I got tired of having Cinelerra hang while create table of contents files for a list MPEG videos. So I wrote my own script to create the table of contents files for a list of MPEGs in a directory. Once you specify the LIST, the script plops the table of contents files in your users' .bcast directory in ~/.bcast.
I output the file name, start time, end time and elapsed time for the table of contents creation. This output looks like this:
*** Update 2010/05/28 *** I added a little logic to the beginning of the script that checks for the presence of a single argument to the script. That argument is the name of a single file to convert. So now, you have two choices: 1) convert a list of mpeg files as specified in the LIST variable or 2) convert a single file via the command line like so: ./tocMaker.sh videoToConvert.m2t *** end update ***
Here is the script:
#!/bin/bash # # tocMaker.sh # Author: SCF # Website: http://crazedmuleproductions.blogspot.com # # Purpose: This script builds the table of contents files for a list of mpeg files # # Input: List of mpeg files as specified in the LIST variable or a single file via command line like so: # ./tocMaker.sh videoToConvert.m2t # # Output: Runs mpeg3toc and puts toc's in your users' .bcast directory # DEBUG= if [ $1 ] then LIST=$1 else LIST="*.m2t" fi echo "converting $LIST" TOTALTIME=0 function dt { DATE=$(date +"%Y%m%d_%H%M%S") echo $DATE } function st { SECONDS=$(date +%s) echo $SECONDS } PWD=$(pwd) DIR=$(echo $PWD | sed 's/\//_/g' | sed 's/^_//g')
As a follow up to this original post, I thought I'd give a few hints on working with the batch render function in Cinelerra. I've been using the batch render to prepare me for my new Fedora 12, x86-64 system. Specifically, I am using batch render to profile the capacity and speed of my current Fedora 10 system. After I install the new Fedora 12 system, I can then use the Fedora 10 performance baseline to tell me how much faster (or slower) the new system is in relation to the old system.
Selecting a Cinelerra Project
The first thing I did was use a short (~1 min) clip of a project that I had been working on:
I then used this project to output a short clip as the basis for the batch render:
Batch render is accessed by typing Shift-B within Cinelerra:
You can save a batch render list to XML format. This format is similar to the XML of the edit decision list (EDL) that Cinelerra stores when a project is saved. You can then load that XML to use later.
Batch Render Gotchas
I used a previous batch list to render out that short clip to about twenty different file formats. The batch render blew up a few times, so I had to get over a few obstacles:
1) the batch list I had saved months ago was out of date and the directory pointers in the XML were incorrect. I fixed the incorrect pointers by going into the vi text editor and doing a wholesale conversion with a sed construct: :1,$ s/videos\/oldpath/videos\/newpath/g
1,$ says look for and make replacements on all lines. s means "search for.." The matching expression looks for the string "videos/oldpath" in the file and replaces it with "videos/newpath". g means "do the replacement globally"
2) my project had an improperly formatted Quicktime video track on the timeline. I kept getting Quicktime errors when the batch ran, which would crash Cinelerra. Once I removed that errant track, the batch render worked correctly.
3) once I got the batch running, my disk would fill up quickly as I was rendering to a few uncompressed formats. I deleted some very large, extraneous files and the batch was able to complete.
So you can see that you need to prepare both the base project and your system if you expect your batch render to run properly.
Command Line Batching
Once you get your batch working, another nice feature of batch rendering you can take advantage of is the ability of Cinelerra to do this at a command line. I just kicked off a batch job at the command line and see that it does work on a box with X installed. Nicely, it also give you an ETA:
[sodo@tbear ~]$ cinelerra -r /mnt/videos/cinelerra/batch/batchList.xml /mnt/videos/cinelerra/batch/renderCompatibility.xml Cinelerra 2.1CV (C) 2006 Heroine Virtual Ltd. Compiled on Sat Jan 23 01:32:17 EST 2010
Cinelerra is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. There is absolutely no warranty for Cinelerra. Render::run: /mnt/videos/cinelerra/batch/renderCompatibility.xml Render::render: starting render farm [mpeg4 @ 0x7ff313d3ba00]warning: first frame is no keyframe [mpeg4 @ 0x7ff313d3ba00]concealing 1 DC, 1 AC, 1 MV errors 12% ETA: 0:07:45
Batch Results
So my goal was to profile my current system's capacity by capturing the system's cpu and disk utilization while the files rendered. To really see what is going on though, I thought it might be nice to have a graphical representation of the render as it occurs over time. So I spent some time writing a gnuplot script to plot the system utilization as the files rendered:
A bit of explanation may be required. I captured the output of vmstat to a file. VMstat has CPU load and wait i/o (disk utilization) statistics. While capturing that output, I kicked off the batch render.
In the graphic, you can see that different types of renders have different utilization profiles. For example, the mpeg4 renders were generally lower in CPU utilization (red line), while h264 renders used a lot of CPU. Similarly, the uncompressed formats like rgb/rgba/yuv420planar stress out the disk quite a bit (green line). Please excuse the fact that the filenames aren't perfectly lined up with each file's render profile..this was my first effort at graphing render times.
It will be interesting to see how the new Fedora 12 install affects the CPU. Also, I am planning on installing a new hardware RAID set, so I expect those green lines to go to zero (hopefully)!
I was compiling Cinelerra today and noticed that CPU usage was very low during the compile..around 10-15% utilized. I have a dual CPU, quad core box. This makes for a total of eight processors. So with all those CPUs, I figured there must be a way to make compiling faster.
Actually, this low CPU use during compiles was something I had noticed the first time I installed Cinelerra. Ashamedly, I've forgotten to investigate this issue in the two years that I've had the box. So today I googled for "make compiler see multiple CPUs" and found the -j switch to "make" the program that does the compiling: http://blogs.koolwal.net/2009/04/20/tip-compile-your-programs-fasters-with-multiple-processor-machines/
This article also mentions the CONCURRENCY_LEVEL environment variable, but that variable did not work for my box, a Dell SC1430. So I used the -j switch to make instead: [mule@ogre my_cinelerra]$ make -j7 (CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/sh /home/sfrase/my_cinelerra/missing --run autoheader) rm -f stamp-h1
Here are my results.
the time it took to compile Cinelerra normally, without -j: 3min 45s the time it took to compile Cinelerra with the -j8 (for eight cores): 1min 16s
Holy crap! That's a 300% speed improvement!
Glad I finally researched this. scott
ps - one other note: 7z is a multithreaded version of tar. On Fedora, use 7za Installing : p7zip-4.65-2.fc12.x86_64 7za - A file archiver with highest compression ratio
To be clear, I don't do upgrades. I will tar up my /home directory to USB, install the new OS from scratch and then blast my /home directory onto the clean new OS and RAID array.
Thinking Hard I've spent quite a bit of time planning this upgrade. One of the big things I am doing is to profile the performance of my system before and after the OS and hardware upgrades. Of course, I won't be able to determine whether or not the performance gain is coming from the OS or the new RAID array, but at the end of the day, I simply want to be able to say "my system is now X% faster."
I will be looking at the performance of the system from the OS, Cinelerra and encoder perspectives.
Virtual Machine Testbed One of the things that has helped me in the process is using VMware Server to test out Fedora 12. I've caught a couple things right off the bat: as it is a proprietary format, FAAC is not installed with FFmpeg by default. I was able to resolve this through Doran's excellent post here: http://fozzolog.fozzilinymoo.org/tech/2009/11/recompiling-ffmpeg-for-fedora-12-to-add-faac-support.html
Also, H264 encoder magic has changed a bit. Other than that, my output testing to various formats (MPEG-PS, HDV, DVD, iPod/iPhone) has worked very well.
General prep work work out bugs with Fedora 12 virtual machine clean up old F10 system backup F10 system files via script backup /home directory via tar to external drive
Installation steps Install new F12, Developer's edition Install RPMs via script Build and install FFmpeg RPM with faac support from nonfree RPM Fusion repo via script Install favorite programs Install Cinelerra dependencies Install Cinelerra
Audio synchronization problems are the bane of video editors everywhere. Here's what I do in Cinelerra to try to prevent them.
The first thing to understand is that playback in Cinelerra is a separate process from rendering (exporting) your project to a final format. So it is a good idea to calibrate the audio-video synchronization of your files when they are played back in the Cinelerra compositor as well as the a/v sync of rendered output. More specifically: 1) You need to calibrate the audio you see being played back in the compositor. Do this by setting Audio Offset in the Preferences -> Playback -> Audio Out so that your playback audio is in sync with the video. I generally use ALSA as my audio driver under Preferences -> Playback. (You might also want to try to use OSS or ESound to see if those audio drivers help your playback sync issues.) For reference, my offset is .1 seconds.
2) You'll need to calibrate the audio and video once it is rendered out of Cinelerra. This doesn't have to be the absolute final format of the file. Just an intermediate file format to determine that audio is not drifting and in sync. You'll do this by setting Nudge on your audio tracks so that your rendered audio and video are in sync. Here's an article on nudge: http://crazedmuleproductions.blogspot.com/2007/05/nudge-avidemux2-and-reminder-about.html
Let me give you an example.
My projects usually run about an hour on the timeline. When I start a project, I will concatenate a bunch of clips or raw videos on the timeline. I will then render a short segment of video, about 15 seconds long, from about 45 minutes into the timeline. The segment I take should have key features like someone speaking or someone hitting a drum that I can use to determine audio synchronization. I then render the file to a format that works well in the various Linux media players: -Quicktime for Linux using MPEG4 video compression and MP4 or twos complement (pcm) audio compression -Quicktime for Linux container using JPEG video compression and MP4 or twos complement
The side benefit of using MPEG4 video compression is that it renders damn fast.
The good thing about these steps is that if you work with the same format of video over and over, you won't need to change #1, the Audio Offset.
Here are notes regarding specific source formats and output formats.
MPEG SOURCE FILES to MPEG RENDERED OUTPUT After much study and pain, I've found that when working with MPEG-TS HDV files as my source (and MPEG formats in general), the best pipe to render from Cinelerra in order to alleviate sync headaches is mpeg2enc. I'm not five channel yet, so I just export my 48.1Khz stereo audio as MPEG Layer II audio MP3s @ 384kbps. I then use mplex to combine my audio and video streams into a program stream. Further on, I use VLC to convert that PS into MPEG-TS (HDV) and then use FFMPEG to convert the HDV into DVD and iPod/iPhone compatible formats.
QUICKTIME (QT) SOURCE FILES to QT RENDERED OUTPUT If my source video format is Quicktime for Linux (like from a screen capture), then I have no sync problems rendering out to QT for Linux.
MPEG SOURCE FILES to QT RENDERED OUTPUT If I render my MPEG-TS, HDV projects using Quicktime for Linux as an output format (MPEG4 or JPEG video w/twos complement or MP4 audio compression), then I will get sync problems in the rendered output. I then follow the above best practices to alleviate the problem.
TEST YOUR OUTPUT In Fedora, I've found that the best media players to test final output are mplayer, vlc and ffmpeg.
Finally, there are options to mplex and ffmpeg to offset audio and video sync problems: mplex -O ; eg, mplex -O 350 ffmpeg -itsoffset ; eg, ffmpeg -itsoffset 0.35
The format of the video edit is roughly the same for every video:
intro titles
staff titles
songs
end credits
Since the Cinelerra EDL (edit decision list) is an XML text file, it occurred to me that I should be able to add the titles by editing this text file.
Cinelerra's Edit Decision List (EDL) Let's look at Cinelerra's EDL. This is an XML text file that gets written when you save a Cinelerra project. The easiest way to review the EDL is to open it in a browser, as the browser will let you expand and collapse the XML elements. Here's a screen cap of the EDL in Firefox:
When I create my monthly video podcast, I place title effects on Cinelerra's timeline for each song of about a dozen songs. The title effect looks like this:
Placing title effects is a real time sink, especially when I have to place one for each song. I place title effects about 20 seconds into each song. I separate each song by placing a label between the clips. So the labels indicate where in the video track I will need to place title effects. You can see that the title effect is placed after the appearance of a label in Cinelerra's timeline below:
The Task My goal is to have a shell script find the position of the label in the EDL and plop the title effects about 20s after the appearance of the label. After I add the labels programmatically, I can do fancier edits to the project later on.
In today's post, I will describe a bit more about my workflow and the EDL, and also show you a basic shell script command to insert a title effect into the EDL.
The Procedure First, I put my basic edit together assembling the clips on the timeline, doing my audio and video fades and placing labels between the clips. This is my first round of edits. At this point, my timeline is very simple. I only have a video track and two audio tracks:
I will use a script to insert a title effect into the EDL file. In the Cinelerra EDL, a title effect (designated by the PLUGINSET XML element) looks like this:
Let's inspect the EDL in order to find the block of labels. If you grep for 'LABEL TIME' in the EDL, you'll find an XML code block that looks like this:
I will need to edit the video TRACK that appears directly beneath the LABELS in the EDL. Since there is quite a bit of confusing repetition in the EDL file, finding the list of LABELs helps identify which TRACK we need to edit. I've capitalized the words in the above sentence to highlight the XML elements you'll need to look for when editing EDL.
Using a bash shell script, I'll add a title effect after the last MASKAUTOS XML tag, but before the closing TRACK tag. Again, this is the closing TRACK tag of the video TRACK:
I've created a second file containing just one title effect, the PLUGINSET example from above:
I will then use a bit of sed (stream editor) magic to insert that file into the right spot in the main Cinelerra EDL for the project. Here is the script command that will place the PLUGINSET into the correct position in the XML: sed -e "/LABEL TIME/,/^<\/MASKAUTOS/{ /<\/MASKAUTOS>/r titleEffectPluginset.txt }" cinelerraEdl.xml > cinelerraEdlNewEdit.xml
Let's break this apart: sed -e "/LABEL TIME/ # This finds the first occurrence of "LABEL TIME" (a label) in the EDL. Remember that the appearance of labels in the EDL tells us that the video TRACK in which we need to insert our title effect comes next in the file.
,/^<\/MASKAUTOS/{ # This finds the first occurrence of a closing MASKAUTOS tag after the string "LABEL TIME". Our title effect, the PLUGINSET, will be inserted after this closing tag.
/<\/MASKAUTOS>/r titleEffectPluginset.txt # Insert the title effect boilerplate after the end MASKAUTOS XML tag
}" cinelerraEdl.xml > cinelerraEdlNewEdit.xml # Have sed perform the edit on "cinelerraEdl.xml", but save the output to "cinelerraEdlNewEdit.xml
The Result - Before
The Result - After
You can see that sed has inserted the title effect (the PLUGINSET XML boilerplate) in between the closing MASKAUTOS tag and the closing tag of the video TRACK. Pretty cool!
This is a simple example of editing Cinelerra EDL, but is the first step to helping me automate otherwise manual tasks in my monthly video podcast creation.
I will try to expand upon this subject in future posts.
I was interested in trying out Cinelerra 4, Heroine Warrior's latest version of Cinelerra. To reiterate, there are two versions of Cinelerra available: Heroine Warrior's, the original coded by Adam Williams (bow down to the man!) and Cinelerra Community Version, Cinelerra CV
On my Fedora 10, x86-64 setup, I gave the latest HV Cinelerra 4 version a try. I encountered a bunch of hurdles, mainly the GCC 4.3 changes that caused missing header information in mjpegtools-1.9.0_rc3 listed here: http://bugs.gentoo.org/show_bug.cgi?id=200767
I got the software to compile. At the final "make install" step, the installation starts, but does not complete with only the first 40 or so lines of the install finishing: [mule@ogre bin]# make install make -f build/Makefile.cinelerra install make[1]: Entering directory `/usr/src/cinelerra-4' make -C plugins install make[2]: Entering directory `/usr/src/cinelerra-4/plugins' mkdir -p ../bin/fonts cp fonts/* ../bin/fonts mkdir -p ../bin/shapes cp shapes/* ../bin/shapes cp ../thirdparty/mjpegtools*/mpeg2enc/mpeg2enc ../bin/mpeg2enc.plugin make[2]: Leaving directory `/usr/src/cinelerra-4/plugins' DST=../bin make -C libmpeg3 install make[2]: Entering directory `/usr/src/cinelerra-4/libmpeg3' cp x86_64/mpeg3dump x86_64/mpeg3peek x86_64/mpeg3toc x86_64/mpeg3cat ../bin make[2]: Leaving directory `/usr/src/cinelerra-4/libmpeg3' make -C po install make[2]: Entering directory `/usr/src/cinelerra-4/po' mkdir -p ../bin/locale/de/LC_MESSAGES cp sl.mo ../bin/locale/sl/LC_MESSAGES/cinelerra.mo make[2]: Leaving directory `/usr/src/cinelerra-4/po' make -C doc install make[2]: Entering directory `/usr/src/cinelerra-4/doc' mkdir -p ../bin/doc cp arrow.png autokeyframe.png camera.png channel.png crop.png cut.png expandpatch_checked.png eyedrop.png fitautos.png ibeam.png left_justify.png magnify.png mask.png mutepatch_up.png paste.png projector.png protect.png record.png recordpatch_up.png rewind.png singleframe.png show_meters.png titlesafe.png toolwindow.png top_justify.png wrench.png magnify.png ../bin/doc cp: warning: source file `magnify.png' specified more than once cp cinelerra.html ../bin/doc make[2]: Leaving directory `/usr/src/cinelerra-4/doc' cp COPYING README bin make[1]: Leaving directory `/usr/src/cinelerra-4' [mule@ogre bin]#
Therefore, the installation does not copy the cinelerra binary into /usr/local/bin. If I try to run the binary from the source code directory, I get this: PluginServer::open_plugin: /usr/src/cinelerra-4/bin/brightness.plugin: undefined symbol: glUseProgram PluginServer::open_plugin: /usr/src/cinelerra-4/bin/deinterlace.plugin: undefined symbol: glUseProgram undefined symbol: glNormal3f PluginServer::open_plugin: /usr/src/cinelerra-4/bin/swapchannels.plugin: undefined symbol: glUseProgram PluginServer::open_plugin: /usr/src/cinelerra-4/bin/threshold.plugin: undefined symbol: glUseProgram PluginServer::open_plugin: /usr/src/cinelerra-4/bin/zoomblur.plugin: undefined symbol: glEnd signal_entry: got SIGSEGV my pid=3965 execution table size=16: awindowgui.C: create_objects: 433 awindowgui.C: create_objects: 440 awindowgui.C: create_objects: 444 awindowgui.C: create_objects: 447 awindowgui.C: create_objects: 453 suv.C: get_cwindow_sizes: 744 suv.C: get_cwindow_sizes: 774 suv.C: get_cwindow_sizes: 800 suv.C: get_cwindow_sizes: 821 editpanel.C: create_buttons: 177 editpanel.C: create_buttons: 303 editpanel.C: create_buttons: 177 editpanel.C: create_buttons: 303 mwindowgui.C: create_objects: 192 mwindowgui.C: create_objects: 195 mwindowgui.C: create_objects: 199 signal_entry: lock table size=6 0x357dc80 RemoveThread::input_lock RemoveThread::run 0x64d0370 CWindowTool::input_lock CWindowTool::run 0x64f3640 TransportQue::output_lock PlaybackEngine::run 0x3441940 TransportQue::output_lock PlaybackEngine::run 0x3442420 MainIndexes::input_lock MainIndexes::run 1 0x3442f80 Cinelerra: Program MWindow::init_gui * BC_Signals::dump_buffers: buffer table size=0 BC_Signals::delete_temps: deleting 0 temp files SigHandler::signal_handler total files=0
Even though I have an NVidia graphics card, the error lines above were related to OpenGL. Thus, I thought I might have better luck compiling without OpenGL enabled. After I removed those lines from from hvirtual_config.h, I did a make clean;make. This time around, I was able to get Cinelerra 4 to start properly. Though, it soon locks up when viewing my 720P MPEG-TS files: [mule@ogre cinelerra-4]$ ./bin/cinelerra Cinelerra 4 (C)2008 Adam Williams
Cinelerra is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. There is absolutely no warranty for Cinelerra. [mpeg2video @ 0xeafd00]slice mismatch [mpeg2video @ 0xeafd00]mb incr damaged [mpeg2video @ 0xeafd00]mb incr damaged [mpeg2video @ 0xeafd00]invalid cbp at 14 37 [mpeg2video @ 0xeafd00]slice mismatch [mpeg2video @ 0xeafd00]ac-tex damaged at 25 40 [mpeg2video @ 0xeafd00]ac-tex damaged at 6 41 [mpeg2video @ 0xeafd00]ac-tex damaged at 7 42
So much for that experiment! I'm going back to the CV version for now. the mule
Occasionally, I'll need to send a video stream into mpeg2enc. Mpeg2enc doesn't take an input file; it only accepts a yuv4mpeg stream. In order to send a yuv4mpeg stream to mpeg2enc, I do this using ffmpeg and the -f yuv4mpegpipe command line switch. Also, for best quality, I will send the stream using the FFMPEG variant of the Huffyuv lossless compression algorithm. ffyhuff is an enhanced version of Huffyuv that compresses better than Huffyuv.
Update 2009/05/19As per Dan Dennedy's comment below, ffmpeg's yuv4mpegpipe command will ignore the -vcodec option and pipe the video stream to mpeg2enc using an uncompressed C420jpeg stream, which is an uncompressed YUV format. Certainly good enough for the likes of me! *** end update ***
Here is a sample command to reencode a 720P video stream as a yuv4mpeg pipe to mpeg2enc: ffmpeg -threads 4 -i INPUT.M2V -f yuv4mpegpipe - ¦ mpeg2enc --verbose 0 --multi-thread 4 --aspect 3 --format 3 --frame-rate 4 --video-bitrate 18300 --nonvideo-bitrate 384 --interlace-mode 0 --force-b-b-p --video-buffer 448 --video-norm n --keep-hf --no-constraints --sequence-header-every-gop --min-gop-size 6 --max-gop-size 6 -o OUTPUT.M2V
Note that I am taking advantage of the eight processors in my dual quad core using the multithread switches in the commands to both ffmpeg and mpeg2enc. Note that the eight threads have been split evenly, four to each encoder, to avoid CPU context switching. (Thanks again, Dan!)
Here's another trick: to see the header information of a YUV4MPEG stream, pipe the FFmpeg conversion stream to head -1 like so: ffmpeg -i intermediate.mov -vcodec mpeg2video -f yuv4mpegpipe - | head -1 ffmpeg -i intermediate.mov -pix_fmt yuv420p -f yuv4mpegpipe - | head -1
The FFmpeg output should show you some very important information, bolded below:
the output format: YUV4MPEG2 stream
height and width: 1280x720
framerate: 30001:1001 (or 29.97fps)
colorspace: C420JPEG
not sure what IP: 1 or XYSCSS is
As a final note, I am a bit confused on the differences between FFMPEG compression algorithms: ffyhuff and ffv1. If someone has pointers to the documentation on these, I'd be interested in finding out more. A google search just added to my confusion.
Over the weekend, I created a VMware Partner Account and got my Fedora 10, x86-64 virtual machine approved to be listed on VMware's Virtual Appliance listings: http://www.vmware.com/appliances/directory/148183
If you want to try out Cinelerra and you use 64-bit VMware Player, Workstation or Server, this is an easy way to get started. I'd appreciate someone giving it a shot and letting me know how it works.
Got my presentation worked out for the Trenton Computer Festival (http://www.tcf-nj.org/web/) tomorrow. It looks ugly, but the content of the presentation should outweigh the aesthetic of the slides:
This was fun. I spent the day perfecting a way to automate a line on a map in Cinelerra. You might think that was a somewhat pedantic exercise, but I think the image I used was very pretty and that the moving line, ala Raiders of the Lost Ark, came out great. What would make it even better would be to use an ancient map of some sort.
Update 2009/04/07 For some reason, this video is not playing as embedded on this page. Please visit my crazed mule profile on Vimeo to view.
Thanks! *** end update ***
Using Gimp to Spice Things Up I created the graphics in Gimp: -the line representing the route and its shadow -the circle representing the route's start -the star representing the route's end
The circle and the star were created using Gfig, the Gimp add-on utility that let's you create geometric shapes. Also note that the shadow of the line matches the position of the light source in the photo of the globe.
Note that the circle and the star are not flat, 2D creations, but they look like stickers pasted on the side of the globe. I acheived that effect by using Gimp's Perspective and Shear tools. Here's a resource that discusses Perspective in Gimp: http://gimp-university.blogspot.com/2008/03/perspective-and-layers.html
I created four images to import in Cinelerra: 1) globe with no Gimp object overlays 2) globe with just the circle as start of route 3) globe with the circle and the line 4) globe with the circle, line and star representing the full trip
Assembling the Images in Cinelerra The tracks in Cinelerra looked like this: Top Video Track image 1 (plain globe) at beginning of timeline and image 4 (all objects) at end of timeline Bottom Video Track image 2 (globe and circle) and image 3 (globe, circle and line)
Gradient Created for Line's Movement The key to the movement of the route was a screen wipe that travelled from the upper left corner of the screen to the lower right, mimicking the direction of the line's travel. Since Cinelerra does not have a built in wipe that moves in this direction, I had to create my own gradient using Gimp and plop it in /usr/local/lib/cinelerra/shapewipe. I then used that gradient in the Shape Wipe video transition tool:
In the timeline picture above, you can see the Shape Wipe transition effect that I used between the image of the map with the circle and the image of the map with the circle and the line.
I love the way this turned out, because it looks a s*1tload better than most of the other animated routes I've seen out there. In fact, it blows away the lame route create with Photoshop and After Effects that I read about in VideoMaker magazine this month. http://www.videomaker.com/article/14206/
I realized I never mastered a repeatable method to stabilize shakey video using Cinelerra's motion tracking tool. The Motion effect is very powerful, but also difficult to understand. At least for me. In addition, the manual isn't much help because it is couched in confusing terminology.
The motion tracker can do a lot of different things. However in this post, I am going to keep it simple and only describe how to stabilize shakey video. I made it easy for myself and chose a sample piece of video that bounces around pretty badly:
What You Need to Make It Work Since the manual's description of motion tracking is cryptic, I'm going to try to clarify the muddy waters. In order to stabilize a section of video, you're going to need a few things: 1) a easily identifiable object in your video that will be used to track motion 2) a box that encircles that object. The following is important: this box needs to be wide and tall enough to encompass the range of motion of the shakey video. 3) a video track (master layer) with the range of motion that needs to be stabilized 4) a video track (target layer) that will be stabilized
I would suggest starting small. Just try stabilization with a clip of video that is short (<10seconds) and needs stabilization throughout.
Step 1: Apply the motion effect to the video track you want to stabilize Like so:
Step 2: Open the Motion tracker effect dialog. In order to simplify the configuration process for the motion tracker, I've divided the configuration box into the only three sections you'll need to worry about:
Step 3: Enable Draw Vectors (in Section 2 of the graphic above) You may leave Track Single Frame selected. Also, Frame Number set to 0 means that the motion tracking of your video will start at the beginning of the timeline.
Step 4: Use Translation Block and Search Radius and Block X/Y to fit a box around an easily identifiable object in your video that will be used to track motion (in Section 1 of the graphic above) In the above picture, you'll notice there are two boxes around the Budweiser sign. The center box around the Budweiser sign is the Translation Block. You'll make the Translation Block fit neatly around the object you're tracking. The outer box is the Translation Search Radius. For the purpose of this tutorial, we'll make the Translation Block always fit within the Translation Search Radius. Below is a graphic depicting these components:
The Translation Search Radius needs to be as large as the range of motion of the video. In other words, the Search Radius needs to be large enough to accommodate all the shaking of your video. If the shaking extends beyond that box, strange things happen, like the motion tracker will start tracking another object in your video. Remember that.
Finally, Block X and Block Y represent the X/Y coordinate location of where you will move your Translation Search Radius.
In sum, you will configure those objects just discussed in the Motion Tracker effect dialog. To review: 1) Encircle the object you want to track with the Translation Block 2) Encompass the entire range of motion of your shakey video inside the Translation Search Radius 3) Use the Block X and Block Y coordinates to move the Translation Search Radius (including the Translation Block) around the screen
It is cumbersome to move the boxes around and X/Y coordinate plane using a round dial. The Translation Block and Search Radius should be drag and drop. The motion tracker interface can definitely be improved upon in this respect.
Step 5: In Section 3 of the Motion config screen, set Action to Track Pixel and set Calculation to Save Coordinates to /tmp The reason why we do this is that we are going to Track the motion of the video around our selected object (the Translation Block). The coordinates of the movement will be saved in temporary files, which we will later apply to a second track.
You may now either playback the video or render out a test video to see the results of the motion tracking. As the motion effect is very CPU intensive, I would recommend just doing a few seconds of playback or rendering, just to make sure the motion tracker is working properly. I also recommend rendering to a file, as it will be at the same speed as a playback, but will also give you some output that you can replay at will.
Reviewing Vector Paths and Translation Block object Once you've rendered out a test file, review the vector path to make sure the Motion tracker is always centered on the Translation Block, the object you want to track. I have found that the Motion tracker is easily confused if the object you've chosen to track is a similar color to the background. You'll know it loses track when the arrow on one end of the vector path no longer points to the original object in your translation block.
Also, the Motion tracker will lose track if the Translation Search Radius is not wide enough to capture the entire range of motion of the camera movement. In my Budweiser example, I found that I needed to widen Search Radius to more than half the width of the video so that the Motion tracker would stay on track.
Step 6: In the Motion effect on the original track, deselect Draw Vectors, set Action to Stabilize Pixel and set Calculation to Load Coordinates from /tmp
Step 7: Make a duplicate of your original track Once you have good motion tracking, you will then be able to apply your saved coordinates to another track or Target Layer. In my Budweiser example, I simply made a duplicate of the track that I generated the coordinates from. One way to make a duplicate of the original track is to: * in the patch bay of the original track, set both the playback and record to on * select the entire track (key "a") * press "c" for copy * create a new video track (Shift-T) * in the patch bay of the new track, make sure playback and record are both set to on * press "v" for paste
This procedure *should* copy the motion effect as well, with the settings from Step 6. If the settings from Step 6 are not in the Motion effect dialog, manually set them.
Step 8: Set the original track to not playback. The Target Layer (duplicated track) should already be set to playback from the last step.
Step 9: Playback or Render the Video Again, I suggest to render out the video to a file, as playing back or rendering will take the same amount of time.
Step 10: Analyze your results You'll find that with motion stabilization, your video will tend to bounce around and you'll see black borders appear along with the motion removal. The easiest way to remove them is to experiment with different zoom levels (Z axis levels) using the Projector (NOT the Camera). For my Budweiser video, I found I needed to zoom in 1.6x. Of course, the side effect is that it may ruin whatever cinematic effect you were trying to achieve. So be advised!
Here were my results from earlier today: 1) The original video:
2) Motion vectors being generated to /tmp:
3) Motion stabilized
4) Video zoomed in to crop after stabilization. Note this crops out most of the interesting content of the video:
Advanced Use I had a second video that bounced around quite a bit:
This time, I followed my own directions from above, but the resulting video came out jittery and jumpy:
Therefore, I increased the sensitivity of the Motion tracker by increasing Translation Search Steps from 256 to 1024:
This still was not sufficient, as I saw a couple jitters and jumps. I increased Translation Search Steps from 1024 to 8196. Be advised that this took about four times as long to render as having Translation Search Steps set to 1024. But it did remove the jitters and jumps!
The only thing in the tarball is libflashplayer.so. To install the 64-bit Flash plugin, simply move libflashplayer.so into your user's .mozilla/plugins directory and restart Firefox.
Even more amazing, the bloody thing works on my Fedora 10, x86-64 virtual machine running in VMware Fusion on my MacBook Pro! Yee haw! This will definetly help me as I'm preparing a presentation on Cinelerra for the Trenton Computer Festival in April.
Update 2009/02/24 You can avoid having to build Cinelerra from source by using Nicolas Chauvet's (Kwizart) precompiled Cinelerra installs: 1) install the Kwizart yum repositories http://rpms.kwizart.net/kwizart-release-10.rpm
2) install cinelerra-cv [mule@ogre doc]$ sudo yum install cinelerra-cv* --enablerepo=kwizart [sudo] password for sfrase: Loaded plugins: refresh-packagekit Setting up Install Process Parsing package install arguments Resolving Dependencies --> Running transaction check ---> Package cinelerra-cv.x86_64 0:2.1-21.git20081103.fc10 set to be updated --> Processing Dependency: bitstream-vera-fonts for package: cinelerra-cv --> Processing Dependency: libmpeg3-utils for package: cinelerra-cv ---> Package cinelerra-cv-debuginfo.x86_64 0:2.1-21.git20081103.fc10 set to be updated --> Running transaction check ---> Package libmpeg3-utils.x86_64 0:1.8-1.fc10 set to be updated ---> Package bitstream-vera-fonts.noarch 0:1.10-8 set to be updated --> Finished Dependency Resolution
Dependencies Resolved
================================================================================ Package Arch Version Repository Size ================================================================================ Installing: cinelerra-cv x86_64 2.1-21.git20081103.fc10 kwizart 6.3 M cinelerra-cv-debuginfo x86_64 2.1-21.git20081103.fc10 kwizart 9.6 M Installing for dependencies: bitstream-vera-fonts noarch 1.10-8 fedora 345 k libmpeg3-utils x86_64 1.8-1.fc10 rpmfusion-free 19 k
Transaction Summary ================================================================================ Install 4 Package(s) Update 0 Package(s) Remove 0 Package(s) Total download size: 16 M Is this ok [y/N]: y Downloading Packages: (1/4): libmpeg3-utils-1.8-1.fc10.x86_64.rpm | 19 kB 00:00 (2/4): bitstream-vera-fonts-1.10-8.noarch.rpm | 345 kB 00:00 (3/4): cinelerra-cv-2.1-21.git20081103.fc10.x86_64.rpm | 6.3 MB 00:05 (4/4): cinelerra-cv-debuginfo-2.1-21.git20081103.fc10.x8 | 9.6 MB 00:11 -------------------------------------------------------------------------------- Total 688 kB/s | 16 MB 00:24 warning: rpmts_HdrFromFdno: Header V3 DSA signature: NOKEY, key ID 5b01f801 kwizart/gpgkey | 1.7 kB 00:00 Importing GPG key 0x5B01F801 "Nicolas Chauvet (kwizart) " from /etc/pki/rpm-gpg/RPM-GPG-KEY-kwizart Is this ok [y/N]: y Running rpm_check_debug Running Transaction Test Finished Transaction Test Transaction Test Succeeded Running Transaction Installing : libmpeg3-utils 1/4 Installing : bitstream-vera-fonts 2/4 Installing : cinelerra-cv-debuginfo 3/4 Installing : cinelerra-cv 4/4
Building from Source Though editing video on Linux is never easy, I'm happy to say that Fedora 10 is finally stable, after I've resolved or worked around the various bugs I've encountered.
I built Cinelerra from the CVS repository (not Heroine Warrior's) on Fedora 10 x86-64 about a month and a half ago, but haven't had time to post the steps. I can say I've put the Fedora 10 build through its paces by editing all different formats in the context of 1080p video. I will add the caveat that Cinelerra is very choosy about the formats it likes, as shown in my testing results below: * Note that I haven't tested all combinations of containers and compression schemes, but this is a good first step
The steps are the same as the steps I ran to build Cinelerra on Fedora 9. Though this post will be rather short, consult my Fedora 9 post for all the details. FYI - the Fedora 9 system and Cinelerra build was so fraught with problems that I opted to move on to Fedora 10. I suggest you do the same.
Detail The below steps should all be run as "root" or sudo 1) install Fedora 10 I usually select the Developer's package, as it will include many of the developer libraries necessary to build Cinelerra from source. Be aware that this install is rather large, weighing in at around 7GB.
Update 2008/02/17 After reviewing the storage consuming "Developer" install, I decided to build out a "Custom" install of Fedora. The base + Cinelerra dependencies yielded a slimmer install, at about 3.5GB.
However, for ease of use, it is probably easier to go ahead and install the "Developer" install. I did not do this, and even with all the Cinelerra dependencies checking out as "Found", I encountered three problems: 1) g++ was missing (go ahead and do "yum install gcc-c++" to resolve this) 2) libXv-devel was missing (the Cinelerra make process failed on a libxv header file) 3) libXxf86vm-devel was missing (the Cinelerra process failed on "/usr/bin/ld cannot find -lXxf86vm")
3) install the dependencies for Cinelerra For this step, I've provided a script below that installs all dependent programs for a Cinelerra installation from two repos: Fedora base and RPM Fusion.
Paste the below text into a file, save it and run it as a script. Don't forget to "chmod a+x yourFile" in order to make your script executable. The script will install all the dependencies in order to build Cinelerra yum install gsm-devel \ libvorbis* \ libogg* \ libtool* \ libtheora* \ libpng* \ libjpeg* \ libtiff* \ esound* \ audiofile* \ libraw1394* \ libavc1394* \ freetype* \ fontconfig* \ nasm \ e2fsprogs* \ OpenEXR* \ fftw \ fftw-devel \ libsndfile* \ libiec61883* \ libdv* \ libquicktime \ ffmpeg \ xvidcore* \ lame \ lame-devel \ a52* \ faad2* \ x264* \ mjpegtools* \ faac* \ vlc*
4) get the Cinelerra source git clone git://git.cinelerra.org/j6t/cinelerra.git cinelerra_source
5) in the Cinelerra source directory, run ./autogen.sh
6) in the Cinelerra source directory, run ./configure
7) As long as configure shows no errors, go ahead and run "make"
8) As long as make showed no errors, run "make install"
The Mule has been working long hours for himself and you, valued video compatriots!
That sounds positive, as it should be. Though in truth, I am feeling less positive than that message implies. Personal and professional life has got me down, but is par for the course these days. Oh well. A pithy quote to pick myself up would be rather nice here. Instead, let me regale you of the past weeks activities, as some of the tribulations may help individuals in similar need.
Sh*t Storm This week, as I look back at my notes, I see a hailstorm of problems that I've dealt with: -Fedora 10, x86-64 spontaneous system lockups/reboots (workaround: noapic on kernel cmd line) -pulseaudio screwing up my audio -usb keyboard stops working (workaround: disable keyboard acceleration) -Gnome session saving broken (the workaround seems more of a pain than its worth) -1080p editing eats RAM! (bought more RAM) -Belkin firewire card causing reboots -I didn't order my RAM in matched pairs, so I'm stuck waiting until Monday for RAM! (finally got it!) -Evolution has trouble fetching mail from Comcast's POP servers, so I've reverted to use Pine (now "Alpine")
Needless to say, my productivity dropped and frustration was running high.
The Good News Knock on wood, I think I was able to workaround the spontaneous reboots using "noapic" boot option to the kernel. Whereas the box was rebooting every six hours, now it has been up a full two days without a reboot! Of course, this isn't a true fix and I will have to submit a bug to the Fedora team. And the other problems still exist.
Most importantly, I've discovered a new scheme for solid, fast 1080P editing in Cinelerra: 1) convert Canon 5D video to MPEG2-TS 2) import into Cinelerra 3) render to any format you need
A Couple of Options In my initial post on editing Canon 5D video, I found that the easiest way for me to get content from the Canon 5D into Cinelerra was using a conversion to MJPEG. However, the drawback with using mjpeg is that the image quality is lacking. Specifically, the output is darker than the original content. So over the past week, I found two solutions to convert the beautiful output of the Canon:
Now, this H264 content is beautiful, will import into Cinelerra and is editable. However, I found that when I went to render the final output, four minutes of the 1080p, H264 content took SIX HOURS to render!! That is unacceptable. I believe the lengthy render time has something to do with the color space or internal conversion that Cinelerra is doing. This bears further research.
Converting Canon to 1080p, MPEG2-TS Now, there are a few steps here.
a. Take a file from the Canon and use ffmpeg to pass a lossless yuv4mpegpipe stream into mpeg2enc, with the result a video stream with no audio: ffmpeg -i INPUT.MOV -threads 8 -s 1920x1088 -f yuv4mpegpipe - | mpeg2enc --multi-thread 8 --verbose 0 --aspect 3 --format 13 --frame-rate 5 --video-bitrate 24000 --nonvideo-bitrate 384 --interlace-mode 0 --force-b-b-p --video-buffer 448 --video-norm n --keep-hf --no-constraints --sequence-header-every-gop --min-gop-size 6 --max-gop-size 6 -o OUTPUT.m2v
Next, render out the audio: ffmpeg -y -i INPUT.MOV -acodec mp2 -ar 44100 -ab 256k -ac 2 OUTPUT.m2a
Using mplex, mux the video and audio streams together: mplex -f 3 -b 2000 OUTPUT.m2a OUTPUT.m2v -o OUTPUT.ps
Using VLC, convert the MPEG2-PS into an MPEG2-TS: cvlc OUTPUT.ps --sout '#duplicate{dst=std{access=file,mux=ts,dst="OUTPUT.m2t"}}' vlc://quit
Update 2009/02/13 I've found that VLC is not writing proper keyframes at the beginning of the converted MPEG-PS video output from mplex. This is only for 1080p video. The VLC command for 720p video still works. For the 1080p, I've found a workaround using our savior, ffmpeg: ffmpeg -y -i OUTPUT.ps -acodec copy -f mpegts -qscale 1 OUTPUT.m2t *** end update ***
The quality and the colors are definetly improved upon over the old version. However, the larger file size is a drawback (479MB for 4m16s of video). So I'd like to get the H264 output without compression artifacts during the scenes with a lot of motion. So now its time to figure that out. Erg.
In general though, I think this is some good news!
"Mules are not really stubborn. They can seem lazy because they will not put themselves in danger. A horse can be worked until it drops, but not so with a mule. The "stubborn" streak is just the mule's way of telling humans that things are not right. Mules are very intelligent and it is not a good idea to abuse a mule. They will do their best for their owner, with the utmost patience." About Mules
with 
video to MPEG2-TS
