PDA

View Full Version : Dead Pixel on JPEG But Not RAW


martinledford
03-19-2005, 10:44
I purchased an RD-1 this week and shot RAW format for two days. Today, I shot B&W JPEG (H) and noticed a dead pixel in the exact location on each B&W JPEG image. However, when I switched back to RAW the dead pixel was no longer there. Should I be concerned about this and return for an exchange? Or is this related to some internal algorithm regarding noise reduction?

Thanks,
Martin

Sean Reid
03-19-2005, 12:20
If everything else is working well on the camera, I wouldn't return it for just the one dead pixel.

Cheers,

Sean

FrankS
03-19-2005, 14:11
Return it and get a film camera.

martinledford
03-19-2005, 15:46
Return it and get a film camera.
Hey, Frank:

Thanks for the reply. I already own several film cameras (Leica M7, Leica CM, Canon 1v, Hasselblad H1).

Martin

laurentvenet
03-19-2005, 15:51
maybe I would return it thinking of the high price of it, it should be flawless

FrankS
03-19-2005, 15:53
Martin, if I were you I'd read Jorge's post on the slow death of digital. ;) (Just funning with ya.)

Sean Reid
03-19-2005, 16:06
Martin,

Frank wasn't serious. He's just messing with you.

Sean

Jim Watts
03-20-2005, 01:58
Martin,
If you have read the other posts about Stuck/Hot pixels you will see that this is not uncommon with the R-D1. I have three stuck pixels and personally have not seen an R-D1 without at least one. Given that many exchanged cameras have been worst in this and other respects I would keep the one you have unless there are any other serious faults.

Sean Reid
03-20-2005, 02:19
If one has an R-D1 with accurate rangefinder focusing, a viewfinder mounted truly square in the camera, just one hot pixel and no other faults...hold onto it for dear life. As much as I love this camera (and I do) one cannot deny that some of them have had defects of one kind or another. But a good copy of an R-D1 is certainly worth having. Should there be better QC for $3000? Absolutely. But...the situation for now is what it is and there are some very good examples of the R-D1 out there - I have one of them.

Jim,

What do you shoot that makes having three bodies desirable? news? weddings?

Cheers,

Sean

Jim Watts
03-20-2005, 02:36
Sean,
Guess I should always get someone to proof read my posts so they make sense. I have now edited to read three "stuck pixels". I have checked a total of three R-D1's at my dealer and on the Epson stand at Focus on Imaging and all have had stuck/hot pixels when the iso is pushed to 1600.

Now the thought of owning three R-D1's each with a different "Star" lens from your ongoing tests. :) :) :) It's just the money thats the problem. :bang:

Thank's for pointing out the lack of clarity.

Jim Watts
03-20-2005, 02:53
I have just read Frank's thread on Good things coming in pairs.

Perhaps I do need 3 R-D1's just to up the ante! :p :)

Sean Reid
03-20-2005, 03:07
Based on what I know so far, if money were no object, I'd indeed carry three R-D1 bodies for working in fast changing circumstances:

R-D1 with Voigtlander Ultron 28/1.9 $3000 + $440.00
R-D1 with Leica 35/1.4 ASPH $3000 + $2795.00
R-D1 with Leica 50/1.4 ASPH $3000 + $2495.00

Total cost: $14, 730 (although I've already got the first body and the Ultron so that's a start)

One of my tasks today is to start to get a feel for how close the Nokton 1.5 comes to the Leica. The CV 35/1.2 is a good lens but, so far, it's not a match for the Leica 35/1.4.

Cheers,

Sean

martinledford
03-21-2005, 08:50
Hey, Guys:

Thanks for your responses to my question. However, I am still confused as to why when shooting 1600 I would see a hot/dead pixel on a the JPEG image but not on the RAW image at 1600. These shots are the same image at the same time.

Thanks,
Martin

pfogle
03-21-2005, 09:09
Martin, can you please clarify? Do you see the 'non'-hot pixel on a tiff or a jpeg you've processed from the RAW?

I imagine that the hot pixel is below the noise threshold for the RAW conversion (there will be some background noise reduction), but is seen as a 'feature' by the on-board jpeg algorithm, which can then be amplified in the compression process.

I have a 1-pixel hotspot in RAW, which becomes a 5-pixel star in jpeg. Not quite the same case, but perhaps a similar process?

cheers
Phil

martinledford
03-21-2005, 09:49
Martin, can you please clarify? Do you see the 'non'-hot pixel on a tiff or a jpeg you've processed from the RAW?

I imagine that the hot pixel is below the noise threshold for the RAW conversion (there will be some background noise reduction), but is seen as a 'feature' by the on-board jpeg algorithm, which can then be amplified in the compression process.

I have a 1-pixel hotspot in RAW, which becomes a 5-pixel star in jpeg. Not quite the same case, but perhaps a similar process?

cheers
Phil
Hey, Phil:

Thanks for the response. My test involved shooting a dark wall twice. The first time using RAW at 1600. I then use the Epson RAW converter in PS (Mac version) and convert with standard settings 16-bit into PS. Then, I shoot the same wall scene with JPEG High @ 1600. I use "standard" settings in the camera set-up. I've also used NR set to highest setting. I open this image in PS as well. When I compare the RAW version to the JPEG version, I see a hot spot (same place every time) in the JPEG version but it's not there in the RAW version. I find it odd that the RAW version never shows the hot spot but the JPEG does.

Thanks for your help.
Martin

pfogle
03-21-2005, 10:09
well, I think you're lucky it's only in the jpg! It's a drag though, when you need jpg. Still, they ususally aren't visible in prints.

As far as NR goes, I'm not really sure, but I think that the noise reduction in the built in film settings are for shutter speeds of 2 secs an over, using dark field subtraction. At least, I think that's true if the NR is set to 'high'. Obviously, they'd use a different method in PhotoRaw.

Out of interest, do you see the hotspot if you make a jpg from the RAW, rather than 16bit tiff?

cheers
Phil