Guennadi Liakhovetski wrote:
On Wed, 1 Jul 2009, Valentin Longchamp wrote:
Guennadi Liakhovetski wrote:
While trying all possible skipping / binning combinations of mt9t031 I
came across a problem, that in some configurations the sensor produces
regular horizontal stripes. They depend on window geometry, with some
skipping factors they can be eliminated by using properly aligned left
window border, but with some other AFAICS valid parameter combinations
stripes persist. And - they seem to depend on lighting conditions... I
think, I'll try to ask Aptina again... Or does anyone have an idea what I
might be doing wrong?
It may be completely unrelated, but we had quite similar problem with
our hardware. With a part of the image "saturated", we had some
artefacts (missing data in fact) on some horizontal lines: pixclk wave
was poor and the i.MX31 could not read data correctly. In order to
resovle this problem, we had to change the bus drivers (we now use some
SN74LVCH16244ADGGR that work well with the 3V3 signals from the camera).
Now even with high contrast and some parts saturated, pix clk looks nice and
we don't loose pixels anymore.
Hm interesting. What exactly did those missing pixels look like in your
case? In my case these are periodic horizontal lines of 1 or 2 (approx.)
pixels wide. Sometimes they are repeated one such stripe every x pixels,
sometimes they are repeated in pairs.
Mine were a bit different: they started at a the pixel where there was a
border between a saturated and unsaturated area and then the pixels were
missing until the end of the line. The line below usually was correct
(but not always) and it did happen again 2-3 pixels below if the border
was still present.
But since yours are dependant to lightning conditions and also happen on
horizontal lines, it may also be something related to what has happened
for us. I would suggest that you really control the pixclk signal.
Val
--
Valentin Longchamp, PhD Student, EPFL-STI-LSRO1
valentin.longchamp@xxxxxxx, Phone: +41216937827
http://people.epfl.ch/valentin.longchamp
MEA3485, Station 9, CH-1015 Lausanne
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
