I've been using Evaer for Skype to record video calls. I realized (very late) that this program really sucks, and switched to Pamela, but in the meantime I recorded a number of video calls with Evaer that present several problems. The one I want to deal with first is images partially rotated. The video shows two images side by side. The one in the left is properly displayed, but the one in the right is rotated clockwise, as you can see in the snapshot below.
I don't have a clue of how to fix this problem. Is there a way to edit this video and make it look right (i.e. keeping the left part as is and rotating the right part 90 degrees)?
+ Reply to Thread
Results 1 to 27 of 27
You should be able to do it with any multi-track video editor. Basically you would make a split-screen project and put a copy of your video on two tracks. then rotate one and use a track mask to reveal both tracks. If you don't already have one, you can use Aviutl (free, very advanced) for that.Got my retirement plans all set. Looks like I only have to work another 5 years after I die........
I'm a newbie, so I'm not following you very well. Is there any documentation (ideally with step by step instructions) that I can use to do this?
Each editor does it in their own way, so documentation would vary for each one. You'll have to check help files or tutorials on split screen projects for the editor that you'll use.
If you post a small 10 sec clip, I'll whip up a video capture of how to do it in Aviutl. You can use something like AviDemux to cut a 10 sec clip out.
Last edited by racer-x; 11th Jul 2015 at 18:36.Got my retirement plans all set. Looks like I only have to work another 5 years after I die........
Uff, I had a quick look and this seems way too advanced for me. I could deal with it if there was enough documentation, but it's not the case. What other software can I use to do this? I don't even know the meaning of "multi-track video editor"...
BTW, I'm 99% sure that app didn't improperly encode/rotate it. Most likely the person was resting or just tilting their head when they were on camera (as you'll notice the normal horizontal rectangle around the face on the right). If true, you'll have to Zoom In a little to compensation for and crop out the black triangles that will appear with a non-perpendicular rotation of that material.
I guess that Evaer has a weak excuse here: Apparently the origin of the problem is that she's using her phone to Skype. It's likely that the rotation of the video comes from her phone. However, that excuse is very weak. If Skype can detect it and fix it, Evaer should be able to do the same.
It's pretty easy to do in the free aviutl. There's a bit of a learning curve for any software, but here is a quick tutorial. Another benefit of aviutl is it will import almost any format.
1) Drag your video onto the advanced editing timeline onto track 1. Click the video layer, and image file [standard drawing] box will pop up if it's not up already. That is where you make adjustments.
2-4) Push the "+" button in that box in the top right corner (that's where you add filters), and select "clipping". Adjust the crop values so only the right half shows. The dotted line shows the border in the preview and move as you adjust the top/bottom/left/right values. There is a reason why you start with the right side - in aviutl, layers on the bottom "cover up" layers on the top. If you're used to photoshop and layers, it's the opposite
5) Next, Select the clip on the timeline, right click copy, right click , paste onto layer 2. If you miss aligning them, just left click and drag it over. This copies everything, including the "clipping" (cropping) effect.
6) If it helps for the next step, you can temporarily toggle the visibility of layer 1 by hitting the layer switch (it will grey out)
7) So with the layer 2 selected, adjust the crop value so only the left shows.
8) Enable visibility of layer 1 again and select layer 1 clip (it will show dotted line in the timeline when you have a clip selected)
9) adjust the rotation parameter of clip 1, and x, y position (easier to drag it with the mouse in the preview window). You can adjust the zoom value if you want to scale up or down, to cut off the corners as you desire
I realize the OP probably isn't going to want to learn AviSynth. But just to show how easy it is:
(load video, then:) left = Crop(0,0,320,-0) right = Crop(320,0,320,-0).Rotate(-55) StackHorizontal(left, right)
I also lightened up the right half a bit to make it more visible, but for clarity did not include that in the script.
Last edited by jagabo; 11th Jul 2015 at 22:00.
Thank you so much, poisondeathray! Amazing tutorial. I managed to follow it with little difficulty, and I'm a real newbie.
I have a general question about the output: aviutl offers many formats and options to export the video. What format and options would you advise to have minimal quality loss and keep a reasonable file size? By reasonable size I mean up to three or four times the original size.
I attach a screenshot with information from GSpot so you know what kind of source I'm dealing with:
You can brighten up one or both sides together, or separately with different values (because you have them in 2 layers) in aviutl by adding the "color compensation" filter., just like you added the "clipping" filter. If you use window=>histogram, it will show you a RGB histgram that you can use to guide your manipulations, or you can just "eyeball" the preview. You probably at least want to adjust the "lightness" on the right side like jagabo did in avisynth
I would use the x264adv export GUI, with AAC audio . I think there are guides to getting it set up (I think either racer-x or maverick tse in the aviutl thread) . You can access it by using File=>Export with plugin => Adv. x264 Export GUI
You have to set up the executable paths for x264 and the AAC encoder in the GUI. I would use either QAAC or NeroAAC for the AAC Encoder. Everything else you can probably leave at default, except you might want to adjust the CRF factor . Lower values yield higher quality and larger filesize. Default is 23 which I think is a bit too high for general use (too low quality) . Or if you need a specific filesize you can use 1 or 2 pass bitrate mode. If you get lost on how to set up the x264adv GUI , then just ask either here or in the aviutl support thread
Actually, for the audio , you could just replace with the original audio by remuxing instead of encoding in aviutl , this way there is no audio quality loss. You can use a GUI like Yamb beta 2, or mymp4boxgui if you want MP4 container, or MKVToolnix if you want want MKV
To rotate at arbitrary angles you'll need to download and install the third part Rotate() filter.
Hi, I ditched this whole thing quite a while ago because things weren't really working with the lovely woman in the video. Now I want to give it another try. I'm following again the nice mini tutorial posted by poisondeathray but I'm having some issues:
When I open my video file I receive this error message: "[Warning]: Detected CTS duplication at frame 3". After that the program seems to work well, but I wonder if this has something to do with the next issue.
When I try to export the edited file the program starts to work, but after all the video rendering is complete I receive the message "File export failed". The program creates a file with the video, but with no audio.
The console displays this info:
Looks like there is a problem with the audio. When I was exporting I did select the paths to the video and audio encoders, and also the paths to the muxer, the remuxer and the timeline editor just in case. I left the rest as it was:
So, am I doing something wrong? Or maybe it's a problem with my video file?
Last edited by Alonshow; 1st Jun 2016 at 19:36. Reason: Adding info
@alonshow - the error message points to qaac (audio encoding) . Some versions of qaac either require quicktime with apple application support installed, or coreaudiotoolbox.dll plus some other files in a subfolder called qtfiles . (There is a workaround thread regarding qaac specifically on the doom9 forum , if you don't want "apple stuff" installed on your computer - basically it involves extracting a bunch of dll's from the apple installer and putting them in a subfolder). I would make a post in the aviutl support thread referencing this thread , as Maverick will know which version is being called for that aviutl version, and workaround
Or it might be easier to use another audio encoder like neroaac (but it usually requires you downloading the nero encoder separately), or fdk-aac. I don't recall offhand if Maverick bundled any of them with aviutl. Or even lame MP3 (which won't require any other steps). In terms of AAC, Apple is generally considered the "best" by most people, but at high bitrates, it's difficult to tell the difference between the top AAC encoders
Or if you weren't making any time edits (in terms of cutting or adding footage, ie changing the length of the video) , it might be better to remux the original audio into the new video , that way you don't lose any audio quality
Last edited by poisondeathray; 2nd Jun 2016 at 10:13.
I don't think I'm going to do any time edits, so I tried remuxing and it did work (with AC3, other formats were problematic). However, now I have realized that the video that Aviutil created is too long. The lenght of the original clip is 3:24, but the length of the processed clip is 4:07.
Any ideas of what might be the reason?
Edit: I think I've found the origin of the problem. For some reason Aviutl seems to convert the frame rate automatically to 25 fps. The original is 30.215 fps. When I open the video with Aviutl it displays the window "create a new project" and the value for frame rate is the same as in the original, 30.215 fps. However, the title bar says "640x480_25fps_16000Hz". And when I export, the exporting window shows a summary of the properties of the generated file which also says 25 fps. How can I fix this?
I attach the original file in case it helps. I removed the last minute (with Avidemux) because it contains personal information:
Last edited by Alonshow; 2nd Jun 2016 at 19:50.
What does mediainfo (view=>text) say about the original file ? gspot isn't as good for information on MP4 files
Set the project to match the original framerate. If you drop & drag it to the adv. editing timeline, it should show a popup where you can choose to match settings of the source file. Or you can manually match the characteristics by creating a blank project and entering the values (right click on the adv. editing timeline => new project. Then enter the dimensions and framerate)
EDIT: ok I see the sample you uploaded show "variable frame rate" . These can be problematic to edit if the full , original video was VFR, not just the cut sample (some cutting tools might cause VFR , at least in mediainfo reports)
The short explanation is timecodes control the display time per frame. Variable means this value changes instead of being a constant value. You can think of it as the video speeding up/ slowing down in sections. If you have not made any frame edits in terms of duration (no cutting or adding of footage) , ie. the frame count is the same as the original - you would need to extract the original timecodes then mux in those in the new container to recreate the same VFR timing as the original. For mp4 sources, I typically use mp4fpsmod. It's a commandline application
Alternatively you can try to convert to CFR. There is an option in newer Aviutl versions to do this in the l-smash import options, but I'm not sure how well it work. Maverick knows more about this option and what it does
Last edited by poisondeathray; 2nd Jun 2016 at 20:19.
I didn't notice thread until someone cross-post into the AviUtl support thread...
(pls PM me or post into the AviUtl support thread if I'm called for... as I'm only watching the support thread.)
The AviUtl Installer SHOULD already installed the Apple components. So the error is unlikely due to the Apple stuff unless there was an error during install.
I can't read Spanish(?) so please translate the red error message(last line) for me.
CTS error suggest file corruption or non-standard-conforming data.(common for TV/realtime recordings)
Try swapping the input plugin order for "L-Smash Works File Reader" and "DirectShow File Reader"
if that doesn't solve the problem, re-encode the source files once with FFmpeg, Mediacoder, Handbrake,Megui or anything of your choice, play them once to see if they are fine, than feed into AviUtl again.
CTS error if left alone may lead to serious AV-sync problem
// Alternative get around
since the error concerns audio, you may also use the "Export WAV" function to try extracting ONLY the audio as *.wav;
than open another project, tuck in the corrected video, the wav, adjust the sync, than encode again.Stopping development until someone save me from poverty or get me out of Hong Kong...
Thank you for your support, guys. The last line in Spanish says:
Error: CoreAudioToolBox.dll: Can't find the specified module.
Of course, this can be said in many different ways in English, so it is possible that my translation is not exactly the same as the standard message in English.
I'm thinking that this might add some significant information to the problem, so I'll wait until you check it before moving on. Please confirm that you still recommend doing the same after reading this or let me know if it changes something.
if the translation is correct, there is something wrong with Apple's software during installation.
Try installing iTune manually and then encode again. If that doesn't work, please reply here as that would mean I need to update the installer.
If that really doesn't work, go grab NeroAAC (http://www.free-codecs.com/download/nero_aac_codec.htm). Then in the encoder GUI, the right side, change the audio encoder to neroAAC and set the right path.Stopping development until someone save me from poverty or get me out of Hong Kong...
Everything works now. Thank you!
The console error disappeared after I installed iTunes manually and the resulting file has the audio now.
I'm not 100% sure of how I fixed the CTS error, but I think it was what you suggested: swapping the input plugin order for "L-Smash Works File Reader" and "DirectShow File Reader". This is my current plugin order:
So everything is fixed now! There are a couple of things that I might want to improve, though. The quality of the original files is low, but I still would like to lose as little quality as possible in the editing process. In my first attempts I used the default quality (Q=23), but now I have improved it a bit to Q=15. I think the quality is noticeably better now. The resulting file is much bigger, but that's OK because it's still only slightly bigger than the original file. For some reason, the size with the default quality was just a half of the original. So I'd like to know if adjusting the value of Q is the best way of improving the quality, and if so, how to choose the most appropriate value.
Also, I have about 20 files to process. I imagine the best way of dealing with them is just manually editing them one by one, which will be tedious. However, @jagabo said something about using AviSynth, which sounds like processing the files automatically in some kind of batch process. So, do you think it might be worth trying that, or would it be too much hassle for a relatively small number of files?
"Garbage In, Garbage Out", so changing the Q-value would at best not to further degrade the image. Q19 should be enough, Q17 is almost loseless. You may also install the UtVideo/Lagarith codec and export as compressed AVI (loseless).
Whether you should use AviSynth/VaporSynth or FFmpeg on the command line depends on the nature of editing. If you are applying some changes to the whole frame or a fixed part of the frame for all the files, the Synth tools and FFmpeg may be more convenient.
If the modification is more elaborate, not on fixed location, non-constant parameters, or simply no analogous plugin/feature in Synth, then you may have no choice but to use NLE to process them one-by-one.
20 files is still manageable by hand, but if all your videos are of the same resolution and have the same area rotated, you'd better ask for help using Synth.Stopping development until someone save me from poverty or get me out of Hong Kong...
Actually, some of the videos are a bit different, so I think I'll just process them manually. It will probably be much quicker than learning how to use a script. Thank you for the advice anyway
I just installed UT Video and it is displayed in AviUtl in the list of avaiable codecs. But I don't know how to use it. There are several UT video options in the list of codecs.
If you are using the NLE:
File> Plugin Export> Adv.Edit AVI/BMP > choose any one Ut format > OK
If you are not using the NLE:
Just Ctrl+S, then choose any one Ut format
All Ut formats are loseless and does not really matter unless you want to keep transparency info in the video
I'd suggest the following format if in doubt:
UtVideo YUV420 BT601 (for SD material)
UtVideo YUV420 BT709 (for HD material)
UtVideo RGBA (for transparency)
Last edited by MaverickTse; 15th Jun 2016 at 18:33.
The size of the video generated by the lossless codec is too big, about 20-30 times bigger than the original. I'm using the x264 codec with Q=17, which keeps the size roughly the same as the original and has good quality.
Thank you all for the help, it's been so useful. Unless something unexpected comes up, I think I've got everything I need