RE: [PATCH v4 00/12] Intel IPU3 ImgU patchset

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Mauro and all,

> > Subject: RE: [PATCH v4 00/12] Intel IPU3 ImgU patchset
> >
> > Hi Mauro,
> >
> > > > > Subject: Re: [PATCH v4 00/12] Intel IPU3 ImgU patchset
> > > > >
> > > > > Hi,
> > > > >
> > > > > Em Fri, 17 Nov 2017 02:58:56 +0000 "Mani, Rajmohan"
> > > > > <rajmohan.mani@xxxxxxxxx> escreveu:
> > > > >
> > > > > > Here is an update on the IPU3 documentation that we are
> > > > > > currently working
> > > > > on.
> > > > > >
> > > > > > Image processing in IPU3 relies on the following.
> > > > > >
> > > > > > 1) HW configuration to enable ISP and
> > > > > > 2) setting customer specific 3A Tuning / Algorithm Parameters
> > > > > > to achieve
> > > > > desired image quality.
> > > > > >
> > > > > > We intend to provide documentation on ImgU driver programming
> > > > > > interface
> > > > > to help users of this driver to configure and enable ISP HW to
> > > > > meet their needs.  This documentation will include details on
> > > > > complete
> > > > > V4L2 Kernel driver interface and IO-Control parameters, except
> > > > > for the ISP internal algorithm and its parameters (which is
> > > > > Intel proprietary
> > IP).
> > > > >
> > > > > Sakari asked me to take a look on this thread, specifically on
> > > > > this email. I took a look on the other e-mails from this thread
> > > > > that are discussing about this IP block.
> > > > >
> > > > > I understand that Intel wants to keep their internal 3A
> > > > > algorithm protected, just like other vendors protect their own
> > > > > algos. It was never a requirement to open whatever algorithm are
> > > > > used inside a hardware (or firmware). The only requirement is
> > > > > that firmwares should be licensed with redistribution
> > > > > permission, ideally merged at linux-firmware
> > > > git tree.
> > > > >
> > > > > Yet, what I don't understand is why Intel also wants to also
> > > > > protect the interface for such 3A hardware/firmware algorithm.
> > > > > The parameters that are passed from an userspace application to
> > > > > Intel ISP logic doesn't contain the algorithm itself. What's the
> > > > > issue of documenting the meaning of each parameter?
> > > > >
> > > >
> > > > Thanks for looking into this.
> > > >
> > > > To achieve improved image quality using IPU3, 3A (Auto White
> > > > balance, Auto Focus and Auto Exposure) Tuning parameters specific
> > > > to a given camera sensor module, are converted to Intel ISP
> > > > algorithm parameters in user space camera HAL using AIC (Automatic
> > > > ISP
> > Configuration) library.
> > > >
> > > > As a unique design of Intel ISP, it exposes very detailed
> > > > algorithm parameters (~ 10000 parameters) to configure ISP's image
> > > > processing algorithm per each image fame in runtime. Typical
> > > > Camera SW developers (including those at
> > > > Intel) are not expected to fully understand and directly set these
> > > > parameters to configure the ISP algorithm blocks. Due to the
> > > > above, a user space AIC library (in binary form) is provided to
> > > > generate ISP Algorithm specific parameters, for a given set of 3A
> > > > tuning parameters. It significantly reduces the efforts of SW
> > > > development in ISP HW
> > > configuration.
> > > >
> > > > On the other hand, the ISP algorithm details could be deduced
> > > > readily through these detailed parameters by other ISP experts
> > > > outside
> > Intel.
> > > > This is the reason that we want to keep these parameter
> > > > definitions as Intel
> > > proprietary IP.
> > > >
> > > > We are fully aware of your concerns on how to enable open source
> > > > developers to use Intel ISP through up-streamed Kernel Driver.
> > > > Internally, we are working on the license for this AIC library
> > > > release now (as Hans said NDA license is not acceptable). We
> > > > believe this will be more efficient way to help open source developers.
> > > >
> > > > This AIC library release would be a binary-only release. This AIC
> > > > library does not use any kernel uAPIs directly. The user space
> > > > Camera HAL that uses kernel uAPIs is available at
> > > > https://chromium.googlesource.com/chromiumos/platform/arc-
> > > > camera/+/master
> > > >
> >
> > The AIC library (in binary form) is available here.
> > https://storage.googleapis.com/chromeos-localmirror/distfiles/intel-3a
> > -libs-
> > bin-2017.09.27.tbz2
> >
> > Licensing information can be found in ./LICENSE.intel_3a_library file
> > after unzipping the tar file.
> >
> > >
> > > Just pinging to know your thoughts on this.
> > >

To follow up further on this, we have posted the preliminary documentation on
bnr (bayer noise reduction) parameters as below.

https://www.mail-archive.com/linux-media@xxxxxxxxxxxxxxx/msg128683.html

We would like hear your early feedback and apply the same towards the
documentation of remaining IPU3 parameters.

Thanks
Raj




[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]
  Powered by Linux