John Yerhot – Weblog

I’ve now completed 11 screencasts for FOSSCasts.com, and I must say it has been a great learning experience.  I’d like to share the workflow that I’ve settled on.

Getting Started

First, the equipment:

• recorded on a Mac using Snapz Pro to record the desktop.  Lots of other Mac screencasters will recommend iShowU, but I already had a Snapz Pro license and it works well enough for me.
• Shure SM55S Mic
• Tascom USB Audio interface.
• Final Cut Pro to edit the video, add transitions, and overdub audio.  The first couple episodes were edited with iMovie.
• Keynote to create slides.
• Sun VirtualBox for virtualization
• Levelator to normalize the audio.  Levelator is a great piece of software.

Now, I actually got some advice from Ryan Bates  of RailsCasts fame as to actually recording the screencasts.  For the Quicktime videos I use the Animation codec.  One thing that Ryan pointed out is that the fewer pixels that change, the smaller the file size.  You many notice I don’t do a lot of moving my mouse and try to not use arrows when scrolling.  Just jumping to a section of a document is much better with the Animation codec.  Scrolling text, fades, or anything with lots of movement will make your file size grow quickly.

Now, the Ogg Theora codec is a different story. It is much closer to H264 than Quicktime Animation.  Generally I get smaller files with Ogg Theora, but they tend to have slightly less accurate color and detail.

For both I usually set the frame rate to 15 frames per second and keyframe every 160 frames with Quicktime animation and 24 frames with Ogg Theora.  I hear that Ryan has set this as high as 600.  Geoffery Grosenbach of PeepCode uses a lower value and has now started using H264 for many of his screencasts.  I may switch, but for the time being, Animation has given me better results, though a slightly higher file size.  I record at 800×600, same as Ryan Bates does for RailsCasts.

I thought about whether to upload the screencasts to something like Vimeo or YouTube, but in the end decided against it.  One thing I want FOSSCasts to be is well produced and high quality.  Once you convert them to Flash, the quality drops considerably, thus I decided against it.

Now, when it comes down to actually putting the screencast together, my workflow is as follows:

1. Research the topic.
2. run through what I want to do.
3. re-run through it while recording the desktop and talking into the mic.  The talking is just so I can roughly gauge myself and get a feel for what I need to say.
4. watch what I just recorded, keep what works, and re-record what doesn’t.
5. Load everything into Final Cut.
6. Go through slicing everything into sections so I can overdub the audio.
7. Go through, overdubbing audio, extending parts that should be longer, shortening others.
8. Create the slides in Keynote, exporting them to PNGs.
9. Import the slide PNGs, putting them into whatever order I need.
10. Record over them, shortening and extending as necessary.
11. Finally, add transitions, the ending slide, and recording over the ending.
12. Create the title slides, export it, and add to Final Cut.
13. Add the into and outro music clips.
14. Export only the audio to an AIFF and use Levelator to normalize it.
15. Import the normalized audio into Final Cut and export the entire movie to Quicktime and Ogg Theora!

Looking back at that list,  it is quite a bit and there is some room for improvement in my workflow.  I can usually get through the whole thing in 3-4 hours if I’m aiming for a 5-6 minute FOSSCast.

The hard part

Initially it was dealing with all the strange things I would say or noises I would make while recording the audio that I didn’t realize I was making.  For example, I would make this clicking noise between sentences.  Go listen the first episode and I’m sure you’ll hear some.  I also tend to do a lot of “umms” and “so’s”.  I’m still working those :)

Of course, you have to get over listening to yourself talk.  I found that after the first two weeks this was no longer an issue.  Learning how to talk into a microphone is also fun and something I’m still perfecting.

Also getting over “putting yourself out there” takes some time.

Thoughts

I might have to write a version 2 of this post in a year and see what changes in my work flow and how I feel about FOSSCasts then.  In the mean time, I’m having a blast.