Screencasts with Ubuntu

Ubuntu disappoints again… although I don’t exactly know who is to blame: Ubuntu, the programmers at Wink, or me.

Today, I was preparing some basic tutorials for VMware Player.
So far, the mini-course, if that’s what you want to call it, consists of four parts:

1) Downloading the VMware Player from the VMware website;
2) Creating a virtual machine with easyvmx.com;
3) Mounting an ISO file directly modifying the definition file of the virtual machine, and
4) Obtaining the VMware Tools for Linux from VMware Server.

The truth is that—more than creating a mini-workshop—I wanted to test whether it could perform screencasts in Linux just like it did in Windows.

The answer is NO.

When searching on Google, I found over and over again the same few options for live screen captures: RecordMyDesktop, Istanbul and XVidCap (for a summary, see the article on Linux.com). Because I am used to doing screencasts with Camtasia, I found it very inconvenient that none of these options allowed me to edit captures. Suppose I captured a video in OGG format, or Theora, how would I edit it? Remember, I am VERY new to Linux,but my needs are simple (e.g. cut out pieces of recording where I made mistakes or where progress bars consume too much time; add some titles, and perhaps make notes on the screen). To do all this, do I have to install Cinelerra? It occurred to me once to do this, but I simply did not know where to begin (it was not in the Ubuntu repositories).

To show that I’m not afraid of trying new combinations and that I do not have to use Camtasia, you should know that for a while I used the following stack of applications:

Camstudio: Small but powerful. It does not require installation; it’s a FreeWare (possibly also OpenSource, as I have just read in the comments), recording whatever happens on the screen (or on part of the screen). This is what I wanted it for.
VirtualDub: Another small portable wonder. It allows me to cut out mistakes, shorten progress bars, and so on. As I’m new to this, I used it only a little at first until someone told me that I could select the “direct stream copy” mode (or something like that) so that I did not have to re-encode the entire video.
Riva Flash Video Encoder: Using this easy and free encoder, I converted the video from FLV to YouTube.
JW FLV Video Player: This incredible player allowed me to embed videos on wiki websites in which the screencasts were posted to document my hardships with the software that I had mastered (HP OpenView Select Identity and HP OpenView Select Access, now discontinued by HP). JW FLV Video Player lets you add subtitles to videos although the process of creating them is completely manual. However, theoretically, there is a freeware product that helps you create them: Subtitle Workshop (but it never worked for me).

With this background—and an eagerness to learn new things on Linux—I thought that a couple of Google searches would be enough to solve the problem, though with another stack of applications similar to those for Windows. Unfortunately, it was not.

Of course, there’s always a Plan B. In my case, Plan B has always been Wink. This software has always surprised me by how simple and powerful it is. It is ideal for posting small demos or giving instructions to someone on how to do something with their computer. Today, after the initial fiasco, I recorded the screencasts with Wink in a Windows virtual machine.

But, to demonstrate how VMware Player works, I had to capture the activity of the physical computer, i.e. that which runs Ubuntu, which is where VMWare Player is installed (I don’t think you can install VMware Player within VMware Player).

So, I decided to install Wink for Linux. I searched on Synaptic, and bingo! I installed it and the first problem I encountered was that it is unable to open files created by Wink Windows. I thought it was because I recorded them compressed, so I opened the virtual machine and I tried to record them uncompressed. Still, nothing.

I decided to go ahead and capture the rest of the series on Linux, and to worry about making them look the same later. Then, the second fiasco occurred. I tried to perform a “timed capture”, but it did not work. Actually, it captured what it wanted to capture…

So, I opened Wink again, indicated that I wanted to capture an area of 800×600, clicked on “Choose” and an outline of the selected area appeared. As I selected the area, the outline disappeared so that I had no visual reference for when something was within the area being captured and when it was not.

Nevertheless, I tried one more time, but again, it didn’t work! It captured a huge quantity of meaningless frames, in which nothing happened (no key press, no mouse clicks, etc.) In addition, there was a part of the recording that was cut because it was outside of the area to capture (as can be expected when there are no references).

But the worst was still to come. I attempted to remove the first five frames by clicking “Delete”, but nothing happened. Did I not hit “Delete”? So I tried again, but again, there was no response. While I wondered what was happening, the five frames were suddenly erased… how did that happen? I thought maybe the first time is always a bit slow, but on my second try, removing just one frame, the same thing happened.

I tried to change a cursor to see that it was in fact clicking on something when an error popped up. This is why I think it is a problem with the program that has been poorly designed.

In the end, I had to uninstall Synaptic (in fact, I chose to “Delete it completely”), but not before capturing the error of accessing the icons.

And now it remains in the Google history cache, as a shameful reminder to the Wink programmers.

Thank you to Self for providing the main inspiration for this article.

Speak Your Mind

*