Re: [PATCH 0/3] gspca: kinect cleanup, ov534 port to control framework
On Wed, 16 May 2012 23:42:43 +0200 Antonio Ospite osp...@studenti.unina.it wrote: The second patch removes the dependency between auto gain and auto white balance, I'd like to hear Jean-Francois on this, the webcam (the ov772x sensor) is able to set the two parameters independently and the user can see the difference of either, is there a reason why we were preventing the user from doing so before? Hi Antonio, I added this dependency by the git commit 2d19a2c1186d86e3 on Thu, 12 Nov 2009 (the original patch was done under mercurial). Looking in my archives, I retrieved this mail I have sent to you, Max Thrun, kaswy, baptiste_lemarie, Martin Drake and Jim Paris: On Thu, 12 Nov 2009 13:24:43 +0100 I wrote: On Thu, 12 Nov 2009 11:13:32 +0100 Antonio Ospite osp...@studenti.unina.it wrote: On Wed, 11 Nov 2009 19:13:51 -0500 Max Thrun bear2...@gmail.com wrote: *I get a weird effect, something like a mosaic effect caused by a picture shift, at such high frame rates (also at 640x480@60), I need to verify if it is my usb host which is weak.* [snip] Maybe the End Of Frame detection logic is still imperfect, but I have to admit I haven't looked at it lately. You are heading to face/object tracking, aren't you? Very interesting. When adding the ov965x, I removed the check of the image size. May you try to set it back? (sorry, I have no patch - the check must be done at 2 places, just before adding the LAST_PACKET - the 2nd is enclosed in #if 0) * * Brightness control in guvcview doesn't seem to work.* Confirmed. Easy fix though: [snip] Thanks, please send patches, they are so easy to create from Mercurial that I don't think you have many excuses for not doing so :) Thanks also from me. I already did and uploaded the fix. * * AWB doesn't have any effect?* I notice its effect if i start uvcview, enable auto gain, then enable awb. If there is a strict dependency between these two settings, shouldn't the driver enforce it? [snip] It should! This asks for a change in the main gspca. I will try to do it quickly. Otherwise, you are right, the ov7670 and ov7729 datasheets do not talk about a possible AGC and AWB dependency... -- Ken ar c'hentaƱ | ** Breizh ha Linux atav! ** Jef | http://moinejf.free.fr/ -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 0/3] gspca: kinect cleanup, ov534 port to control framework
On Fri, 18 May 2012 09:08:29 +0200 Jean-Francois Moine moin...@free.fr wrote: On Wed, 16 May 2012 23:42:43 +0200 Antonio Ospite osp...@studenti.unina.it wrote: The second patch removes the dependency between auto gain and auto white balance, I'd like to hear Jean-Francois on this, the webcam (the ov772x sensor) is able to set the two parameters independently and the user can see the difference of either, is there a reason why we were preventing the user from doing so before? Hi Antonio, I added this dependency by the git commit 2d19a2c1186d86e3 on Thu, 12 Nov 2009 (the original patch was done under mercurial). Looking in my archives, I retrieved this mail I have sent to you, Max Thrun, kaswy, baptiste_lemarie, Martin Drake and Jim Paris: [...] * * AWB doesn't have any effect?* I notice its effect if i start uvcview, enable auto gain, then enable awb. If there is a strict dependency between these two settings, shouldn't the driver enforce it? Here I made a wrong assumption at the time, the bug must have been somewhere else, forgive the younger me in that email :) Otherwise, you are right, the ov7670 and ov7729 datasheets do not talk about a possible AGC and AWB dependency... OK, thanks. Regards, Antonio -- Antonio Ospite http://ao2.it A: Because it messes up the order in which people normally read text. See http://en.wikipedia.org/wiki/Posting_style Q: Why is top-posting such a bad thing? -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 0/3] gspca: kinect cleanup, ov534 port to control framework
Hi, Thanks for the patches. I've added them all to my tree, so they will be included in my next pull-req. In the mean time you can find them (unmodified) here: http://git.linuxtv.org/hgoede/gspca.git/shortlog/refs/heads/media-for_v3.5-wip On 05/16/2012 11:42 PM, Antonio Ospite wrote: Hi, the first patch just removes traces of the gspca control handling mechanism from the kinect driver; this driver does not have any controls. The change is trivial and can be applied right away, or postponed to when the gspca_main code is removed, you decide. The second patch removes the dependency between auto gain and auto white balance, I'd like to hear Jean-Francois on this, the webcam (the ov772x sensor) is able to set the two parameters independently and the user can see the difference of either, is there a reason why we were preventing the user from doing so before? The third patch is the conversion of the ov534 subdriver to the v4l2 control framework, I tested the code with a PS3 Eye (ov772x sensor) and it works fine (now disabling automatic exposure works too, yay), maybe someone else can give it a run on a webcam with OV767x. NOTE: in patch 3, in sd_init_controls(), I left multiple checks if (sd-sensor == SENSOR_OV772x) just to preserve the order the controls were declared in struct sd, if you feel the order is not that important I can aggregate the checks, just let me know, it just looked neater to me this way. From a purely aesthetic point of view maybe the gspca mechanism of defining controls was prettier, more declarative, but the control framework really looks more correct even from userspace, qv4l2 can now display labels of control classes in tabs automatically while before we had empty labels, disabled controls in clusters work beautifully, and disabled controls with associated automatic settings can show the value calculated by the hardware on every update, very instructive if not super-useful. I'm glad to hear you like the control framework. Regards, Hans Thanks, Antonio Antonio Ospite (3): gspca - kinect: remove traces of gspca control handling gspca - ov534: make AGC and AWB controls independent gspca - ov534: convert to v4l2 control framework drivers/media/video/gspca/kinect.c |9 - drivers/media/video/gspca/ov534.c | 590 2 files changed, 261 insertions(+), 338 deletions(-) -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PATCH 0/3] gspca: kinect cleanup, ov534 port to control framework
Hi, the first patch just removes traces of the gspca control handling mechanism from the kinect driver; this driver does not have any controls. The change is trivial and can be applied right away, or postponed to when the gspca_main code is removed, you decide. The second patch removes the dependency between auto gain and auto white balance, I'd like to hear Jean-Francois on this, the webcam (the ov772x sensor) is able to set the two parameters independently and the user can see the difference of either, is there a reason why we were preventing the user from doing so before? The third patch is the conversion of the ov534 subdriver to the v4l2 control framework, I tested the code with a PS3 Eye (ov772x sensor) and it works fine (now disabling automatic exposure works too, yay), maybe someone else can give it a run on a webcam with OV767x. NOTE: in patch 3, in sd_init_controls(), I left multiple checks if (sd-sensor == SENSOR_OV772x) just to preserve the order the controls were declared in struct sd, if you feel the order is not that important I can aggregate the checks, just let me know, it just looked neater to me this way. From a purely aesthetic point of view maybe the gspca mechanism of defining controls was prettier, more declarative, but the control framework really looks more correct even from userspace, qv4l2 can now display labels of control classes in tabs automatically while before we had empty labels, disabled controls in clusters work beautifully, and disabled controls with associated automatic settings can show the value calculated by the hardware on every update, very instructive if not super-useful. Thanks, Antonio Antonio Ospite (3): gspca - kinect: remove traces of gspca control handling gspca - ov534: make AGC and AWB controls independent gspca - ov534: convert to v4l2 control framework drivers/media/video/gspca/kinect.c |9 - drivers/media/video/gspca/ov534.c | 590 2 files changed, 261 insertions(+), 338 deletions(-) -- Antonio Ospite http://ao2.it A: Because it messes up the order in which people normally read text. See http://en.wikipedia.org/wiki/Posting_style Q: Why is top-posting such a bad thing? -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html