Yes, it can, but since you're making a DVD and since only 29.97fps is acceptable for NTSC DVDs, there's no point at all in doing that.I am also wondering if DGPulldown can do 18 -> 24 because there is an old post here, that describes doing just that:
Fine for PAL DVDs.So DGPulldown with custom 18 -> 25 will be fine.
+ Reply to Thread
Results 61 to 90 of 95
VirtualDub to open the .AVS file (with the AVISynth filters that we've been discussing), saving it to an uncompressed AVI, then I ran the AVI through DGPulldown which produced the .M2V file. But you are saying that I should be using the .M2V file directly in the authoring program. I guess I was under the impression that the .M2V file had to be encoded to DVD-compliant MPEG2 video.
Last edited by szilagyic; 5th Apr 2011 at 21:56.
24fps is legal , but telecine 3:2 makes it 29.97 frames per second (or 59.94 fields per second) for the DVD player. The content is still 24fps. The DVD player or display inverse telecines (reverses the process) so you "see" the orignal 24fps
"Soft" telecine means this is done by flags, e.g. using DGPulldown or similar . So this allows you to use progressive encoding instead of interlaced encoded (it's much more efficient, higher quality)
DGPulldown that I should be using in this case. But this is what I'm gathering so far based on manono's suggestions: first I apply the AVISynth filters and open the AVS with VirtualDub, and save the uncompressed AVI file @ 19.98 fps progressive (I'm actually running the VirtualDub "sharpen" filter at 50 which helps clarify the picture). Then I compress/encode the AVI file to MPEG2 video with Premiere @ 23.976 fps progressive, then run this MPEG2 file through DGPulldown set to custom 19.98 -> 29.97, and then use the resulting .M2V file from DGPulldown in the DVD authoring software.
Whew. Please correct anything above that is not right. And thanks again.
Last edited by szilagyic; 6th Apr 2011 at 12:48.
Last edited by szilagyic; 6th Apr 2011 at 13:42.
AviSynth has very good sharpening filters of its own. You don't need VDub for anything at all. You can use its built in Sharpen, or one of the third party sharpeners. I use LimitedSharpenFaster:
And even if you do use VDub to make an intermediate AVI, I hope you're making a lossless file (Lagarith, HuffYUV, etc), and not uncompressed. With all the filtering done in the script you can use the faster Fast Recompress rather than Full Processing.
Yes, you did it right. Make the script for 19.98fps (and some encoders will require that you actually make it for 23.976), encode for 23.976fps, apply DGPulldown with the 'Custom' box checked and set for 19.98->29.97.
Last edited by manono; 6th Apr 2011 at 14:30.
The Motion.dll? This should work to give you 2 new blended frames in every 20 :
This goes after you've already made the framerete 18fps.
Well I went ahead and processed some videos. I ended up encoding the AVI file (which runs at 19.98 fps) as .MPG video @ 23.976 fps progressive with TMPGEnc so that I could get separate video (.M2V) and audio (.MP2) streams. Next, I ran the video stream (.M2V) through DGPulldown with the custom 19.98 -> 29.97 setting. Then I muxed the resulting .M2V video with pulldown, and the original .MP2 audio back together with TMPGEnc MPEG Tools to get the final .MPG file.
But in the final .MPG file, I noticed that the audio is not in sync with the video. The video is actually playing slower than the audio, and as a result the audio gets ahead of the video and gets worse as the video plays. I also compared the original .M2V file with the .M2V file processed with DGPulldown, and they do play at different speeds (the one with the pulldown applied is definitely slower). So I must be missing something here. The settings I'm using in DGPulldown are:
Custom: 19.98 -> 29.97
Top field first: checked
Set timecodes: checked
Write timecode to text file: unchecked
Set drop frames: checked
Remaining settings all unchecked.
Thanks again in advance....
Posting the complete script you used to create the 19.98fps AVI would be helpful also.
DVD Styler), by muxing the M2V from DGPulldown with the MP2 audio. Now I guess I should ask, when I mux the M2V (with pulldown applied) + MP2 audio together with TMPGEnc, is that possibly stripping out the pulldown that DGPulldown applied? I didn't re-encode anything, just muxed them together for the DVD authoring program.
I processed a test clip to get some sample timings:
19.98 fps AVI file : length 3 min 19 sec
... encoded the AVI file to separate video & audio streams with TMPGEnc (M2V @ 23.976 fps / MP2)...
23.976 fps M2V : length 3 min 19 sec
MP2 audio : length 3 min 19 sec
... ran the M2V file through DGPulldown, custom 19.98 -> 29.97...
29.97 fps M2V (after pulldown applied) : length 3 min 19 sec
... muxed the M2V after pulldown + MP2 audio together with TMPGEnc MPEG Tools...
muxed MPG : length 3 min 59 sec
However still the same issue, when I play the MPG video in VLC as a test, the audio is OK right at the beginning, but starts to get ahead of the video the further the video plays. I also tried authoring a test DVD and tried playing it on two settop DVD players and they play awful, just like they do in VLC. I will try using Muxman, as I am starting to wonder if TMPGEnc's muxing tool is doing something strange.
The current AVISynth script I am using to produce the AVI files is:
LoadPlugin("c:\program files\avisynth 2.5\plugins\DGDecode.dll") LoadPlugin("c:\program files\avisynth 2.5\plugins\Decomb.dll") LoadPlugin("c:\program files\avisynth 2.5\plugins\Motion_12Dec05.dll") LoadPlugin("c:\program files\avisynth 2.5\plugins\Masktools.dll") video=MPEG2Source("T:\movie_project\test\avisynth.d2v") audio=WAVSource("T:\movie_project\test\avisynth_proc.wav") AudioDub(video,audio) #Decomb filter AssumeTFF() Telecide(guide=1) Decimate() AssumeFPS(18000, 1000, true) # Convert frame rate, also adjust audio. SSRC(48000) # Restore audio sample rate to a standard rate. BlendFPS(19.98,Aperture=0.2)
Since TMPGEnc can't do 19.98, put assumefps(23.976) at the end of the AVIsynth script (after what you already have).
(I could be very wrong - it's a long time since I've tried this!).
szilagyic: Stop right now and read this post before going any further....
About a decade ago, I used the exact same model GAF projector to transfer 24fps Super8 Sound films to video. I set the projector switch to 24fps, then slowed the variable speed knob down until the video flicker vanished. (The video camera shot the images through a mirrored telecine lens box. Projector light bulb was diffused with frosted glass.) The transfers came out perfect, except for the sound. I took the audio portion (a .wav file) and adjusted the pitch to sound like normal.
Here's the deal: A film shot at 24fps was slowed to roughly 20fps for the transfer to video. You know what? Not a single viewer (among dozens) noticed that the film speed was slightly slowed down. Because I had restored the pitch of the voices and music to normal, nothing seemed out of sorts for the viewers.
Many Hollywood movies contain scenes that were shot slightly off-speed, just to achieve a certain mood. Unless the speed change is dramatic, it usually goes without notice.
Please do a short test as follows:
Try setting your camera to 60i (NOT 24p cinema mode -- you have a film-look source to start with) and set your projector to the original speed of the film (either 18fps or 24fps), then adjust the variable speed knob until the flicker vanishes.
Fix the pitch (but NOT the time duration) of your audio until the voices sound normal. Audacity is a good tool for this.
I am not saying this will work for you. None of my films were shot 18fps. (In that case, you would need a slight speed-up.) But, please experiment with a short sample. You may get satisfactory results without killing yourself with all the AviSynth scripts and renderings.
See which method provides the most satisfactory results. If exact speed is not really a critical issue, you may find the process in this post far more efficient.
TMPGEnc so I have no idea what it does. In CCE you set the framerate in the encoder and it doesn't follow what the framerate of the script is. So, after you have your script made for 19.98, afterwards add:
2BDecided mentioned doing that. Either that or feed your AVI into TMPGEnc using an AVISource line in an AviSynth script, with that same AssumeFPS(23.976) just below so TMPGEnc doesn't go around behind your back changing things without telling you. The M2V fresh out of TMPGEnc should be shorter in length than the 19.98fps AVS or AVI.
AviSynth has a bit of a learning curve. He's almost there, and after he passes this last little bump in the road he'll be using AviSynth for all his encoding projects and wondering why he didn't know about this before. You do it your way and he and many of the rest of us will do it ours.
The resulting video looks excellent. Many thanks to manono and everybody else for the help with this. I'll be sure to post back any further updates but hopefully this thread also helps others doing the same thing. I've seen other posts about capturing 18 fps film but none so far that include sound. This process seems to be very good.
So at 17 fps, I am guessing that the aperture setting needs to be changed for BlendFPS, but I don't exactly understand how the aperture is calculated. What should it be set to now, 0.3?
I'm not at my encoding computer to figure it out for you. You set the aperture so that it (now) gives you 3 evenly distributed blended frames in every 20. Instead of 18fps going to 20fps (roughly) you're now going 17fps to 20fps. My guess is that the same aperture will work, but it's very easy to figure out yourself. If it's too low you get duplicated frames instead of the blended frames. If it's too high you'll get more than 3 blended frames in every 20. Pick a number (I'd start with the 0.2), open the script in VDub(Mod), scroll to a place with motion, and start looking and counting.
Just trying to help, no other motive.
Apparently they film elephants at ~40fps to make them look really cool when watched at 24fps/25fps.
But generally, for home movies, I'd want the frame rate to be correct. I wouldn't want to pitch-shift the audio either - just a straight resample to where it should be is best. Though I've never dealt with an 18fps movie with synched audio - I have 18fps home movies with audio, but it was laid down later (on magnetic tracks added after the film was developed), so it's pretty tolerant to messing around.
Here's a new version of FilmExtractor. A few additions and some bug fixes. I also included the full autoit3 script for your use.
FilmExtractor v1.1 - Used for extracting progressive frames from a camcorder capture of film projected at a slow speed using a projector without a shutter (aka, the poor mans telecine).
Last edited by Khaver; 15th Jan 2014 at 07:43.
Thanks Khaver - I'm going to have to try this - it looks very clever.
I'll need to mod a projector - I think it may be wise to replace the lamp too as projecting more slowly will increase the chance of damaging the film. One of my eumig projectors already has a special slow projection mode and it puts a screen in front of the lamp to prevent film damage (it reduces the brightness of the projected image).