[slightly OT] driving a Sony PAL CCD in progressive way

It's a bit off-topic here, sorry, but I think many of you have a good
knowledge of CCD sensors, and resources are so scarce...

I'm developing my second camera, and I need both good live video quality and
high vertical resolution still frames.
I need also high sensititivy (very low llight levels...). Easy, isn't it ?
:-)

Progressive scan CCDs as ICX204AK have too low sensitivity. Just tried.

The ICX419AKB (PAL, 1/2', 752x582) has a very good sensitivity and nice
round package.

Usually is interlaced scanned, binning two lines in each field; color mosaic
is made in a way that is possible to derive Y e chroma from binned G+Cy and
Mg+Ye cells.

Do you think it's possible to change vertical clock sequence in order to
avoid the binning, and have a progressive full frame scan ?

I need to sample each pixels without binning, at half the vertical rate, but
without interlaced scanning.

Given that I have a whole frame buffer, I could also live with interlaced
scanning, if all the pixels (of both fields) are exposed at the same time. I
can't deinterlace two fields taken at different times with the final quality
I need.

Of course I would halve the frame rate, and I would need an FPGA to generate
the timings and the color demosaicing.

But, even if the logic needed is complex, do you think it's possible ?

How could I tweak the vertical sequence ?

Has anybody tried something similar ?

I can send the datasheets if someone is interested, and share ideas, also.
Even samples :-)

Antonio Pasini wrote:
The ICX419AKB (PAL, 1/2', 752x582) has a very good sensitivity and nice
round package.

Usually is interlaced scanned, binning two lines in each field; color mosaic
is made in a way that is possible to derive Y e chroma from binned G+Cy and
Mg+Ye cells.

Do you think it's possible to change vertical clock sequence in order to
avoid the binning, and have a progressive full frame scan ?

The problem is, that the 419 has only the half number of vertical interline
registers, so you have to readout a full resolution pic in two sequences.


Given that I have a whole frame buffer, I could also live with interlaced
scanning, if all the pixels (of both fields) are exposed at the same time.

For that you need a mechanical shutter.
Without, the two half-frames would have different exposures.
That is only possible with progressive scan sensors, which have the full
number of vertical interline registers.
An exception is only for very long exposures, where the two half frames
have quite the same exposures.


But, even if the logic needed is complex, do you think it's possible ?

How could I tweak the vertical sequence ?

For a frame readout instead of a field readout with vert. binning,
you have only to disable one of the two readout pulses in the vertical
readout sequence.
In this case, the sn ratio is a bit lower, because of the less fullwell.


Has anybody tried something similar ?

Did an astro-camera with the ICX249AL (exview), the successor is the
429AL. The driving is quite the same.

I was looking for a high sensitivy (exview) progressive scan sensor by
myself, but Sony did it only to the rel expensive 285AL 2/3" 1.4mpix.

Jens