Get 10% Off Eversure Camera Insurance Today!

*ist D bad pixels


malcolmk

Link Posted 16/10/2003 - 08:56
As revealed in another thread, I'm on my second *ist D, the first had a bad clock. The current one has three annoying stuck pixels, one red, one blue, and one white. These show at all normal shutter speeds, I've not tried excessively long exposures where these things reputedly get worse.

Shots I still have from my first *ist D are entirely free of bad pixels. What's other users experience with this? Should I take the camera back and insist on another exchange?

Example image at http://www.malcolmknight.co.uk/lightstraw/tower.jpg look in shadowed paved area.

Don't look at it without a broadband connection! 3MB

MattMatic

Link Posted 16/10/2003 - 11:39
Malcolm,

Thankfully I don't have any (yet). All digicams have issues with stuck pixels (yes even the Canons!). Some of the cameras can detect a stuck pixel, or profile their own sensor and map it out. AFAIK, pixels can grow defective too, so it's not a case of just getting a 'perfect' sensor.

Note for Pentax: Couldn't you do that in the camera? Have a "sensor profile" mode after cleaning that checks the pixels and maps them out before JPEGing?

As the pixels are unusually bright, they are easily detectable. I have seen a Photoshop filter floating around to automatically map these pixels out. Obviously if the image is a JPG then the area of defect can appear larger because of JPG haloing.

BTW, I presume you've given the sensor a good puff of air? I bought a supersized Hama puffer from 7dayshop which is fantastic, and cost 5.

If I come across the filter page, I'll post a link

Matt
http://www.mattmatic.co.uk
(For gallery, tips and links)

MattMatic

Link Posted 16/10/2003 - 12:07
Malcolm,

Here's a link: http://www.starzen.com/imaging/ There's a tester for dead pixels for free. (Found the link from here: http://www.kernphoto.com/Resources.html - excellent list)

Some say use dust & scratches filter, others suggest NeatImage, some say return the camera if it's in warranty

It's a pain, but I would return the camera.

Matt
http://www.mattmatic.co.uk
(For gallery, tips and links)

malcolmk

Link Posted 16/10/2003 - 13:50
I've taken a number of test shots today, plain subjects (eg. newsprint) calculated to show up coloured specks. No bad pixels revealed themselves. Today's shots have been RAW or TIFF, the bad ones were all highest quality JPEG and all in the same session (no camera timed off between them).

This subject is obviously more complex than it looks. I'll be doing some more tests but at present I'm not so concerned about it.

malcolmk

Link Posted 16/10/2003 - 13:55
Sorry Matt, didn't notice your second reply until after I replied to the first. Looks useful, I'll look into it after doing some JPEG tests.

MattMatic

Link Posted 16/10/2003 - 15:15
Malcolm,

Hot pixels occur because of current leakage in the CCD, ie electrical noise larger than signals created by the image. It is very temperature dependent too

You will also find that ISO rating will affect it, as does shutter speed. As a rule of thumb, the colder it is, the faster the shutter speed, and the lower the ISO rating the lower the probability of hot pixels (and indeed noise). Conversely, the worst situation would be: ISO3200 on a hot day with a long shutter speed (and no in-camera sharpening).

Hope that helps

Matt
http://www.mattmatic.co.uk
(For gallery, tips and links)

malcolmk

Link Posted 16/10/2003 - 19:01
I tried your dead/hot pixel tester program, carefully following the instructions. TIF file tests of the 'black' files (slow shutter speed/ISO200) revealed no problems. JPEG files (the help file predicted these may appear noisy) produced the expected long list of hot pixels.

I repeated the shot of the BT Tower this afternoon in TIF mode (1/350th 200ISO) and the three coloured blobs are clearly visible in the same place as on the first JPG shot (1/2000th ISO800).

Where do we go from here? Collect a third *ist D.

MattMatic

Link Posted 17/10/2003 - 11:08
Malcolm,

Looking at your JPG again, it's peculiar. Have you tried overlaying the TIFF on top of the JPG to confirm they are exactly the same pixels? I would also say to try turning the camera 180 degrees and taking another shot - I wonder if it's because those pixels are in deep shadow?

Also, what happens if you manually set the same exposure but put the lens cap on? Same pixels or nothing?

There's something not right, but it doesn't seem clear-cut

Matt
http://www.mattmatic.co.uk
(For gallery, tips and links)

malcolmk

Link Posted 17/10/2003 - 14:24
I've been refraining from further posts because I keep doing more tests. Dark subjects certainly appear to be a key factor.

I have established that there are four sensor problem areas and that the file format, ISO rating, shutter speed (short of burning everything out!) all make no difference. The four areas are always in the same place and I can now reliably go to three of them on every suitable picture. The fourth is probably a single pixel and quite hard to see on my 17" monitor. The others are small clusters of pixels all permanently blue (2 with 9 pixel cluster) and red (13 pixels cluster). These aren't hard to see at all.

In practice they tend to get lost in lighter areas but I have a shot of a silver coloured car and it's easy enough to see the blue spot on the door.

As I've said before, I came back to the Pentax fold after an absence of 20 years, right now I rather wish I hadn't bothered. I took your advice to buy the 360 flash too. It's certainly no encouragement to buy any more accessories.

I emailed the supplier about the problem, (Park Cameras) but no reply as yet. If necessary I'll go there yet again. I suppose Pentax uk won't have a clue if I ask them how they'd go about repairing something so new.

MattMatic

Link Posted 17/10/2003 - 14:41
Malcolm,

Just to allay your fears (a little at least), sensor problems run across the board, irrespective of make. The CCDs are cutting-edge technology and sometimes they do go wrong It's not Pentax, per se. You could just as easily have had the same experience with any of the high end DSLRs.

The image sensor is made by Sony (same as in the Nikon D100). Working in the electronics industry I know that chips sometimes just go AWOL on their own - even after they've been built, tested, and dispatched. Additionally, I know that first production runs can be "interesting" to say the least. It's one the risks of buying absolutely brand new kit.

That said, I would telephone Park Cameras if I were you and arrange a replacement. I bought mine from there, and all has been well so far (except for dust on the sensor which produced unsightly blobs - easily blown off in the "clean sensor" mode).

You might be right about Pentax UK, though there's no harm ringing them

Matt
http://www.mattmatic.co.uk
(For gallery, tips and links)

malcolmk

Link Posted 17/10/2003 - 22:06
It was as you may have suspected, digital photography that I was beginning to regret not Pentax. I think the camera's great and lots of fun though I suspect I'll have to regard it (and similar cameras) as short life items and that I won't be keeping it for 20 years like my previous purchases.

Anyway, things are beginning to look up. Park Cameras behaved impeccably when the first camera had the clock fault and they contacted me today to say they will swap the current one too as soon as they have more stock which should be next Tuesday. So another little drive down the M23 looms.

BTW, I knew of the Nikon/Sony links, I researched the *istD extensively for two weeks before purchase, and I've seen stuck pixels in pro grade Sony video gear too. It happens.

Thanks for your interest and support Matt.

Anonymous

Link Posted 17/10/2003 - 22:43
Hello Malcolm,

Just a little anecdotal evidence which, I think, will back up Matt's comments.

"Stuck pixels" have featured a lot recently in my wanderings. Among my other vices is that of "flying" MS Flight Simulator on the PC. Recently a number of the members of the PC Pilots' Club of Ireland ( http://www.pcpilotsireland.com ) have purchased flat screens and this topic has featured on our club emails. Last night I was visiting a friend who has spent his life in electronics - plasma screens, MRI scanners etc etc and I happened to mention your problem and I commented that it was the first I had heard of the problem being raised in relation to digital cameras. Before I could go any further his immediate response was "Well they are CCDs after all". It was as simple as that i.e. they are an inherent possibility of any CCD device. Not any consolation to you of course, or anybody else so effected. But I think it shows that it is not something the Pentax can be blamed for. As Matt inferred, it doesn't matter whose name is on the camera - if it's a CCD it's a possibility.

Interestingly, Matt raised another topic which has been exercising my mind of late. You may have gleaned from my previous posts that I am considering the *ist D myself. However floating around in my brain somewhere is the question of "first generation" Matt referred to first production runs sometimes been a risk, but while trying to come to a decision I have been wondering whether I should go for the *ist D i.e. first generation or wait for the second generation from Pentax. But I am also aware that even that course of action might not necessarrily guarantee freedom from problems. However, having waited so long so far, I suspect that the desire to get my hands on a digital SLR will probably win the arguement.

By the way Matt, I was with my "friendly dealer" today and decided to price a "go for broke" package including the "360" flash. When he rang the distributor for prices he was unable to speak to his contact. Half an hour later he rang me to say that they did not have a 360 in stock but that they would "do me a deal" for an AF500FGZ for marginally more than the cost of the 360. How much temptation can a guy be expected to resist

Anonymous

Link Posted 17/10/2003 - 22:49
Hi Malcolm,

Just a quickie to say that I hadn't seen your post timed at 10:06 before composing my "tome" But, yes, I think you may be correct in saying that SLRs of the digital age may not give us as many years of service as our old windie-uppy and auto-focus film SLRs.

Anonymous

Link Posted 18/10/2003 - 00:12
Sorry folks,

For some reason those last two posts - 10:43pm and 10:49pm - didn't carry my Username.

Lurach

Link Posted 18/10/2003 - 00:20
And again???

something funny going on, I am appearing as logged-in. But when I send a post my Username drops out??

Lurach
Lurach
Add a Comment
You must be registered or logged-in to comment.