Episode 035 - Perspective Reflections
August 30th, 2007 by rfquerin
I’ve uploaded my latest Inkscape screencast (Episode 35). This one describes a neat method of creating reflections using a simple black to white mask. I discovered this neat trick at Nicu’s great blog.
While I’ve made reflections using Inkscape in the past, I’ve always had two problems. First, if the group of objects to be reflected consisted of multiple colours, using a simple opacity gradient was quite a pain. Second, creating a realistic looking reflection for an object in perspective (particularly photos) was very difficult. The one that we use on the screencaster LCD monitor thumbnail (at the top of this post) is really faked with a white opacity gradient. The problem being that if I changed the background to anything but white, you’d suddenly see the retardedness of it all. 
So the mask method eliminates these two problems and makes creating reflections of more complex objects (including perspectives) much easier and more elegant.
Hope you enjoy it.
For those who care about the mechanics of screencasting, this was actually the first one that I created solely with recordmydesktop, mencoder, and Audacity - although I did use an ffmpeg one-liner to extract a screenshot for the thumbnail (anybody know how to do that with mencoder/mplayer?). I think the quality is better and the workflow is definitely easier and more logical. Big thanks to Heathenx for helping me out with this stuff.
But it wasn’t completely without problem. You will notice some weird artifacting that occurs about 2 or 3 seconds into my sped-up intro. This wasn’t in the original ogg capture, but occurred somewhere during the post-processing. I thought it might disappear by the end - but no such luck. I’ve decided to leave it there for the time being and will correct it and replace the files when I can.. ie. not tonight.
One other thing is that I haven’t yet put this one up on YouTube. It’s about 12min long, so I’m gonna have to split it up like I did with Episode 26. I’d like to do it with mencoder as well this time, so in the next couple of days I expect to have the two-parts finished and up on YouTube as well.
August 31st, 2007 at 1:03 am
Hi there,
Since you have the capacity to host files, have you considered offering the
option of ogg/theora? Having usefull content, like yours, available in open media
formats, is crucial for their viability.
Anyway, the script bellow is what I use to extract thumbnails from ogg files,
with mplayer (and resize with imagemagick).
The -ss 15 is to define which frame will be the thumbnail.
#!/bin/bash
for i in `ls *.ogg`;
do
curfile=`echo $i|sed -e “s/.ogg//”`;
mplayer -nosound -vo png -ss 15 -frames 1 $i;
convert 00000001.png -geometry 256×192 _thb_$curfile.png;
rm 00000001.png;
done
August 31st, 2007 at 3:50 pm
Yes, we have considered using OGG format. At some point I would like to offer these but we are not quite there yet. We actually record into this format on our Linux boxes before converting them to H.264. Actually, someday I would like to offer even more formats.
Thanks for the bash script. I have used a similar script but I found that it was difficult to know which frame number to grab. Especially since the frame that we wanted could be anywhere in the video. Is there an easy way to determine what frame to grab? The FFMpeg script that we use works with time rather than frame. So we can watch up to a point, see where it is on time, and plug that into FFMpeg.
Thank-you, John
August 31st, 2007 at 11:29 pm
well, it seems I forgot to RTFM :).
the -ss option is for time actually(hhmmss format) , but it is kinda bugged when used
with -vo png and it always goes for the first frame.
Sorry for the misinformation …