Hi all,
I've been thinking a bit on the problem about incorrect picture width and height (see my earlier post and MichaelK's in this forum about black borders around the picture).
Answer from some contributors was : Tv picture contains garbage at the beginning and at the end of each line and this garbage is encoded by the TV card. OK. I agree with this, and Tv overscan allows not to see this garbage when using a TV Set alone. But when digitizing a PAL signal according to CCIR601 recommendations, in YUV4:2:2, there must be 720 valid pixels per line. Digitizing process must manage to produce 720 valid meaningful pixels per line, and leave garbage out, which means it should start capturing at a given offset on each line, and produce 720 Pixels for each line.
Obviously this is not the case with Hauppauge PVR150 and one would have to compensate for this by cropping garbage out and resizing to 720x576 afterwards.
This should be done in the capture driver, but to be honest, I don't know if it is feasible from some kind of SDK or API, or else.
This could be done in MP by allowing the user to define an offset before capture.
For example, if there are 16 pixels garbage at the beginning of each line, one could specify : image resolution 720x576, offset 16. This would start capturing each line only at the 17th pixel in the frame buffer and capture 720 pixels only for each line, leaving garbage out.
Pictyure would then be correct
Would it be complicated for MP to allow some cropping to be done (for example let the user preset cropping values in the configuration application) and then do a small resize to fill the entire 720x576PAL (or 640x480NTSC) with the correct picture ?
MCE2005 handles this without any problem on the same TV card so I guess MP can do...
I've tried to add ffdshow cropping but it just gives a scrambled picture Any comments from the MP developers ?
Am I wrong here ? Talking nonsense ?
(Sorry for confused English, I'm French
Pierrelo 8)
I've been thinking a bit on the problem about incorrect picture width and height (see my earlier post and MichaelK's in this forum about black borders around the picture).
Answer from some contributors was : Tv picture contains garbage at the beginning and at the end of each line and this garbage is encoded by the TV card. OK. I agree with this, and Tv overscan allows not to see this garbage when using a TV Set alone. But when digitizing a PAL signal according to CCIR601 recommendations, in YUV4:2:2, there must be 720 valid pixels per line. Digitizing process must manage to produce 720 valid meaningful pixels per line, and leave garbage out, which means it should start capturing at a given offset on each line, and produce 720 Pixels for each line.
Obviously this is not the case with Hauppauge PVR150 and one would have to compensate for this by cropping garbage out and resizing to 720x576 afterwards.
This should be done in the capture driver, but to be honest, I don't know if it is feasible from some kind of SDK or API, or else.
This could be done in MP by allowing the user to define an offset before capture.
For example, if there are 16 pixels garbage at the beginning of each line, one could specify : image resolution 720x576, offset 16. This would start capturing each line only at the 17th pixel in the frame buffer and capture 720 pixels only for each line, leaving garbage out.
Pictyure would then be correct
Would it be complicated for MP to allow some cropping to be done (for example let the user preset cropping values in the configuration application) and then do a small resize to fill the entire 720x576PAL (or 640x480NTSC) with the correct picture ?
MCE2005 handles this without any problem on the same TV card so I guess MP can do...
I've tried to add ffdshow cropping but it just gives a scrambled picture Any comments from the MP developers ?
Am I wrong here ? Talking nonsense ?
(Sorry for confused English, I'm French
Pierrelo 8)