HELP - Should I return G2

[rmjanzen]

Hi everybody,


I've got a question about hot/dead/stuck pixels and when to return your camera. I bought a canon G2. When i took an night shot me and my girlfriens
notices a star we couldn't place. Having read a lot off posts here i thought it might be a hot/dead/stuck pixel. I took a picture in a dark room(with the lens covered) at several iso settings, apeture settings and exposure settings. On that picture i found it to. It seems the pixel is only noticable in the original shot and only in
night shots

Black shot (hot pixel is on the right side of the picture):

http://image.pbase.com/u4/rmjanzen/upload/744816.IMG%5f0173.JPG

I tried the starzen test. I took the following procedure:

- all pictures are taken in a dark room with lens covered;
- pictures taken in TV mode;
- apeture on 2;
- tried iso 50,100 and 200;
- tried tv 1/30 , 1/4, 0"3 , 0"6 and 1 sec;
- setting on RAW;
- converted to TIFF 8-bit;
- Raw image conversion settings - use false color filter disabled.

i got the following results:

- no stuck/dead pixels;
- all hot pixels centered(until 9) around pixel 2118x538;
- when i got more than 9 hot pixels, they showed up in other places.

Pixel 2118x538 gave the following result:

TV #hot LUM

iso 50
1/30 0
1/4 0
0"3 0
0"6 1 61
1"0 2 79

iso 100
1/30 0
1/4 2 76
0"3 2 86
0"6 5 110
1"0 9 126

iso 200
1/30 3 95
1/4 9 125
0"3 9 131
0"6 11 146
1"0 17 146

I can change the camera for a new one (i have to return mine and i will get another in a week). Reading posts down here i'm a bit scared that this spots are normal and i will get back one with more of these pixels. Can you give me some advice on what to do. Exchange it or not (or maybe even send mine in for repair instead
of exchange). And have i used the correct method to find these results. I feel that i should have one without at this price.

With kind regards,

René

[kd6lor]

My concern would be if I saw them in the picture when actually using the camera. Since you do, I would return the camera.


FYI, in my G2, I ran the test and found at 4 seconds, ISO 50, I had 1 pixel with a 5, 2 pixels with a 4, and a handfull of 2's and 3's. I see no effect of this in my images.


Paul

[rmjanzen]

kd6lor wrote:
My concern would be if I saw them in the picture when actually using the camera. Since you do, I would return the camera.


FYI, in my G2, I ran the test and found at 4 seconds, ISO 50, I had 1 pixel with a 5, 2 pixels with a 4, and a handfull of 2's and 3's. I see no effect of this in my images.


Paul



I can't place your findings. I guess you mean yoour camera has a better result then mine ??

If got a example up here where it's visible:

http://image.pbase.com/u4/rmjanzen/upload/744762.IMG%5f0138.JPG

René

[kd6lor]

Rene, beautiful night shot. To be honest, I did see the hot pixels on the right middle side of your picture before I looked at the prior link where it was even more easily seen in the dark field picture. If it were my camera, I would return it or prepare to edit each of my night pictures for that area. While editing is not that big of a problem, if the camera is new and it is possible to return and exchange, I would.

My test was not as thorough as yours. I tested only one ISO (50) and three shutter speeds. The longest was 4 seconds and when I ran the DeadPixelTest program it showed no hot or dead pixels until I lowered the hot pixel threshold from 60 to 2. At a very low threshold, it found some hot pixels, but the hottest was a 5.

So, to answer your question, while we didn't run the test the same way, with a 4 second exposure, my camera has significantly fewer problem pixels. My camera may go crazy with bad pixels at higher ISO ratings, but I didn't test the higher ISO's because I only shoot at 50.

Good luck with your decision, and by the way the night shot was beautiful. A testimony to your skills and the quality of the camera. If not possible to return, please know that the picture can always be edited to remove the hot spots. I remember seeing a program on the WWW that would allow you to map an area and remove the spots automatically, but I can't find my link.

Paul Jaruszewski

[rmjanzen]

kd6lor wrote:
My test was not as thorough as yours. I tested only one ISO (50) and three shutter speeds. The longest was 4 seconds and when I ran the DeadPixelTest program it showed no hot or dead pixels until I lowered the hot pixel threshold from 60 to 2. At a very low threshold, it found some hot pixels, but the hottest was a 5.


Have you tried the test with for instance 0"3, 0"6 and 1 second. Because if you have a G2 the automatic noise reduction kicks in with slower speeds. If i check my picture with an exposure of 2 secs i get none (as everybody shoot get.)

Hope to hear back from you soon...

btw thanks for your comments...other pics are on www.pbase.com/rmjanzen

regards,

René

[kd6lor]

At one second, which is before the noise reduction filter from what I understand, I have 7 "hot" pixels. One has a lumenescence of 5, the others have 3's. This was done with a threshold of 2 for the "hot" pixel setting on the software. My understanding was that the noise reduction software kicked in at 1"4. But I may be wrong.

If you would like, I can email the RAW file for any of the three dark field images I sent you and you can inspect them yourself.

Paul