[libcamera-devel] libcamera on Librem 5, phase detection auto focus

[Jacopo Mondi]

Hi Pavel,

On Tue, Nov 22, 2022 at 12:46:29AM +0100, Pavel Machek via libcamera-devel wrote:
> Hi!
>
> > > > So far I'm working with contrast detection, but samsung chip should
> > > > support phase detection auto focus, too. So far I hacked
> > > > drivers/media/i2c/s5k3l6xx.ko not to filter out those bits and took
> > > > photos (and looked at them in darktable), but I don't see any obvious
> > > > "defects" indicating presense of focus data. Do you have any pointers
> > > > to more information by chance?
> > > >
> > > > I took deliberately defocused photo to make "defects" easier to see. I
> > > > guess I should take another one with less gain to keep other noise
> > > > down. Any ideas how to locate those bits?
> > >
> > > The PDAF pixels will be located at fixed positions in the pixel array,
> > > documented by the sensor vendor. Some sensors will process the PDAF
> > > pixels and provide the processed information in the CSI-2 embedded data.
> > > I don't have access to the sensor datasheet, so I can't help much I'm
> > > afraid :-(
> >
> > Yep, I think I'm figuring it out with some help.
> >
> > In case of L5, pixels are in the pixel array, and it is blue
> > pixels. So... it is easy to take blurry 13M photo of barcode (with
> > hacked driver, see above), remove red and green channels, enhance
> > contrast... and pattern emerges. Pairs of pixels emerge, and they
> > should be those AF pixels.
> >
> > Now I guess I'll need to select two lines of special pixels, compute
> > correlation between them, and adjust focus according to that.
> >
> > Does not sound too bad...? But if someone already has code / tools to
> > help, let me know.
>
> Ok, so I'm pretty sure I have right pixels, I'll need to compute the
> correlations, but it no longer looks as easy as it looked before.
>
> Am I right that I should be able to do the AF using subset of special
> pixels? I'm thinking about single horizontal line in the middle of screen...

Are you still talking about PDAF (for which I'm afraid I cannot help
with reverse engineering) or are you looking at contrast-based
algorithms here ?

>
> I pushed work-in-progress code to
> https://gitlab.com/tui/millipixels/-/merge_requests/new?merge_request%5Bsource_branch%5D=pdaf
> .

br0ken it seems :)

>
> Best regards,
> 								Pavel
> --
> People of Russia, stop Putin before his war on Ukraine escalates.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.libcamera.org/pipermail/libcamera-devel/attachments/20221122/b5f6dae0/attachment.sig>