xvidcap = great program, another reason not to leave Linux for Windows to be able to do something. (lately i thought about going Xp because of fraps and playing WoW, cedega or wine is fine but the movie making was a problem)
I have a question though.
1 - Could it be possible to make the encoding a bit more transparent? I am nub at the MPEG1 MPG MAVI or whatever.. like writing there Raw Avi, compressed Mpeg, Divx-like, frames jpeg, png, bmp?
2 - Could fullscreen option be added upon posting the target resolution somewhere into a dialog box? If cedega for example runs at 800*600 a game it uses 0.0 as 0.0 while not displaying the rest pixels over the game resolution till say 1280*1024. Thus the program would start automatically at this offset 0.0 + 800+600? I guess there is a way how to do that via cmdline options.. Maybe the program could have some profiles saved.. Like profile - fullscreen 800*600, fulscreen 1280*1024 or whatever.. maybe add 3 customizable ones to the interface.
3 - last question, could the program make automatically say recording in 1/2 size? Not sure how does the picture from screen get into the file so maybe that would reguire some postprocessing that would seriously make the process more CPU demanding..
Karl H. Beckers
one at a time:
(1) while I feel that this has considerable improved since 1.1.3, there is certainly room left for improvement. I'm thinking about reworking the preferences GUI so that the command line will stay the same (do auto-detection based on filename) but the GUI will have a main switch (on each of the tabs for single-frame or multi-frame capture, or potentially bridge the distinction) for file format, auto-detect stuff from there and at the end ask you for a filename. But that will certainly be after 1.1.4.
I don't understand what you're talking about at all.
You're trying to record the game right?
Then cedega will start the game in a window, no?
In that case, you should click the frame selection button and (with the cross-hair cursor) single-click the window. Now, if the window always appears at the same position, you might want to do:
right click filename button->preferences
in the "General" tab select "Save Capture Geometry", then "OK"
right click filename button->Save Preferences
Close xvidcap, reopen and behold it use your previous capture area.
Now, where does fullscreen step into the equation? And what's with this fullscreen that's not fullscreen completely? Are you saying that cedega still runs in a window but zooms in till nothing outside the windows is displayed? In that case the above should still work, no?!?
svn trunk has had this feature since yesterday. Atm you can go to the preferences GUI and set the "Rescale Output Percentage" in the "General" tab to define how big the output frames should be relative to the original, e. g. 50 will give 50 percent, 25 will give 75 percent reduction (output will be 25 percent the size of the original).
Aahhh, and about additional CPU requirements here, I am actually hoping it will generally REDUCE the load on the system as the rescaling (I hope) is less demanding than the encoding.
Oh i see now my 2nd question was just stupid enough :)
Cedega runs the game in fullscreen.. When i switch to desktop by alt-tabbing or changing to another desktop i see the first 800 pixels from 0.0 and 600 pixels down from the for example wallpaper. So it uses every time this pseudo fullscreen. but i guess i could just move the rectangle to the very topleft corner and make it save its position.
Anyway i have 1.1.3 version as it is the only available in standart gentoo portage tree, so i will try to upgrade to the very-recent version.
Thank you for the hints!
hmmm, you're talking about switching desktops ...
two things here:
To the Xserver there is only a single desktop. Switching desktops is typically done by switching what's called viewports.
Normally the wallpaper is drawn on the background of a display. Therefore, you can normally only have a single wallpaper for all viewports (not one across all). Now, you appear to have one of those programs running that simulate a single backdrop by catching you typing alt-tab and adjusting the portion of the backdrop being shown on the display.
All the same you only have one display and one set of coordinates.
And that means, if you tell xvidcap to capture say 800x600+100+100 it will always be on the current desktop. In other words, the capture input will follow your desktop moves, even if the main control does not. Otherwise, e.g. you wouldn't be able to record cube movement in compiz.
It could be interesting to find out whether it is possible at all to capture an image from an offscreen viewport. That should be possible, because things like the desktop switcher for luminocity could show contents of offscreen viewports real-time, too.