Mirage Recorder

[Take Vos]

Jose,

Are you sure your camera doesn't use some factory calibration data to compensate for non uniformity?
In any case I have contacted the reseller of the Pike, I wonder what they will reply.

Cheers,
Take

[Take Vos]

Hi,

I contacted the dealer, and they have a firmware update for me to do.
So with a little bit of luck my fixed pattern noise will be over soon.

Now I'll have to get my hands on Windows Vista, as I can not update the firmware with Mac OS X. A little bit sad that I will have to get an OS just to upload firmware :-(

Cheers,
Take

[Take Vos]

Hi, I measured the linearity of the sensor in the pike.
It seems that any value below 100 is trash, but above that it is linear. I guess it would be possible to use two flat field images to compensate for anything above 100. But I am scared that it won't look right, when clamping that much data.

here are the results from 10 neighboring green pixels in a single column. every forth pixel is to hot (thus three pixels in this case).

http://www.vosgames.nl/products/Mira...lLinearity.png

[Take Vos]

I've made an other measurement using the camera flash light (in continues mode) of my mobile phone. This makes for a much more stable measurement.

You can see here that 7 pixels are pretty much linear. But three pixels are slightly parabolic. I may be able to compensate for it.

I guess I will need to build some form of calibration light that can be screwed on the camera, to make it easy to do accurate automatic calibration.

http://www.vosgames.nl/products/Mira...Linearity2.png

[Take Vos]

It has been some time since I last posted, but I have an update.

I seem to have solved the dark pixel non uniformity, using a per-pixel 7 point lookup table with cubic interpolation.

However there are still some bugs in my code so I can not let you see a good image just yet.

Cheers,
Take

[Take Vos]

Hello everyone,

I've been able to fix most of the problems now. Here is a picture of the results of today.
http://www.vosgames.nl/products/Mira...oodPicture.png

This algorithm includes bad pixel removal, marking of bad pixels is depended on brightness. So in dark patches more pixels are removed, this is to solve unfixable non-linearities.

It seems to not have solved all the double tap problems yet, which it should, maybe I need a more accurate LUT table at higher brightness levels.

Cheers,
Take

[Jose A. Garcia]

That image looks just great Take! Keep it up!

[Solomon Chase]

Great job, thats a really smooth image :)

[Take Vos]

Hello everyone,

I've made a new picture, there isn't much sun as it is pretty early in the morning and overcast.

http://www.vosgames.nl/products/Mira...cingSolved.png

Here you can see that there is no longer any visible difference between the left and right side of the image.

I now use a 15 point LUT and separate average black left for the left and right side of the picture. The black level is first subtracted from the image before the 15 point LUT is applied. The black level does drift and right now you will have to manually (cover the lens, press button) measure this before each take. I hope with new firmware on the Pike I will be able to access the left and right light shielded pixels.

If one would look closely to flat field target with a flashlight pointing to it, I am seeing color banding. I am not sure why this is caused, maybe by the sharp edges in the LUT which would be solved by using cubic interpolation instead of linear interpolation.

I also like to mention that the 15 point LUT will now pull the pixel values to the theoretical linear photo response. I do this according to this formula:
when average intensity is around 90 %:
intensity_per_second = average_intensity[90] / shutter_time[90]

for all the measurements done I will calculate 0% <= k <= 100%:
theoretical_intensity[k] = intensity_per_second * shutter_time[k]


Cheers,
Take

[Take Vos]

I've made a new picture that was better calibrated.

http://www.vosgames.nl/products/Mira...ingSolved2.png

I've done some white balancing and set the black level lower (sun was shining into lens).

[Solomon Chase]

Originally Posted by Take Vos

I've made a new picture that was better calibrated.

http://www.vosgames.nl/products/Mira...ingSolved2.png

I've done some white balancing and set the black level lower (sun was shining into lens).

The white seems to be clipping a bit too much, like on bike lady's shirt. I did some color grading with your first picture, pushed it pretty far... (see attached JPG)

[Take Vos]

Originally Posted by Solomon Chase

The white seems to be clipping a bit too much, like on bike lady's shirt. I did some color grading with your first picture, pushed it pretty far... (see attached JPG)

Yes, I noticed the ladies shirt, I was not really paying to much attention (still have to add a zebra kind of thing). But I wanted something a little bit more bright than the image you color graded.

As you noticed it had still a lot of latitude left. You may also have noticed the amount of black offset, there was quite a bit of light falling into the lens then as well, and the calibration I did was done very shoddy (The white target was not evenly lit).

The second image should already be much better, even if it is clipping a little bit to much. Although it looks to me that this camera somehow handles white clipping more gracefully as the DV cameras I am used to.

although that corner stone for that arced window, in the foreground, is not really pretty.

strange thing is that I seem to have compensated for what normally would be considered dead pixels (there are around 8 bad pixels in this image according to how the manufacturer would measure them, and really visible on an uncompensated image). I am doing statistical analysis of each pixel after compensation compared to the theoretical value and I come up with zero bad pixels for an average light intensity above 10%.

[Solomon Chase]

Originally Posted by Take Vos

Yes, I noticed the ladies shirt, I was not really paying to much attention (still have to add a zebra kind of thing). But I wanted something a little bit more bright than the image you color graded.

As you noticed it had still a lot of latitude left. You may also have noticed the amount of black offset, there was quite a bit of light falling into the lens then as well, and the calibration I did was done very shoddy (The white target was not evenly lit).

The second image should already be much better, even if it is clipping a little bit to much. Although it looks to me that this camera somehow handles white clipping more gracefully as the DV cameras I am used to.

although that corner stone for that arced window, in the foreground, is not really pretty.

strange thing is that I seem to have compensated for what normally would be considered dead pixels (there are around 8 bad pixels in this image according to how the manufacturer would measure them, and really visible on an uncompensated image). I am doing statistical analysis of each pixel after compensation compared to the theoretical value and I come up with zero bad pixels for an average light intensity above 10%.

Yeah, Your 2nd image (with the clipped highlights) has tons of latitude and detail in the blacks, leading me to believe that you can underexpose a good bit without worrying about noise.

I'm really interested in getting a Pike 210c, as I have some ultra fast c-mount lenses made for a 1" image sensor. (Canon f0.95, Fujinon F0.85, Schneider f0.95, and some others) Would definately give some beautiful shallow DOF shots.

[Solomon Chase]

What is the maximum resolution you are getting right now at 14-bit, 24fps?
I read that AVT is releasing firmware with 12-bit support for pike 210c. What kind of resolution at could you get at 12-bit instead of 14-bit?

[Take Vos]

Hello Solomon,

I wish I had one of these fast lenses, I only have a 4.5, and I can not really use it in my room with just the lights. But it looks like the depth of field I get with this old SLR minolta lens is pretty nice.

As for the firmware, I was aware of this and I have requested it.

With 14 bits (16 bit transfer) I can do the following resolution:
1800 x 750

With 12 bits it should do:
1920 x 800 (maximum 1920 x 930)
With 12 bits and the larger Pike:
2048 x 850 (maximum 2048 x 870)

I want to get access to the light shielded columns, but for both these resolutions you would then still be able to get the full 2.40:1 ratio.