Re: Deinterlacing?
Bj|rn Englund <[email protected]> Wed, 28 Sep 2005 18:47:14 +0200
| Newsgroups | gmane.comp.video.ogle.devel |
|---|---|
| Message-ID | <[email protected]> |
Wed Sep 28 2005, [email protected] wrote: > I think that this is where I'm not getting something. I thought that, for > the normal (not deinterlaced) case, the entire xv_image was always used. > Is there a feature that actually zooms into a *part* of the picture, > rather than juzt zooming the entire picture into some other size? If so, I > did not notice it and don't understand it yet. Yes there is a feature that zooms into a part of the picture. When not using this feature, everything worked fine, even without my last patch. To zoom you press the middle button in the top left corner of the area you want to zoom/crop and then drag to the bottom right corner and release. If you are in windowed mode this will crop the picture (so you can remove letterbox borders for instance) and if you resize the window you will still only see the cropped part so it will work like a zoom. In fullscreen mode the cropped part is always scaled to fit the screen so it will zoom into the crop area. To get back to the original picture press and release the middle button without moving the mouse (or I think you can do a "negative" drag too). With the last patch the zoom works as it should. > Now a real hit for this method of deinterlacing would be finding a way of > not even *rendering* the dropped field. This would impact CPU usage and > make ogle usable for DVD playback on low-end machines. But this is a > future challenge. I don't think it will make such a big difference. How low end machines are we talking about? A PII 450 should work ok for most films. Anyway if I remember correctly, you have to decode both fields because one field can use information from the other field of a previous/following picture in the decode stage. /Bj|rn