

There is a video "container" called MKV which avoids this limitation, but it doesn't apply to YouTube, or as far as I know, any other web video.

Also bear in mind that YouTube limits your videos to 10 minutes, I don't know how long your clips will be, so for that venue it may have to be broken up anyhow. I was assuming in my mind that there would only be a few full scenes showing a range of emotion on your part, with you as the dominate active speaker/emoter, but what do I know about acting?

As for the scripting, we'll help you with that. Let's start with the easiest clips and fully prepare your first half dozen clips, completely ready for further editing.

I've been thinking ahead, and one of the clips which you've analyzed with the weird pulldown pattern - it would be a lot easier if someone were to develop the proper script for you, but that would mean releasing some sample of it. We respect your privacy, and you could put it on some filesharing site (like www.sendspace.com) and give out the link only in a private message. To further protect your privacy, you can submit a clip that *doesn't* have you it, but is representative of the problematic pulldown pattern.

Personally I don't watch TV and shockingly enough, have never seen ER, and I'm really quite focused on the project so I couldn't care less if you were Nicole Kidman. But, I would probably have a new appreciation for the art "knowing" someone in the field, and might try to catch a few episodes sometime.

Last edited by jmac698; 07-30-2012 at 01:48 AM.
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
Compression with MMC ATI, P frames and B frames for MPEG-2 ? John Capture, Record, Transfer 1 08-20-2011 10:15 PM
VirtualDub - frames inserted? lookouts Capture, Record, Transfer 3 10-15-2006 09:28 AM
Dropping frames - IRQ conflict or VCR? lookouts Capture, Record, Transfer 3 07-06-2006 06:15 PM
Need pgm to extract frames from avi mlaviolette Edit Video, Audio 3 04-30-2006 11:06 AM
ATI All in Wonder dropping frames allaboutduncan Capture, Record, Transfer 5 06-13-2005 03:10 AM

Thread Tools



 
All times are GMT -5. The time now is 03:40 AM