Re: *ist D sensor noise survey
Hi Rob, On Tue, 10 Feb 2004 13:36:36 +1000, Rob Studdert wrote: I'm interested in making an informal survey of the noise performance of our *ist D cameras. Anyone with access to a PC who has permission to run the little test app at http://www.starzen.com/imaging/deadpixeltest.htm can participate. Good idea! I will run that test later and mail you my results. However, I happened to run a less scientific test yesterday, just to see if there are big problems with my two week old *ISTD :-) I ran a series of long exposures at 200 and 800 ISO settings, and counted the hot pixels at 100% magnification on the screen. All exposures with cap on of course, and NR switched off. Exposure ISO-200ISO-800 1/8 0 2 1/2 1 3 2 3 10 8 8 50 30 50 200 Actually, in that last shot there are probably thousands of 'warm' pixels too, the screen really looked like a coloured version of a clear sky at night :-) I did not do a full test with NR on, but a 30 second exposure at 200 ISO did not show any hot pixels. From what I have read on the list in the past this is a fairly typical outcome, correct ? Regards, JvW -- Jan van Wijk; http://www.dfsee.com/gallery
Re: *ist D sensor noise survey
On 10 Feb 2004 at 16:50, Jostein wrote: Quoting Rob Studdert [EMAIL PROTECTED]: 0=Hot,2798,135,69 Hi, Rob, What was your threshold for hotness? 65? And the other threshold...? Hi Jostein, Default, 60/250 Cheers, Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
RE: *ist D sensor noise survey
Ok, what about actual dead pixels? Mine seems to have a couple. Does anybody else see them, or should I be considering sending it back? Jeff. -Original Message- From: Rob Studdert [mailto:[EMAIL PROTECTED] Sent: Monday, February 09, 2004 8:37 PM To: [EMAIL PROTECTED] Subject: *ist D sensor noise survey Hey it's a while since we had a survey... I'm interested in making an informal survey of the noise performance of our *ist D cameras. Anyone with access to a PC who has permission to run the little test app at http://www.starzen.com/imaging/deadpixeltest.htm can participate. One exposure is all that's required for the test however in order to achieve consistency we need to make sure that each camera is set up the same. I propose that the test shot should be made as follows: 10 seconds manual exposure (lens capped) 200ISO Daylight WB NR off Saturation setting (middle) Sharpness setting (left most) Contrast setting (left most) sRGB CS TIFF L file The tiff file can then be opened and tested under the default settings of the DeadPixelTest application and the information file saved. I ran the procedure above and the results were as follows: [DeadPixelText] Version=1.0 Description= FileType=TIFF NumBadPixels=15 0=Hot,2798,135,69 1=Hot,1954,339,113 2=Hot,1809,585,64 3=Hot,726,610,112 4=Hot,726,611,192 5=Hot,726,612,112 6=Hot,2312,753,121 7=Hot,323,766,94 8=Hot,572,1365,116 9=Hot,1627,1400,64 10=Hot,2163,1958,96 11=Hot,2162,1959,113 12=Hot,2163,1959,145 13=Hot,2164,1959,112 14=Hot,2163,1960,98 The first two numbers is the pixel location and the last number is the heat, 0 being off and 255 being full on. So I have one pixel that's 3/4 on at 10 seconds. If anyone would like to mail me their results I'll collate and publish the data later down the track (I'll keep data sources anonymous if requested). Cheers, Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
RE: *ist D sensor noise survey
On 10 Feb 2004 at 15:22, Jeff Jonsson wrote: Ok, what about actual dead pixels? Mine seems to have a couple. Does anybody else see them, or should I be considering sending it back? Doesn't sound so good, none of the six test results I've seen so far show dead pixels. Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
Re: *ist D sensor noise survey
I show no dead pixels, but quite a few so called hot ones, but I can't see them in PS7 Bill - Original Message - From: Rob Studdert [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, February 10, 2004 6:32 PM Subject: RE: *ist D sensor noise survey On 10 Feb 2004 at 15:22, Jeff Jonsson wrote: Ok, what about actual dead pixels? Mine seems to have a couple. Does anybody else see them, or should I be considering sending it back? Doesn't sound so good, none of the six test results I've seen so far show dead pixels. Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
Re: *ist D sensor noise survey
Hi, Rob. I don't have the camera with me tonight, but I have some results from yesterday with a 30s exposure at ISO 200. [DeadPixelText] Version=1.0 Description= FileType=TIFF NumBadPixels=4 0=Hot,2820,150,77 1=Hot,1865,209,67 Ambient temperature was 20-22°C, and the camera had been turned off for a long time before the test. I'll get back to you with standardised data later. Cheers, Jostein - Original Message - From: Rob Studdert [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, February 10, 2004 4:36 AM Subject: *ist D sensor noise survey Hey it's a while since we had a survey... I'm interested in making an informal survey of the noise performance of our *ist D cameras. Anyone with access to a PC who has permission to run the little test app at http://www.starzen.com/imaging/deadpixeltest.htm can participate. One exposure is all that's required for the test however in order to achieve consistency we need to make sure that each camera is set up the same. I propose that the test shot should be made as follows: 10 seconds manual exposure (lens capped) 200ISO Daylight WB NR off Saturation setting (middle) Sharpness setting (left most) Contrast setting (left most) sRGB CS TIFF L file The tiff file can then be opened and tested under the default settings of the DeadPixelTest application and the information file saved. I ran the procedure above and the results were as follows: [DeadPixelText] Version=1.0 Description= FileType=TIFF NumBadPixels=15 0=Hot,2798,135,69 1=Hot,1954,339,113 2=Hot,1809,585,64 3=Hot,726,610,112 4=Hot,726,611,192 5=Hot,726,612,112 6=Hot,2312,753,121 7=Hot,323,766,94 8=Hot,572,1365,116 9=Hot,1627,1400,64 10=Hot,2163,1958,96 11=Hot,2162,1959,113 12=Hot,2163,1959,145 13=Hot,2164,1959,112 14=Hot,2163,1960,98 The first two numbers is the pixel location and the last number is the heat, 0 being off and 255 being full on. So I have one pixel that's 3/4 on at 10 seconds. If anyone would like to mail me their results I'll collate and publish the data later down the track (I'll keep data sources anonymous if requested). Cheers, Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
RE: *ist D sensor noise survey
I haven't done the test, I can actually see the two pixels in my images. They're stuck on, and appear as bright dots. The same two pixels in every image, at least those with dark enough subject matter that they can be seen. I have dead pixels in my Canon Powershot G1, but the canon manual actually says that a certain few pixels will always die on the chips they used. They show up larger in those 3MP images, than the two pixels in the 6MP images from the *istD. The easiest way to spot dead pixels is to take an exposure with your lens cap ON. I encourage list members to try this, and let's see how many of you have them. If they are to be expected, then we can all just accept them, if not, then those of us that have them can send our cameras back to pentax for warranty repair. Thanks, Jeff. -Original Message- From: Rob Studdert [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 10, 2004 4:33 PM To: [EMAIL PROTECTED] Subject: RE: *ist D sensor noise survey On 10 Feb 2004 at 15:22, Jeff Jonsson wrote: Ok, what about actual dead pixels? Mine seems to have a couple. Does anybody else see them, or should I be considering sending it back? Doesn't sound so good, none of the six test results I've seen so far show dead pixels. Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
RE: *ist D sensor noise survey
May I recommend one additional stipulation. I ran the test for 13 exposures. I found an increasing trend among the number of hot pixels as the number of exposures increased. Perhaps if everyone would run the test with the camera having been at rest for an hour the results might be more consistent. Over the course of my 13 exposures the ratio of hot pixels from the 13th to the 1rst was 1.56. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Rob Studdert Sent: Monday, February 09, 2004 9:37 PM To: [EMAIL PROTECTED] Subject: *ist D sensor noise survey Hey it's a while since we had a survey... I'm interested in making an informal survey of the noise performance of our *ist D cameras. Anyone with access to a PC who has permission to run the little test app at http://www.starzen.com/imaging/deadpixeltest.htm can participate. One exposure is all that's required for the test however in order to achieve consistency we need to make sure that each camera is set up the same. I propose that the test shot should be made as follows: 10 seconds manual exposure (lens capped) 200ISO Daylight WB NR off Saturation setting (middle) Sharpness setting (left most) Contrast setting (left most) sRGB CS TIFF L file The tiff file can then be opened and tested under the default settings of the DeadPixelTest application and the information file saved. I ran the procedure above and the results were as follows: [DeadPixelText] Version=1.0 Description= FileType=TIFF NumBadPixels=15 0=Hot,2798,135,69 1=Hot,1954,339,113 2=Hot,1809,585,64 3=Hot,726,610,112 4=Hot,726,611,192 5=Hot,726,612,112 6=Hot,2312,753,121 7=Hot,323,766,94 8=Hot,572,1365,116 9=Hot,1627,1400,64 10=Hot,2163,1958,96 11=Hot,2162,1959,113 12=Hot,2163,1959,145 13=Hot,2164,1959,112 14=Hot,2163,1960,98 The first two numbers is the pixel location and the last number is the heat, 0 being off and 255 being full on. So I have one pixel that's 3/4 on at 10 seconds. If anyone would like to mail me their results I'll collate and publish the data later down the track (I'll keep data sources anonymous if requested). Cheers, Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
RE: *ist D sensor noise survey
On 9 Feb 2004 at 22:02, Kevin Thornsberry wrote: May I recommend one additional stipulation. I ran the test for 13 exposures. I found an increasing trend among the number of hot pixels as the number of exposures increased. Perhaps if everyone would run the test with the camera having been at rest for an hour the results might be more consistent. Over the course of my 13 exposures the ratio of hot pixels from the 13th to the 1rst was 1.56. Very interesting, I didn't consider temperature as a variable. My test was done very soon after turning on the camera (after setting the menu pre-sets) and was the one an only exposure, the ambient temp (test temp) here is 30dec C. What I failed to mention too was that the brightest pixel wasn't near as irritating in the full 1:1 image as the cluster of medium brightness pixels. Clicking on the left most column of the results table in the DeadPixel app will reveal a magnified view of the problem pixel and its surrounds. Cheers, Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
