[Elphel-support] 10359
Andreas Bean
office at beanbox.com
Sat Nov 21 16:06:57 PST 2009
Hello Oleg!
> Hi Andreas,
>
> Alternating mode with buffering ist still not working. I'm
> attaching the
> screenshots.
> Should I update the x359.bit from cvs? Have you comitted your latest
> verstion?
>
>
> I haven't found the solution yet - I tried various variants and, for
> example, if I output only the buffered frame - the picture is ok - the
> problem is in the frame latencies, I think - for the direct image I
> use sync signals from the sensor and for the buffered I generate these
> signals myself - and the mistake is here.
> I also have a verification testbench with all the models in this
> system and it doesn't have this bug unfortunately.
> I'm working on it, but a bit busy with other things.
Ok, I understand that. So I will have to wait until you can fix the bug.
>
> Commited to cvs the version you already have.
>
> Have you spoken to Andrey about the new mode?
>
>
> Yes, he said it's possible of course but it'll be hard to make an
> ideal alignment because the images can be shifted vertically and they
> also have a distortion near edges. The best place for doing it is a pc.
Just to make sure that we speak of the same mode I will recapitulate it
here. Store the image of both sensors in the ddr-ram. Send a combined
frame to the 10353 board, where
the second image is right of the first. This solves a) the problem that
you can't figure out which frame corresponds to what sensor and b) it
also solves your timing problem I think. I'm not talking of any
geometric reconstruction.
>
> First, why do I have to enter 976, i.e. +16px in frame size input
> boxes.
> If I don't do that, the buffered picture is vertically unstable.
> What is the correct frame size which should be entered? Although the
> picture is vertically stable it moves from time to time
> horizontally, is
> brighter and the brightness ist unstable. Whats the reason for that?
>
>
> You need to enter them because 10359 should know the frame dimensions
> for correct storing and output the buffered image. And it doesn't -
> this input boxes only program 10359 and don't affect the sensor. These
> input boxes must be confusing.
> The sensor frame size (and the size for the comressor also) is set
> through the camvc.
>
Whats the reason for the unstable brightness and why do I have to enter
a size of +16px to get a vertically stable picture?
Thanks,
Andreas
> Oleg
>
> Hello Oleg!
>
> Alternating mode with buffering ist still not working. I'm
> attaching the
> screenshots.
> Should I update the x359.bit from cvs? Have you comitted your
> latest
> verstion?
>
> Have you spoken to Andrey about the new mode?
>
>
> Hello Oleg!
>
> I have done further tests with the following results:
>
> 1) set triggered mode
> 2) set the horizontal resolution to 1/2 -> 1296px
> 3) set the vertical resulotion to 1/2 -> 960px
> 4) tested alternation mode -> works fine
> 5) switched to buffered mode with resolution 1296 x 976
>
> First, why do I have to enter 976, i.e. +16px in frame size input
> boxes.
> If I don't do that, the buffered picture is vertically unstable.
> What is the correct frame size which should be entered? Although the
> picture is vertically stable it moves from time to time
> horizontally, is
> brighter and the brightness ist unstable. Whats the reason for that?
>
> Best wishes,
> Andreas
>
> Oleg Dzhimiev schrieb:
> > Hi Andreas,
> >
> >
> > During the last days I have mounted the 10359 board inside
> the camera
> > below the 10353 board. Two sensor boards are connected via a 6
> > inch flex
> > cable. With your description I was able to switch the board to
> > triggered
> > mode and I was able to test the first sensor. Additionally
> with the
> > phase adjust page I was able to remove wrong pixels which were
> > caused by
> > the longer flex cable. So the first sensor board is working
> fine.
> > I was
> > not able to switch to the second one, simply because I don't
> know
> > how to
> > do that.
> >
> >
> > You can switch to the other channel in two ways:
> > 1. http://192.168.0.9/359/reg_write.php?adr=0x835&data=2
> <http://192.168.0.9/359/reg_write.php?adr=0x835&data=2>
> > <http://192.168.0.9/359/reg_write.php?adr=0x835&data=2
> <http://192.168.0.9/359/reg_write.php?adr=0x835&data=2>> - that
> will be
> > the second channel (J3)
> > http://192.168.0.9/359/reg_write.php?adr=0x835&data=4
> <http://192.168.0.9/359/reg_write.php?adr=0x835&data=4>
> > <http://192.168.0.9/359/reg_write.php?adr=0x835&data=2
> <http://192.168.0.9/359/reg_write.php?adr=0x835&data=2>> - 3rd channel
> > (J4) - the most distant from the cable to 10353
> >
> > 2. or if you have firmware >8.0.5.2 you can try
> > http://192.168.0.9/359/10359_controls.html - and there you can
> find a
> > field for switching a channel - 1,2 or 4, and you also can try to
> > autoadjust phases.
> >
> > What I want to do is to record single framers which are captured
> > simultaneously and then written separately to disk using the
> > camogm
> > utility at a resolution of 1280x1024 at 40fps. Can you give
> me a short
> > description what I will have to do to achieve that?
> >
> >
> > Alternating frames 1280x1024 at 20fps from each channel that in sum
> will
> > give 40fps? Here I need to make a small revision of my previous work
> > while you are adjusting phases I checked CVS - and there's at least
> > one script is missing.
> > Basically, what you'll need to do is:
> > 1. Switch to broadcast channel (0x7) to program both sensors to the
> > resolution you need.
> > 2. Put sensors to triggered mode
> > 3. Switch on the alternation mode with buffering on 10359.
> > 4. and then use camogm to record frames
> >
> > Oleg
> >
> ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Support-list mailing list
> > Support-list at support.elphel.com
> <mailto:Support-list at support.elphel.com>
> >
> http://support.elphel.com/mailman/listinfo/support-list_support.elphel.com
> >
>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: office.vcf
Type: text/x-vcard
Size: 233 bytes
Desc: not available
URL: <http://support.elphel.com/pipermail/support-list_support.elphel.com/attachments/20091122/6b1d3413/attachment.vcf>
More information about the Support-list
mailing list