BlackDots phenomenon must be caused by something like unsharpmask algorithm ... It reminds me so much Photoshop wrong parameters with unsharpmask that I could reproduce that so easy.
dvarnav wrote:
BlackDots phenomenon must be caused by something like unsharpmask algorithm ... It reminds me so much Photoshop wrong parameters with unsharpmask that I could reproduce that so easy.
It has been shown to be a readout issue at the sensor. Probably caused by pixel over-saturation/readout issues.
cdTzabcan wrote:
You can always switch to live view and use the lighting fast contrast detect AF on the periphery of the frame ;-)
Ummm I would hardly call it "lightning fast". My old Dimage 7 (7 years old) acquires focus using this technique as fast and sometimes faster. One of the options I liked with the Dimage 7 was the ability to switch from a square acquisition area to a cross hairs. Made focus a snap on macro subjects.
Regards,
Jon
Andrew Gough wrote:
It has been shown to be a readout issue at the sensor. Probably caused by pixel over-saturation/readout issues.
To be honest if there is an option for us to see what exactly sees a CMOS sensor then maybe we will stop thiking like " It has been shown to be a readout issue at the sensor" The reason is that in CMOS sensors we have a good base of data that software then will produce a usefull RAW output.
Lotusm50 wrote:
What has changed from the 1Ds MkIII iteration of this sensor that would cause this, since it doesn't seem to happen with the earlier version/camera?
good question .... but a quick answer is software and DigicIV
dvarnav wrote:
BlackDots phenomenon must be caused by something like unsharpmask algorithm ... It reminds me so much Photoshop wrong parameters with unsharpmask that I could reproduce that so easy.
no doesn't act like that at all plus it is in the RAW data itself
Lotusm50 wrote:
What has changed from the 1Ds MkIII iteration of this sensor that would cause this, since it doesn't seem to happen with the earlier version/camera?
I believe that there were significant on chip changes to lower the cost of the 5D2 sensor. Readout channels were reduced from 8 to 4 for example, this would change the number of amplifiers etc...
Thats about as much as I know. Perhaps someone with a better understanding could add additional info.
You don't get the IQ from an $8000 camera for $2500 without some corners being cut.... I'm sure the 'fix' will be a typical (for Canon) drawn-out, multi-stumble PITA for all concerned. Nice testing work Canon. Very thorough. Nothing like some needless (and avoidable) frustration.
Conner999 wrote:
You don't get the IQ from an $8000 camera for $2500 without some corners being cut.... I'm sure the 'fix' will be a typical (for Canon) drawn-out, multi-stumble PITA for all concerned. Nice testing work Canon. Very thorough. Nothing like some needless (and avoidable) frustration.
I and a number of other people have been unable to reproduce the black dots after considerable effort to do that very thing. So even though they've appeared in other circumstances, the fact that I can't make them appear leaves me unable to blame Canon for having missed them. The only reason so many instances showed up even now was because the roll-out occurred during the Christmas season in the West.
I've been involved with software testing for a large company, and I know it is impossible to duplicate in a test lab the thousands of test environments and test cases of a roll-out to the world.
Conner999 is there any technical analysis comparing the sensor technology of the one and the other camera. Whats the defferencies of the 2 sensors finally. 1DsMarkIII and 5D MarkII. Because its not so simple.
I remember when I had to develop all my slide film and worked in dark rooms for color printing with chemicals used for embalming.
A few problems here and there? I don't like them either but the huge jump from what was to now is very, very welcome.
And the iso, asa, that used to be maxed at 160 or 320 and now very useable results at 3200. Wow and auto focus? Not just one place but many. Oh I do love my pro Canon body, what a beauty but no more film.
I remember when I had to develop all my slide film and worked in dark rooms for color printing with chemicals used for embalming.
A few problems here and there? I don't like them either but the huge jump from what was to now is very, very welcome.
Anyone ever use the Kodak stainless Steel "High-Speed" <ahem> Darkroom Color Paper Processor?
It was essentially a textured stainless steel drum set horizontally that rotated at high speed over a shallow chemical tray. The drum was filled with temperature-controlled water, and it rotated through the chemical in the tray, pulling up a layer of chemical onto its textured outer surface.
In white light, you set up your beakers of color chemicals at one side of the tray in strict order--because you'd have to reach for the first four (developer, stop, bleach, and fix) in pretty-darned-close-to-total darkness (how close to total darkness was that brown safelight? So close to total darkness that you had to look directly at it and let your eyes adust to tell it was turned on). You had a tray of water next to the processor to do a pre-soak of the paper. In that tray you spread out the web apron that you'd use to flop the paper on that spinning drum.
You turned off the white light and ambled over to the enlarger to expose the color paper, then returned to the sink and laid the paper face-up on the apron in the water tray (all in darkness). Then you'd pour the beaker of developer into the tray, turn on the drum, and let its spin up a layer of developer onto its surface.
Then you'd pick up the print and the apron by two corners and--in one smooth, deft, svelte, precise motion--you'd lay the paper and apron across the spinning drum (in the dark) the way a toreador sweeps his cape before a raging bull. Without a break in motion, you'd reach up to switch on the ubiquitous GraLab 300 timer.
If you were really good at this (and I was good...yeah, bay-bee! Aarr, aarr, aarr!), the paper would be perfectly aligned under the web apron on the spinning drum, hydroplaning serenely over the film of developer.
If you were 'way off, you'd hear the paper fly across the room in the dark and slap wetly in the darkness against the far wall.
If you were a little bit off, the paper would be crooked on the drum (so that the corners wouldn't be developed), but you wouldn't know that for about 20 minutes until you turned on the white light.
At the end of the development, you'd tip the chemical tray and pour in the stop, then the bleach, finally the fix. After half the fixing time, you could turn on the white light. Another 5 minutes and you could remove the print from the drum and see...
...that your color balance was slightly off (beyond the ring-around).
Rinse, repeat.
I'm constantly amazed by how easy it is to create color prints from digital cameras.
Ain't that the truth, so easy. We bubbled C02 through a concoction of chemicals, with formaldehyde as a preservative gassing out into the heavy damp air with the dark room light that made you feel more blind than closing your eyes.
Digital and printer printing. . . bring it on. . :^)
Thanks for that write up, RDKirk.
I always sent my Kodachromes to the developer to have the slides back in 3 to 6 weeks. One of the reasons I switched to Fuji later on was that the turnaround time only was a couple of days.
. Can't on.