Special software is used to capture the video either in one go or in multiple steps depending on the length. Simultaneously you record English voice over (also called "original" voice over) into a microphone with a separate sound recording software. This results in two files (or one avi/wmv), a video and an audio file, which must be merged in the next step.
Regardless, which software you use, you should always end up, with a video that utilizes these minimal features. Use these clues:
Its VERY important to turn off hardware acceleration of your graphic card. This can be done, using the settings for your graphic card subsystem. To do this on Windows 2000/XP, go to the Control Panel, choose Display : Properties : Settings : Advanced : Troubleshooting. Set the Hardware Acceleration slider to None. For other versions of Windows, go to the Control Panel, choose System, (and under the performance tab), choose Graphics : Advanced. Set the Hardware Acceleration slider to None. The performance boost is dramatic.
800x600 pixel screen resolution is recommended. If not using Windows XP, switch to 256 color mode (XP does not support this mode anymore).
The more fps we get, the better, but keep in mind, that we wont need more than ~10 fps. There is a relationship between 'CPU usage during recording' <-> 'fps' <-> 'final file size'. 10 fps give a very smooth looking video and we can use the rest of the CPU for doing the presentation in a stable way, because the CPU is not only loaded with compressing the video.
Record English "original" speak simultaneously in 44.1KHz / 16bit / Mono.
If possible start all videos with a visual "sign" which can be heard on the audio track as well so you have something to sync video and audio with.
Also end each video with 5 seconds silent screen - this makes it more tolerant to make audio dubbing later. Plus the video does not end so abruptly.
After experimenting it quickly turns out that the absolutely best solution is to create as good source material as possible. Better redo the recording it than editing it. And in particular make sure to record commentary while recording the video (so both parts follows each other naturally)! Get it as right as possible the first time!
To capture your Screen, you can choose between several software solutions. These are some I tested (there might be more tools available)
www.microsoft.com/windows/windowsmedia
Microsofts Media Encoder is a very powerful tool, for creating any kind of videos. But the screen capture codec, delivered with this product, is the best of the best (at the time of writing this document) and it is free. Unfortunately the wmv files are only playable on Windows computers.
After launching Media Encoder, a wizard starts up. Click on cancel. We will make manual settings.
Click on 'Properties', a dialog opens. We chose 'Source from Device' and 'Screen Capture' for Video Input. Audio, should be your audio hardware. Here your microphone etc. needs to be correctly connected and configured. See Screen shots.
In the Output Pad, make modifications, to write the video into a wmv file.
Now go on to 'Compression', the most important part. There might already be a streaming profile in place, but we want to make a new one, so click edit. After that use the 'New' button, to create a new profile. Just look at the screen shots for good settings The kBit/s rate is only a maximum value – the encoder always runs in variable bitrate mode. Don't forget to mention the correct language for that video.
Add a new target bitrate of 300 kBit/s
Audio: CBR – For first recording use CBR Windows Media Audio Series 9, 48 kBit/s 44.1 kHz, 16-Bit, mono. You might wonder, why not the voice encoder @ low bit rates. The reason is, that we might want to end up with a HQ audio file, for later post processing. Perhaps an alternative to WMV will be developed by the Open Source Scene, and in that case, it is good for us, to have an archive of high quality audio files. If we want to encode the audio streams into mp3, it is also very important to have HQ footage. I will show you later in this document, how to extract the audio data of a WMV file.
Video: CBR – Windows Media Video 9 Screen Video Size, 'Same as Input', FPS 10, the rest of the values should be left at their default state.
Note: The video bit rate settings are only the maximum allowed bit rate of a variable bit rate encoding. So the encoder decides by itself the optimal bit rate to represent a good quality video. The same is valid for the fps. Media Encoder is one of the few encoders, that support variable FPS – so if nothing happens in our videos, it automatically decreases the frame rate and saves a lot of bandwidth,
Pressing the 'Apply' button, we are ready to capture our screen session. After that, you might want to convert the wmv file into an avi (described later in this document), to make post processing etc.
This is the crème de la crème, of all screen capture products, but its not a cheap software. You can get a test version for 30 day, so you can try it out, before you buy it. It ships with its own codec, which is also very powerful.
http://sourceforge.net/projects/camstudio, www.ehelp.com, www.rendersoftware.com
This is a free application, which is very powerful. It comes with its own lossless and powerful video codec. The Video settings are a little bit tricky – you have to find the best values for your cpu. Since we don't need more than 10 fps, the settings, shown in the screen shot are ok. For audio choose PCM WAV, 16-Bit, mono, 44.1 kHz. So you will end up with a high quality screen capture video, that has ~10 fps and perfect sound.
There are some interesting point in the faq, which I want to mention here:
When I play back a full-screen AVI file using Windows Media Player by double clicking it, the text and graphics becomes blurred.
There are two main reasons for the movie being blurred. One is that you are using Windows Media Player to play back a movie that has a frame size that is as big or bigger than the screen. In this case, Windows Media Player will shrink the picture to fit it on the screen. This cause the text and graphics to be blurry. To view the movie in full quality, you will need to switch Windows Media Player to full screen playback, or switch your monitor to a higher resolution. You may also record a smaller region to avoid this problem. Otherwise you may need to use the Movie Player software that is distributed with the CamStudio package to playback the movie. Another reason for the cause of the unclear image is the use of Lossy Codec as your compressor. This means the compressor will degrade the quality of your picture to reduce the size of your AVI file. To remedy this, you may either set the Quality settings in Video Options to a higher value, or use a Lossless Codec for your compressor (e.g Microsoft RLE is a lossless codec that is available only in the display mode of 256 colors).
I record with the default settings of CamStudio and it uses up all my CPU resources. How can I free up more CPU resources for doing other stuff?
In the default setting, CamStudio uses up all your CPU resources to achieve the maximum recording rate possible on your system. You will need to set this capture rate to be lower than the maximum to free up resources for other tasks. To see how fast your computer can go, try recording with a fixed region size at the maximum frame rate (the auto-adjust slider is set to the leftmost point in Options:Video Options). Take note of the Actual Input Rate when CamStudio is recording. This is the maximum frame rate achievable on your system at this fixed size. To free up CPU resources, go back to Video Options and set the capture rate (or playback rate) -- using the auto-adjust slider -- to be less than the Actual Input Rate noted earlier. If you are recording with the same fixed region size, your CPU meter will now show some improvements.
How can I optimize the video settings to get the best results ? Can you suggest a good video setting?
In Options: Video Options set the Capture Frame Every value to 5, set the Set Key Frames Every value to be 200, set Playback Rate to 200. In general, you should adjust the Set Key Frames Every and Playback Rate to be equal 1000/Capture Frame Every. For example, if Capture Frame Every is 5, then the Playback Rate should be 1000/5 = 200. However, if you are creating a time-lapse movie, (in which your Capture Frame Every is a very large value), you may want to set the Playback Rate to be around 20 to 30 frames/second.
The one Kasper used to produce the videos. Seems to work quite fine. Does not come with its own codec, so you have to compress in real time to an uncompressed video or to a very fast codec (like MJPEG). This is a great disadvantage, because you can only capture at low fps and still a powerfully CPU is necessary. This tool doesn't seem to me of being optimized for the task of taking screen capturings. This is the settings used:
These tips are good to keep in mind when recording the video/audio:
Generally, use a stop watch so you can keep track of time
Always check that the video gets captured correctly afterward: Sometimes it can stall in the end.
If real-time audio recorded: When starting a capture session, start sound recording, start stop watch, press F7 to start Capture and when capture is running a) click a link in the browser or b) mark a text and unmark it by a loud click sound. The point is to create a sound plus visual sign so synchronization is easy later! For the end of the video you should say "Stop" and then stop sound recording. Probably you can hear on the sound signal as noise when the video capture has ended. Check synchronization in Premiere afterward: Beginning, End and Middle.
If real-time audio recorded: Avoid talking when waiting for the system to respond - thus you can edit out waiting periods easily.
Using CoolEdit to record the sound from microphone
To record sound from a microphone in CoolEdit you just click the Record button when no document is open and you get these settings up. Then make sure they are set to 22KHz / 16bit / mono:
When the recording is finished, normalize it:
In the Normalize box all settings are fine right away. Just click OK:
The result is something like this which you save as a Wav file:
Tips to recording audio on laptops
It turned out that I got (significant!) less noise on the recordings from my microphone connected to my laptop if I totally removed the power plug and ran on batteries while recording!
INTRO: Recording sound with microphone in one channel and the computer lineout with mouse clicks in the other is something I have done in a few cases. But it's very unpractical and personally I find that the videos works just as well without the mouse click sounds etc. So why waste time and energy on this... In the end I skipped it.
The recording of sound can also be combined with mouse clicks and keyboard sounds if you are really fancy. See Appendix C for details. But do your self a favor and just forget about it. Takes too much time and hassle.
If you want to record sound while recording the videos I have used Creative WaveStudio or CoolEdit (or whatever, but not too processor intensive...) to record real time audio.
WARNING: With 4fps video capture I experienced that audio and video got out of sync since the video capture was delayed in the neighborhood of 5%. With 3fps - no problems!
I recorded the voice in one channel and the click sound from the computer in another channel. This was to allow audio-mixing afterward and dubbing in other languages.
This required a mixer!
Mixer setup
Microphone in LEFT
ONE channel from line-out of computer into RIGHT
Stereo line-out back into computer line-in
Computer Volume Control
The setting of audio controls (WinXP: Accessories > Entertainment > Volume Control) looked like this:
Recording Control
(Select menu: Options > Properties, then "Adjust volume for" = "Recording")
Main point is that you are recording the line-in signal.
Playback Control
(Select menu: Options > Properties, then "Adjust volume for" = "Recording")
Main point is that you are muting the output of the line-in signal - otherwise that will be sent back to the mixer which is recording the system-sounds of mouse-clicks for you.
Mouse-clicks in WinXP are normally played when clicking links in Internet Explorer or links in the Windows Explorer.
Make a test. Make sure the volume does not clip the signal. Enhancements can be done afterward in CoolEdit.
Machine was WinXP, 2.4GHz. Has successfully been done on 800 MHz Pentium with uncompressed/4fps.
"Uncompressed" will generate very large files which will later be compressed again with MS-RLE codec (RLE because this is screen shots which compresses best with RunLengthEncoding schemes to keep the clarity of detail) or QuickTime/Animation mode. On my WinNT I could not successfully capture uncompressed video! Too slow? Using compression in CorelCAPTURE will make blurry videos.
Don't create files larger than 2GB - that will make them corrupt! 480 seconds/ 3fps gives file of 1.923.786 KB at 800x570
Capture at 800x600 screen res. Open Browser window. MSIE will automatically fill in the 800x570 area. Otherwise scale the "start-window" to have this size! The "start window" is the active application window when capturing begins. This will determine the size of the capture. Make sure the browser window in 800x570 is ALWAYS the active window at start.
A test with 4fps (instead of 3fps) resulted in the 360 seconds video being captured over MORE time. This meant that synchronization with any real-time recorded sound was very hard since the video clip had to be time-stretched. 3 fps recorded the video in full realtime. Probably the lagging behind was due to bad graphics card.
Always check the captured video afterward; does it play back well? In particular check the end of the video! Some times it has been frozen for the last 20 seconds or so. Seems the best is to capture on a computer with a fast graphics card and at a low resolution like 800x600.
