Hot/dead pixels on the 5D - ISO setting implications

When I saw an obnoxious white pixel in the middle of the black nose of a Peke I was photographing (ISO 3200 1/125 sec.), and it persisted in the same spot of the sensor for several shot, I was ready to send it in to Canon (with B&H shutting down for Passover, I'd get it back faster from Canon NJ as a warranty repair than from B&H as an exchange). I downloaded the [URL=http://[URL]www.s​tarzen.com/imaging/dea​dpixeltest.htm]Dead Pixels Test to get an idea of how bad the problem was. Manual exposure, manual focus. I had LENR on, but as times were less than 1 sec., this should have been irrelevant. All shots were RAW converted to TIFF. What I found was very interesting

ISO 100  1/125    0 hot; 1/30    0 hot; 1/8    0 hot; 1/2    0 hot
ISO 200  1/125    0 hot; 1/30    0 hot; 1/8    0 hot; 1/2    0 hot
ISO 400  1/125   12 hot; 1/30   12 hot, 1/8   15 hot, 1/2   11 hot
ISO 800  1/125   39 hot; 1/30   51 hot; 1/8   52 hot; 1/2   17 hot (!)
ISO 1600 1/125   54 hot; 1/30   59 hot; 1/8   64 hot; 1/2   28 hot
ISO 3200 1/125 1405 hot; 1/30 1355 hot; 1/8 1398 hot; 1/2 1261 hot

1600+1st. 1/125  219 hot; 1/30  231 hot; 1/8   252 hot; 1/2   176 hot

Humm.
If you notice them in regular shots, and it's bothering you, send it to Canon Jamesburg and have them "map" them out to near neighbor pixels under warranty. I imagine any camera will have similar number of failed dead pixels. What's 250/12,000,000,000 as a %? Can't be a very big number.

I can't imagine using 3200 as a guide, since it is not really ISO 3200, but ISO 1600 pushed one-stop in camera.

I have an 2-year old 10D with a cluster of 5 dead pixels on the left side of frame. I know I should get them mapped, but it's $145 at Canon Jamesburg. When I "notice" it, I just lasso it, and be gone. Hard to spend $145 on it. But, a new 5D? Get them mapped since a new one is not going to be better.
Jack

What's interesting is the decrease in "hot" pixels as the shutter speed increased, which is counter-intuitive and suggests LENR kicks in earlier than Canon says, and the significantly better performance of ISO 1600 "pushed" in PP when compared to ISO 3200 (in-camera).

Canon NJ told me that acceptable hot pixels in the 5D and 30D were .01 percent of the total pixels. So I assume that 82,000 hot pixels would be acceptable to Canon for the 30D and 120,000 for the 5D. (Seems a bit high to me) Usually I don't even see hot pixels on my 5D or 30D until I get to 10 Second Exposures at ISO's above 1250.

I was told this number is also listed in the camera manual somewhere.

That's 'cause you are two orders off... it'd be 820 and 1200 respectively at 0.01% (or one in 10,000).

sheesh, i hate math!

im not understanding this, i see dead or hot pixels quite regularly with my 5d. like iso 400 sometimes but moreso over 800, they are all over the place. and sometimes show up as a white x i dont know what that is? can be bothersome for critical shots. should i go through the hassle of sending it in? i shoot almost everyday, being without a camera would be deathly.

How many do you have, what shutter speeds are you using, and have you turned LENR on?

is it "normal" to expect to PP hot pixels on a regular basis?
I never noticed any with my 300D in the past few years (probably more so due to lack of experimenting on different settings).
Having just got a 5D (last thursday), i was somewhat concerned last night when an under-exposed pic showed up some hot pixels (as confirmed using the deadpixel test above).
Given that obviously i will be trying not to under-expose pics as a rule, should i still expect to see hot pixels regularly?
Is this just something i need to deal with PP and consider it part of my workflow?

A few's normal. Consider, too, whether they'll show up in your standard print size, or after running a noise-reduction filter.

Jon, you didn't mention what was the threshold set to?
1400 hot pixels at 1/125s and ISO 3200 looks way to much for me, unless the threshold was set too low, so it wasn't only collecting distinct hot pixels, but bite the noise also.

NUMB3RZ

Did people do all this pixel peeping back in the film days as well? Or is it just easier now that we can zoom in to 1600%?

Saw this interesting link to try and solve the Hot/dead pixel problem.
Seems it works for some people.
http://forums.dpreview​.com …age=18151631&ch​angemode=1

-S

Well, you didn't have pixels back in the days of film did you

Dave