Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-31 Thread Bryan Mayland
Bryan Mayland wrote: These work for me to, and give a vertically centered picture as well. For the record, the disassembly sets these (as a write4) to 0x1e1e701a for one set of standards and 0x28244024 for another set of standards. The complete function does some more setup: Oops, missed a

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-31 Thread Bryan Mayland
Chris Kennedy wrote: When I use these values, the picture is really strange looking, smearing horizontally, did these work for you? Thanks, Chris On Mon, May 30, 2005 at 11:37:33PM -0400, Ben Lancki wrote: I did a quick check on i2c dumps I did for my pvr150mce (NTSC) and the Hauppauge

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-31 Thread Chris Kennedy
When I use these values, the picture is really strange looking, smearing horizontally, did these work for you? Thanks, Chris On Mon, May 30, 2005 at 11:37:33PM -0400, Ben Lancki wrote: > I did a quick check on i2c dumps I did for my pvr150mce (NTSC) and the > Hauppauge driver sets the following

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-31 Thread Rune Petersen
Ben Lancki wrote: These values are for a PAL version of the card and I expect everyone in PAL country has a well positioned picture. I wish Rune Petersen --- This SF.Net email is sponsored by Yahoo. Introducing Yahoo! Search Developer

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-30 Thread Ben Lancki
I did a quick check on i2c dumps I did for my pvr150mce (NTSC) and the Hauppauge driver sets the following values: 0x474: 1A (VBLANK of 26) 0x475: 70 0x476: 1E (0x475+0x476=1E7) (VACTIVE of 487) 0x477: 1E The current values in ivtv are 36 for VBLANK and 580 for VACTIVE. These values are for a PAL

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-27 Thread William Powers
For grins, I expanded on the same theme to set the values for 0x474, 0x475, 0x476, and 0x477 to the defaults listed in the datasheet appropriate for the required video mode. I did this against an old version of the driver, so there's no sense patchifying it, but this is what it looks like: i

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread William Powers
That may well be true, but the patch does fix the bizarre 12 pixel wide black border on the bottom of PVR-500 recordings that is far wider than it should be and rather wider than (I believe) you get with the Windows driver. In fact, without the patch, the last good video line ends in mid-scree

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Chris Kennedy
And this patch does break VBI on the pvr150/500, plus you have to set the other register 0x476 to the same value, and they have to be offset of another register too which hasn't been changed. Well basically it seems this isn't what the Windows driver is doing, maybe there's another way to do this

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Sergey Alexandrov
my stupid head, of course I was talking about VACTIVE!On 5/26/05, Sergey Alexandrov <[EMAIL PROTECTED]> wrote: Great ... But is there problem with HACTIVE? I see black border at the bottom in NTSC ... it might be the similar problem.On 5/25/05, Ben Lancki < [EMAIL PROTECTED]> wrote: This was one of

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Chris Kennedy
Actually without that VBI won't work, the Hauppauge driver sets that up exactly as we do for PAL and NTSC, odd but a fact, without that setup VBI doesn't work (actually the last 4 registers set in there, but all in a bundle, and improves picture quality too, seems to override the trigger tolerances

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Ben Lancki
According to the docs VBLANK should be set to 22/34 in autoconfig, but according to the register definitions, it defaults to 20. If disable setting the register, it does in fact get set to 20 rather than 22. I think if this value was manually set to ~24 the vbi data at the top would no longer be vi

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Bryan Mayland
The datasheet also says: "In autoconfiguration mode, the default value for the vertical blanking delay, VBLANK (0x474 an 0x475) is set to 22 and 34, respectively, for each video standard." So maybe we should just let it autoconfigure. Also, I've never noticed this before, but why does vbi_reg

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-26 Thread Ben Lancki
The decreased value in VBLANK should remove the black border at the bottom by shifting everything down (I think this is the border you're referring to). I didn't touch HBLANK or HACTIVE yet since there wasn't any guidelines in the documentation. Currently they're at the default values, but I do not

Re: [ivtv-devel] [PATCH] VBLANK adjustment

2005-05-25 Thread Sergey Alexandrov
Great ... But is there problem with HACTIVE? I see black border at the bottom in NTSC ... it might be the similar problem.On 5/25/05, Ben Lancki <[EMAIL PROTECTED]> wrote: This was one of the last visual things bugging me with the driver. Thescreen always appeared vertically offset by 10-15 pixels

[ivtv-devel] [PATCH] VBLANK adjustment

2005-05-25 Thread Ben Lancki
This was one of the last visual things bugging me with the driver. The screen always appeared vertically offset by 10-15 pixels from the broadcast video. According to the cx25840 documentation, VBLANK should be 22 for NTSC and 34 for PAL/SECAM. This patch sets it to 22 or 34 depending on the card r