VENETIAN BLIND ARTIFACT IN CALIBRATED IMAGES

[Jay W. Butler]

I am using PI to calibrate and process images from an SBIG STXL-16200M camera. Most of my calibrated images have a “venetian blind” artifact consisting of horizontal alternating light and dark bands from top to bottom of the entire image. The effect becomes more pronounced with narrow band filters. The light bands are approximately 30 pixels wide interspersed with dark bands that are only four or five pixels wide. The boundaries between the bands are somewhat irregular and discontinuous.

I have read the single PI users’ forum post I could find on this subject: (http://pixinsight.com/forum/index.php?topic=8516.msg56447#msg56447) but did not find it helpful. I do not want (and am not competent) to go through a cumbersome process of making corrections by finding a PixelMath formula that will address the problem after the fact. I need to find a way to prevent this problem from happening in the first place.

1.   There is no evidence of the artifact in my raw .fits images.
2.   Nor is there is any evidence of the artifact in my raw flats.
3.   The artifact first appears in images which I have calibrated using the PI ImageCalibration process and before further processing.
4.   The artifact is most pronounced in narrow band images and those captured through my blue filter.
5.   There may be some connection to the percentage of full well capacity reflected in the flats used for calibration. (The camera has a FWC of 41000 ADU and I create my flats at about 20000 ADU.)

I have posted an example of a PI calibrated .xisf image showing this artifact in a DropBox folder at:

https://www.dropbox.com/sh/32eqcl8u0zvfyvy/AABmFzh9tdOxqtXr8YeV3s-Ya?dl=0

Please let me know how I can avoid creating this venetian blind artifact.

[pfile]

1st hopefully a mod can move this post out of this section of the forum because it's in the wrong place

2nd, this is a new sensor, right (meaning, it just came on the market recently)? the pattern looks to me to be related to the driver's configuration of the chip. it's not normal to have that huge white block on the side, and the artifact looks like some kind of a readout timing problem.

if i had to guess, i'd say that the artifact is there in the non-flattened image but is not visible upon application of STF because of the vignetting. maybe you can post the calibration masters and an uncalibrated light.

what software are you using to acquire the images, and do you have all the latest SBIG firmware and driver updates? what is the readout mode?

rob

[Jay W. Butler]

1. I didn't know there was some particular place to post this. Please tell me how I would know that and whether I can move it myself.

2. I have now put into my DropBox folder the bias, dark and Ha flat masters as well as the uncalibrated Ha frame of which I previously posted a link to the calibrated frame.. See https://www.dropbox.com/sh/32eqcl8u0zvfyvy/AABmFzh9tdOxqtXr8YeV3s-Ya?dl=0

3. I use TheSkyX Pro, latest version, to acquire the images and have installed and updated the SBIG 64-bit driver checker. I am confident I have the latest drivers from SBIG. I am not confident that they have provided a driver that works with the KAF-16200 chip.

4. I don't know what the chip's readout mode is and don't know how to find out. I attach the manufacturer's specifications

[Jay W. Butler]

I should add that I have now contacted SBIG and asked them to look at the issues of a possible defect in the chip or in their supplied driver.

[Greg Schwimer]

I have the FLI Microline variant with the KAF-16200 sensor.

There was a bug in the TSX X2 driver for the FLI cameras that caused me to capture unusable bias frames. Check here:

http://www.bisque.com/sc/forums/t/27775.aspx

The vertical stripe looks similar to what your flat looks like. Do you get the same effect if you calibrate without bias?

Also, I don't know what the shutter for the SBIG camera looks like, but I found that if I did not capture a flat of at least 3-4 seconds in length I got a notable shutter effect in the final stack. Probably not a factor with HA flats, but perhaps that is contributing.

Hope this helps...

Greg

[pfile]

1. I didn't know there was some particular place to post this. Please tell me how I would know that and whether I can move it myself.

2. I have now put into my DropBox folder the bias, dark and Ha flat masters as well as the uncalibrated Ha frame of which I previously posted a link to the calibrated frame.. See https://www.dropbox.com/sh/32eqcl8u0zvfyvy/AABmFzh9tdOxqtXr8YeV3s-Ya?dl=0

3. I use TheSkyX Pro, latest version, to acquire the images and have installed and updated the SBIG 64-bit driver checker. I am confident I have the latest drivers from SBIG. I am not confident that they have provided a driver that works with the KAF-16200 chip.

4. I don't know what the chip's readout mode is and don't know how to find out. I attach the manufacturer's specifications

on #1 it looks like it's been moved. originally it seems to have been posted in PixInsight Forum » PixInsight » Announcements » Image posting rules: Read before you post! but now it is in PixInsight Forum » PixInsight » General.

on #3b i hear you - i use an STT-8300M and for a while there the drivers/firmware were in really bad shape. i am not convinced the cyanogen thing was a positive development for SBIG.

the readout mode is something that can usually be controlled with the aquisition package - fast vs. slow (meaning the clock rate during readout. generally faster rates are great for focusing or framing but they have higher read noise)

based on the dark and bias in your zipfile there's gotta be something going on in the driver or the camera firmware. there's something wrong with those frames...

rob

[Jay W. Butler]

Thanks, guys.  As nothing has changed in my acquisition or processing software or workflow and as I did not have this problem with an identical camera (STXL-11002), only with a different chip, I am satisfied that this is a problem with the SBIG driver or camera firmware and am therefore addressing the problem with them. I'll let you know when we get it resolved.

[jkmorse]

If the SBIG Driver checker doesn't solve this, keep pressing SBIG (+1 on Rob's comment on the SBIG/Cyanogen merger).  I had a recent problem with two different sets of weird artifacts on a brand new STXL 6303E that took a trip back to the SBIG repair shop to get fixed and ultimately they had to put in a new chip and rewrite the driver.  The SBIG guys like Bill Lynch are great to work with, but I think the quality control on the manufacturing side and the driver side has suffered a hit.  The one that was a driver issue sounds a lot like the problem you are having.  They were able to send an updated driver but I am not sure they have released the patch yet for the public generally.  They will work to solve the issue (mine involved both) but keep pressing.

Best,

Jim

[Jay W. Butler]

I am embarrassed to say that I have now been able to trace the problem to the sets of bias and dark frames I created when I first received the STXL-16200 ten days ago. All of the subs in both of those two sets exhibit the venetian blind artifact. This morning I have taken new sets of both bias and darks and they are all free of this problem. I am unable to explain the source of the problem in the earlier sets as I took both with the same acquisition software (TSX Pro) and under similar conditions.
My apologies for wasting everyone's time on this.