I would wait.
This was posted on the CCD-NEWASTRO Yahoo Group:
Cause for hesitation at least.
Turbo
-----Original Message-----
From:
ccd-newastro@yahoogroups.com [mailto:ccd-newastro@yahoogroups.com] On Behalf Of Daniel Perry (californiastars)
Sent: Thursday, 18 October 2007 11:23 AM
To:
digital_astro@yahoogroups.com;
Canon_DSLR_Digital_Astro@yahoogroup s.com;
Yahoo - CCD New Astronomy
Subject: [ccd-newastro] Hutech Canon 40D - horizontal banding problem
Hi all,
I have a new Canon 40D that was modified by Hutech. In my first outing with it, I've noticed a rather severe horizontal banding problem. This isn't the "dark river" issue seen in many 20Ds. It's inconsistent banding throughout the image that seems to bleed some color along the bands. By "inconsistent"
I mean that the degree to which it's visible varies throughout the image; it's not uniform throughout the image but it IS present to some degree in the entire image and the same pattern is present in every image I took over the weekend.
I had a Hutech modified 20D for about 2 years and never saw a banding issue anywhere near as severe as this one. In fact, I went through some of my old images and majorly compressed the histograms and cannot see any banding issues even remotely similar to what I'm seeing in the 40D. I've provided some of those images for comparison purposes.
The images were taken at ISO 400 at an ambient temperature of about 60F. I took and applied 17 dark frames, 30 bias frames at 1/8000th of a second each, and flat frames at 1/15th of a second. The raw images exhibit the banding but it's not quite as apparent in a single raw image (but the same patterns are clearly present). My curve adjustments in these "processed"
images weren't all that aggressive. The camera was powered with the Canon AC adapter, the same one I used on my 20D. There's also a gradient in every image from upper left to lower right but I'm less concerned about that.
Here are the images I just took with the 40D. These were processed using IP's (v3 beta 9) automatic image set processing, saved to TIFF, then I adjusted the curves in Photoshop. That's the only processing applied to
these:
http://www.californiastars.net/galle...c2239.proc.jpg
http://www.californiastars.net/galle...c2177.proc.jpg
http://www.californiastars.net/galle...c1499.proc.jpg
http://www.californiastars.net/galle...gc1499.raw.jpg
The last one is a single 5 minute exposure with no darks, flats, or bias frames applied, just some curve adjustments.
And here are some images from my 20D with the levels massively compressed.
There are clearly some issues in these images but it's mostly uniform noise, which doesn't even appear problematic until the levels are compressed way beyond any normal processing parameters as I've done here. When processed with normal settings, nothing in the way of banding is even remotely detectable. The M51 images has the most noticeable features that might be referred to as "banding" but they're invisible when the image is processed normally, unlike the 40D images:
http://www.californiastars.net/galle...sed.ic1318.jpg
http://www.californiastars.net/galle...essed.m042.jpg
http://www.californiastars.net/galle...essed.b033.jpg
http://www.californiastars.net/galle...essed.m051.jpg
The only possible contributors I can think of, other than electronic issues within the camera itself, are external electrical noise, improper shielding of a cable near the camera or its USB cable, or something along those lines.
I'm going to run some tests using just a battery, a dedicated USB connection from the camera to the laptop (no hubs), and at different ISOs.
Hutech's response thus far has been that it's pretty normal, software can probably process it out, and to seek support in the Yahoo groups. I find that unacceptable seeing how this is not easily processed out and will prevent me from getting images even close to the quality I had with the 20D.
Unless I find an external culprit, this camera is going to either be exchanged or returned.
If anyone has any insights or suggestions, I'd appreciate hearing them.
Thanks,
Daniel Perry
http://www.californiastars.net/