View Profile
View Forum Posts
Private Message
You can specify frame ranges (not time ranges) using the RemapFramesSimple filter, part of RemapFrames. It's one of stickboy's filters and can be found here:
http://avisynth.org/stickboy/
Also, if you use the full NoLogoM, you can lighten or darken the delogoed part with the final parameter. For example:
NoLogoM(A,B,50494,2,72,44,168,28,1.0,240,8)
That final 8 lightens it up a bit.
And if all you want to do is blur the result a bit, here's one way:
A=Last
B=A.Crop(52,36,-488,-412)#Crop to the location of the logo
B=B.MedianBlur(1)#or your favorite blur filter
Overlay(A,B,X=52,Y=36)Overlay the blurred logo onto the main video
+ Reply to Thread
Results 61 to 90 of 105
-
Last edited by manono; 18th Feb 2010 at 11:29.
-
Originally Posted by spiritgumm
Use ReplaceFramesSimple from http://avisynth.org/stickboy/
in the RemapFrames package.
This will apply Xlogo on frames 7-10, 20-32 and 41-50.Code:A=MPEG2Source("C:\RedLight\VTS_01_1.d2v", cpu=0).ConvertToYUY2() B=ImageSource("C:\RedLight\VTS_01_1 copy.bmp").ConvertToYUY2() NoLogoAuto(a,b,30) filtered = XLogo("LogoFinal1_x_32_y_24_2.bmp", X=32, Y=24, alpha=10) ReplaceFramesSimple(last, filtered, mappings= \ "[7 10] [20 32] [41 50]")
You can add as many ranges as you like.
I skim through the video in AvsP and mark the transitions with bookmarks, Control-B.
Then use a macro to export the bookmarks to a text file, format them and paste into my script. -
Just noticed this earlier message.
Not sure what you mean.Originally Posted by spiritgumm
But there is certainly no need for that long chain of conversions
Ifgives you sound in sync, then use that.Code:AssumeFPS(23.976, true)
If it goes out of sync, use "EnsureVBRMP3sync()"
This makes moving through the file cumbersome, but keeps the sound in sync.
To make DVD compatible AC3, one way:
Using Soundout from http://avisynth.org/mediawiki/SoundOut
-- make sure the Soundout is AFTER any Trims or other operations that affect the play order. Simplest to put it at the end.Code:...source... EnsureVBRMP3sync() AssumeFPS(23.976, true) SSRC(48000,false) Soundout()
This will pop up a GUI and let you export to AC3 when you open the AVS in VDub or any app.
Comment it out after you've made your AC3.
You can Add AmplifydB() (BEFORE Soundout) to boost or reduce the volume.Last edited by AlanHK; 18th Feb 2010 at 13:04.
-
Hi AlanHK - inre to the audio, I'm encoding with HCenc which doesn't accept audio streams, so I wasn't sure where to send it via Avisynth. That's why I used GUI ware, which had some issues when making framerate and format changes.
Based on what you wrote, I thought perhaps I could output altered audio framerate from Avisynth thru Vdubmod, but the script wouldn't run - got an error that there wasn't an audio clip.
Haven't had a change to digest the frame-range and logo advice you and manono posted, but looks promising! I expected to get thrown a bone but got steak instead.
Last edited by spiritgumm; 18th Feb 2010 at 18:16.
-
What is your complete AVS script? Do you perhaps turn off audio in the source command?Originally Posted by spiritgumm
Yes, HCenc only does video. But you can use the same AVS script to create audio, even simultaneously if you use Soundout, but preferable to do it separately (using VDub, say) to open the script, and then #comment it out.
I usually stream the audio from the AVS using wavi and send it to Aften with a batch file:
Which is good for doing things en masse, but for the one movie you're slaving over, Soundout is easier to set up. -
Inre to audio, initially I was using Directshowsource to input the whole movie (mpeg2/ac3). When I inserted video via d2v, forgot to add the ac3 plugin (didn't need it with the avi's). Getting audio now, which is ac3 48 khz. Get an error with SSRC(48000, false) - "could not resample between two sample rates." It runs with ResampleAudio(48000). If SSRC is preferable, how to implement? I did research it but didn't find what I'm doing wrong.
Wondering why I need to mention the samplerate if it isn't being changed?
I usually preview in AvsP.
Here's my script to test for AC3 output:
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\DGDecode.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\NicAudio.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\SoundOut.dll")
LoadCplugin("C:\Program Files\AviSynth 2.5\plugins\yadif.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\mt_masktools-26.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\MaskTools-v1.5.8\MaskTools.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TDeint\TDeint.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\xlogo_25_dll_20030630\xlogo.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\LoadPluginEx.dll")
video=MPEG2Source("C:\DVDs\RedLight\VTS_01_1.d2v", cpu=0)
audio=NicAC3Source("C:\DVDs\RedLight\Redlight delay -860ms.ac3")
AudioDub(video, audio).DelayAudio(-0.860)
EnsureVBRMP3sync()
AssumeFPS(23.976,true)
SSRC(48000, false)
Soundout()Last edited by spiritgumm; 19th Feb 2010 at 12:25.
-
For one thing, you don't need any LoadPlugins if the dll is in the plugins directory.
You'll get an error message if it can't find anything.
As for sample rate, comment out the SSRC and see what the sample rate is without it. Changing the fps should have changed it.
If you need to resample, SSRC is preferable to ResampleAudio, but doesn't work for all ratios. So try it first, if it fails then use ResampleAudio.
AC3 shouldn't need EnsureVBRMP3sync(), as AC3 is CBR, so can omit that too
So just:
Should work and Soundout() will tell you the sample rate.Code:video=MPEG2Source("C:\DVDs\RedLight\VTS_01_1.d2v", cpu=0) audio=NicAC3Source("C:\DVDs\RedLight\Redlight delay -860ms.ac3") AudioDub(video, audio).DelayAudio(-0.860) AssumeFPS(23.976,true) #SSRC(48000, false) Soundout()
Only need to change it if it's not 48000
What fps is the mpeg? -
It's a PAL movie - 25 fps, and am converting to 23.9796 fps. The popup gui reports the sample rate becomes 46034 after the AssumeFPS(23.976), and as you say, needs to be resampled back to 48000. Added the comment line but when I start to save the ac3, I still get the initialize encoder error.
So if I use ResampleAudio, or SSRC with AssumeFPS filter deleted, Soundout works.Last edited by spiritgumm; 19th Feb 2010 at 16:33.
-
What is your COMPLETE script?Originally Posted by spiritgumm
If it includes Xlogo, as I found a while ago, loading it kills the audio format that SSRC uses. There is a workaround, but Resample Audio is okay.
And this is a script fragment that I picked up and modified to convert PAL to 23.976 fps, converting the audio sample rate too:
Comes to the same thing as what you have.Code:#Convert to NTSC film, 48000 Hz Nfr_num = 24000 Nfr_den = 1001 Ar = 48000 ResampleAudio(Ar*FramerateNumerator()*Nfr_den, FramerateDenominator()*Nfr_num) AssumeSampleRate(Ar) AssumeFPS(Nfr_num, Nfr_den, False)
You'd put Soundout() after that to export the audio.
Also, if the sound it staticky you might want to use some noise reduction in Audacity to try to clean it up. You could do all the rate conversions in Audacity too, but might as well use Soundout to export as wave and clean it up, then use Audacity's AC3 export. -
If XLogo was causing an issue, I don't mind doing the audio and video separately. I got the script working with SSRC by deleting "True" from AssumeFPS. Would that effect the new audio's sync?
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\DGDecode.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\LoadPluginEx.dll")
video=MPEG2Source("C:\DVDs\RedLight\VTS_01_1.d2v", cpu=0)
audio=NicAC3Source("C:\DVDs\RedLight\Redlight delay -860ms.ac3")
AudioDub(video, audio).DelayAudio(-0.860)
TFM()
AssumeFPS(23.976)
#SSRC(48000, false)
Soundout() -
Originally Posted by spiritgumm
See
So you need this to keep sync.FPS
If sync_audio (which is false by default) is true, it also changes the audio sample rate to match the duration of the video, the pitch of the resulting audio gets shifted.
But that obviously is not the complete script, you don't have the filters. Though you still have the LoadPlugins which you don't need, even if you did have the filters in the script.
You can't work out the problems and conflicts with a fragment.
Please post the complete script. -
That's the script - I was going to process the audio separately, without logo filters running. I left TFM() in case it had any effect on FPS. I thought I needed some of those plugins for the video to be recognized/processed.
-
You do need the plugins. You don't need to load them in your script, because all plugins in that folder are automatically loaded.Originally Posted by spiritgumm
If anything is missing, you will get an error message telling you it can't find XXX.
Anyway, if you are using two script files it's easy to get them out of sync.
So I try to use the same script for both. But if necessary I turn off some parts with a # comment temporarily.
So ... would still like to see the complete script -- or both scripts if you are doing it like that.
Anyway, it's late here and I'll be offline for at least 12 hours.
Good luck. -
As AlanHK says, SSRC only works with certain resampling ratios, and 46034->48000 is not one of those.Originally Posted by spiritgumm
So you can only use ResampleAudio here. The first resampling to 46034 via AssumeFPS is necessary to preserve audio sync. -
I dont have a working script for the video part of this movie (logo removal yet) as I was focusing on this audio conversion. Makes sense to do audio-video conversion together to keep sync, tho I've read somewhere there are advantages to doing them separately. Would sound quality be better by Resampling versus the multiple conversions I did with GUI ware?
A side issue - I can use Xlogo on this mpeg movie, but can't run it with any AVI - get access violation error. Can't figure out what the problem is. Tested the bitmap in Vdubmod (with it's XLogo filter) and it worked. I reinstalled Avisynth but didn't help.Last edited by spiritgumm; 20th Feb 2010 at 19:21.
-
Every conversion of sample rate or into AC3 or MP3 will do the audio equivalent of blurring the sound a little. So you minimise these, try to do things in one shot or use a lossless format like WAVE if you need an intermediate.Originally Posted by spiritgumm
Is the frame size of the bitmap the same as the video?Originally Posted by spiritgumm
-
edited since posting:
Hi AlanHK, just to clarify, we're talking about a Pal AVI, not the Pal format movie which XLogo did work for (which I still haven't done yet).
The bitmap is the smaller area containing the logo (which is now whitened) surrounded by a black border. Speaking of which, I had a question about your XLogo guide - it says a minimum of 4 pixels of black, but when I use Photoshop, I have to use 40 pixels.
Thanks for the audio info, and I have some more audio questions. Since I couldn't get XLogo to run in Avisynth with the AVI, I only processed it with Nologo. Changed the audio FPS with Soundout alone, and while encoding the video. Haven't hard muxed them, but previewing the new A/V loaded in TDA, the Soundout alone audio was better synced than the audio processed during encoding. Might be TDA isn't previewing correctly, but the audio files are the same size and length, so I dont know why this should be.
Also, is it possible that some PAL versions of American movies retain the original audio pitch? This AVI was in PAL progressive (25 fps) with a European TV logo. Changing the FPS restored it's original USA running time, but the opening music is now too slow (I have the music CD soundtrack). In this instance, I'm thinking the audio framerate needs to be changed WITHOUT changing the pitch. Omitting the "true" in AssumeFPS would only mess up the sync and isn't a solution, right? Seems like a GUI could do this (as it seemed to happen in Belight once), after stripping out the AVI audio with Vdubmod first.
Later: to check for consistency with the CD, I previewed the end of the movie with opposite results - in that section, the slowed-down audio is truer to CD than PAL audio. What would explain this change in pitch throughout the movie? Rather than the PAL speed-up, perhaps it was incorrectly recorded on the film itself, or broadcast slightly off at the start? This is assuming the CD is the correct standard of comparison.Last edited by spiritgumm; 21st Feb 2010 at 11:27.
-
4 or so pixels black should be fine. Sometimes the whole bitmap is only 40 pixels across.Originally Posted by spiritgumm
Most likely your black is not really black. In Photoshop select the black area (eg, using magic wand) and use "Fill" and then select "Black". You may have used a foreground colour that looks like black, but has other colours mixed in. Make sure it is 100% black. You can also change the mode to Greyscale.
What does "audio processed during encoding" mean? Processed how?Originally Posted by spiritgumm
"Audio pitch" is not "speed". What do you mean?Originally Posted by spiritgumm
I think you can change pitch in Avisynth, using SOX filters, but I've never tried those, so can't do more than point you to them.
Otherwise you can adjust the pitch in an audio editor, like Audacity, without changing the length, if you want.
-- Use Soundout, export WAVE, in Audacity Effect/Change pitch. To keep the same pitch as 25 fps, use increase by 4.27% (25/23.976 -1). Then export as AC3 from Audacity.
It's possible they did some audio editing, not just a simple conversion. The credits especially are often different between different movie releases.Originally Posted by spiritgumm
-
Inre to the black pixels, I did fill in with 100% black.
For my mpeg Pal movie, manono made a sample de-logo bitmap (on first page, I think it's LogoFinal_x_32_y_24_2.bmp). His has a 30-40 pixel black border. I experimented trimming the border a couple times, but it wouldn't work unless it had the 30 or 40 pixels.
I used same size border for my AVI, but it still wouldn't work. I'll revisit the project again.
Inre to audio processing, I thought you posted earlier to run the whole shebang - have Soundout process the audio while video was sent to the encoder. When I ran script in HCencoder, Soundout did start, so I did both audio and video. Earlier, I'd done the audio alone when I simply opened the script in AvsP. For whatever reason, that audio seemed better synced.
I previewed thru this AVI, comparing it with the new (slower pitched) audio. New audio seems fine; just the beginning is too slow, so it's not a big issue. Must have been that particular copy of the movie as aired.
To clarify what I meant in previous post, when the audio sounded too slow, I meant the pitch. I should have remembered about Audacity - I previously tried it to change movie audio by 4% and it didn't change the pitch, and I moved on to BeSweet and forgot about it. Also Belight can do a framerate change in WAV while leaving pitch unaltered (if I ever need it). -
Well, I never figured out why I couldn't run Xlogo with an AVI, so I finished de-logoing my movie. The filtering came out fine, but it has an interlaced look (not like previous sample results manono and I have done), so I must have done something wrong. Here's my script (for brevity, I omitted the huge list of ReplaceFramesSimple pairs):
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\logotools\LogoTools.dll")
MPEG2Source("C:\DVDs\RedLight\VTS_01_1.d2v", cpu=0)
TFM()
A=MPEG2Source("C:\DVDs\RedLight\VTS_01_1.d2v", cpu=0).ConvertToYUY2()
B=ImageSource("C:\Documents and Settings\Dave\Desktop\RedLight\VTS_01_1 copy.bmp").ConvertToYUY2()
NoLogoAuto(a,b,30)
filtered = XLogo("LogoFinal_x_32_y_24_2.bmp", X=32, Y=26, alpha=40)
ReplaceFramesSimple(last, filtered, mappings= \
"[231 392][2759 3090][110180 110186]")
A=Trim(0,54)
B=Trim(55,145)
C=Trim(146,0)
D=B.Crop(600,60,0,-444)
E=Overlay(B,D,X=484,Y=50)
A+E+C
LanczosResize(720,480)
AssumeFPS(23.976) -
Interlaced look? And this one has shifted or out-of-phase fields you were realigning with TFM? If so, TFM was ignored after you did that A=, B= stuff right afterwards. If it was interlaced to begin with, it'll be interlaced afterwards. You might move the TFM down below all the XLogo stuff and try again. XLogo works better on the unfiltered source anyway. And test your scripts in VDub(Mod) first before wasting all the time encoding. Maybe you did.
-
I'm using AvsP to preview my scripts in action. I was hoping my script had a more obvious error, as I too thought about moving the TFM line. Anyway, moving it seemed to work, except then the mask would flash when a scene went from bright to dark. I manually changed all my range points to not include any frame in that situation. Afterwards, occasionally the logo was visible for a fraction of a second, so I experimented having TFM included twice - it it's original spot and at the end, and that fixed the issue (tho I had to manually change all my ranges back). Burned a dvd and noticed the audio was slightly off (didn't notice it on the PC)....Fixed that, then noticed sometimes the logo was more visible (more "on top" of the video) in certain grey/black scenes - very difficult to determine on a monitor, but at that point, I decided it was good enough.
So, thank you all for the help!
Inre to not being able to use XLogo with an AVI, looks like it was related to running in AvsP. It worked when I played the script in Media player. I didn't try encoding it, but I'll assume for the moment it would have worked, and I have AvsP incorrectly set up.
Have another downloaded AVI (originally PAL dvd rip) problem, if anyone's interested in helping diagnosing. Progressive 25fps, has abit of blurring - I thought it was blended video, but using TFM() didn't seem to fix it. Is it just the way the original dvd was made? Here's a sample (which I clipped with Virtualdub):
crycity.avi -
Too late for that one. It looks like it was converted from NTSC, but then it was deinterlaced, so it can't be unblended.
Last edited by manono; 7th Mar 2010 at 18:27.
-
oh, well. Thanks for taking a look. Luckily it's still watchable, and the audio was fine. I've seen two differently compressed uploads with the blending so I'm thinking it was the dvd itself. It's an old American movie but was never released in USA; there are 4 different PAL releases. I'm familiar with the actors, and for some reason the audio pitch sounded correct (was too slow when I changed the framerate to 23.976), so I used Belight to change the audio framerate (since it doesn't change pitch when doing framerate change).
-
Possible, but not likely, in my opinion.Originally Posted by spiritgumm
If converted from NTSC by field blending (or even field-blending followed by deinterlacing, as in your example), the length remains the same and the audio hasn't gotten the 'PAL speedup'. You could check IMDB for the real film length and compare it to your PAL version. IMDB sometimes makes mistakes, but if it's correct in this instance it might be a good way to prove to yourself that the PAL version is the same length as the NTSC version. If true, it also means that if you slowed the film to 23.976fps, you made a mistake in doing so and you should have kept it at 25fps....for some reason the audio pitch sounded correct -
The first thing I do with any movie (Pal or Ntsc) is check the official runtime (though IMDB isn't always accurate); the AVI was shorter by the usual 4%.
I checked out a dvd reviewer website which had an extensive comparison of all 4 Pal dvds (apparently the guy is a big fan of the movie). He indicated they were all sped up, as well as the "[spanish] release is from an unconverted NTSC source (verify with times above) and hence there is significant 'combing' (or 'trailing') problems (see examples below)."
I saw one frame at the end of the AVI that was in spanish (the authoring company), so perhaps this is that version. He indicates it's sped up (shorter runtime, in Pal), so I'm not sure what he means by "unconverted NTSC source." -
Have another weird AVI problem - jumpy video, possibly someone deinterlaced it even though it wasn't interlaced (it was originally 29.970 fps). Dont know if I should start a new post or what.
Also couple questions: is there a way to scan an AVI for corruption problems before encoding?
Does an mpeg or avi that is 29.970 fps, and appears to have the 4th frame repeated, require any special processing?Last edited by spiritgumm; 15th Mar 2010 at 10:49.
-
One of the reasons to field blend (if there is any good reason) is to keep the movie the same length and keep the original unspeeded up audio. If the movie is 4% shorter in length, then almost by definition it hasn't been field blended but just speeded up. You don't get both at the same time.
He means that they used an NTSC master for the PAL DVD and the PAL movie on DVD is the same length as is the NTSC/film version, so he's contradicting himself because the 3 with obvious PAL speedups are made from PAL masters. Your earlier AVI sample of Cry Of The City was probably made from that Suevia DVD and may have been converted using an unusual method not involving field blending. They added a frame every second in the form of a field every half a second. On the DVD it'll look like half a second of 'clean' video followed by half a second of interlaced video. When blend deinterlaced by people that don't know any better you get half a second of 'clean' video followed by half a second of blended video. It looks like the blending/clean/blending/clean switches every 12-13 frames. If you had the DVD itself, you could 'IVTC' it and easily pull out the duplicate frames and return it to 24fps. That would explain your audio sounding 'normal' as this method of converting NTSC to PAL also leaves the movie the same length.so I'm not sure what he means by "unconverted NTSC source."
If so, my friend Gary Tooze is incorrect about all 4 having the PAL speedup, and at least one of the 4 examples (the Suevia) can be explained by not having the complete film on the DVD.
Other than removing the duplicate frames so it doesn't play jerky?Does an mpeg or avi that is 29.970 fps, and appears to have the 4th frame repeated, require any special processing?Last edited by manono; 15th Mar 2010 at 21:33.
-
VirtualDubOriginally Posted by spiritgumm
Video/Scan video stream for errors
Not a guarantee though, it can miss some errors.
Also DivFix and DivFix++ can scan and possibly repair AVI files.
I've found Avidemux good for fixing problem AVIs; just open and save with video and audio set to "copy" format.
A file that might crash in Avisynth can be made usable, but you might lose sync if there was a discontinuity. -
AlanHK - thanks for the scanning/fixing ware suggestions. I've noticed a couple times that Vdub would offer to fix some avi files, I think because of "improper audio," so I wondered if it had a video scan - thanks for specifying it.
I just loaded a couple avi's in Avedemux - it offered to fix the index on all of them - is that typical? Vdubmod didn't report this issue, nor did Avisynth.
Manono, I see you followed up on that noir (I figured you could with google the quote if you were curious). While I doubt the divx uploader fooled around with the audio, I could be wrong about the audio pitch. I would think The Beav would have noted any audio discrepancy between the 4 dvds in his review.
Of my first question in my recent post, I found a substitute AVI download which wasn't deinterlaced unnecessarily (or whatever it was), though it seems to have other issues which I'll ignore for now (pixelations in one spot).
For my other question about the repeatad 4th frame, I'm attaching two video samples - a dvd I own, and a divx version of a similar movie from the same company. I believe both are 29.97 fps progressive, so I believe encoding the divx back to mpeg should only require resizing, and no framerate change and no pulldown. Stepping thru Separatefields revealed what seems to be the 4th frame repeated in both, but they play smoothly enough. If I should correct the divx first, perhaps you could let me know?
dvd.m2v
divx.avi
And an additional question:
I have two versions of the same movie - one is 2 hour widescreen, the other 2hour 20 minute fullscreen. I'd like to keep the widescreen (which looks better) and add the various missing scenes to it from the fullscreen. Unfortunately, the widescreen was a divx I encoded already to fill a DVD5 (and I'd rather not have to download it again, and figure out the delogo filters again). What might be the best way to combine these without causing too much quality loss? Just add borders to the fullscreen and reencode the whole thing?
Quote
Similar Threads
-
HCenc and PanScan
By Comp360 in forum Video ConversionReplies: 2Last Post: 14th Mar 2011, 17:35 -
Problem with HCenc 0.24.
By HolyFries in forum Video ConversionReplies: 2Last Post: 17th Jun 2010, 18:03 -
Problem with HCenc 0.24.
By HolyFries in forum Newbie / General discussionsReplies: 0Last Post: 17th Jun 2010, 13:18 -
Frameserve from VirtualDubMod to TMPGEnc
By pchan in forum DVD RippingReplies: 1Last Post: 25th Aug 2007, 08:13 -
HCEnc
By Belfour in forum DVD RippingReplies: 1Last Post: 28th Jun 2007, 09:57