to complement KJbruin's sample of the 1d here is a sample with the 24-105 on a 5d. Developped in DPP sharpend 3, no other post. Go find them if you want to
Thanks for that flash info Woody. Just tried on mine. Didn't work until the flash was on the camera and camera was turned on. Then if I turn the camera off, I can switch back and forth to gnc, but only until I turn flash off.
This setting made me take out the manual, one of my least favorite manuals. I had to dig out the manual (that really small booklet) for the 3502 module to find reference to this function. This is from the M4 version. And reading it did not help me understand it and how it would work on the M8 .
The section on "Flash mode with guide number control" starts with:
"The amount of flash light emitted in the flash mode with guide number control (GNC) is also based, among other things, on the distance setting or the camera's automatic distance measuring system."
This must be one of those "separating the men from the boys" type descriptions, and I know where I fall.
Digital sensors are colorless. What makes the colors is the Color Filter Array and mostly the RAW converters... ...Just change the color temp by 10 degree. Change the saturation, contrast, curve or the calibration parameters... ...There is absolutely no such thing like "Canon color", "Leica color". ;)
In the 1Ds/M8 comparison kindly posted by KJ, the M8's greens seem to have shifted towards yellow/red. That suggests that there is some compression in the green yellow slice of the colour wheel. I don't know if that was done in post processing, but I have seen it in many of the DMR and M8 files. Interestingly, it was also a feature of Kodak film which is why Fuji film was popular with garden photographers.
Could someone with more knowledge of capture and post processing (i.e. probably any of you) tell me if that green compression is happening because of some aspect of the sensor, or if it is happening in post processing? Either way, what would be the best way of ensuring that greens looked more like the 1Ds green, while yellows retain their warmth, i.e. how yellowy greens can be separated from greeny yellows in post processing, if you see what I mean?
I am not asking this to be difficult. I would like to know about a solution if there is one as it affects my readiness to use Leica digital cameras.
What is curious about this colour issue is that we can profile our monitors, scanners and printers to a certain standard but digital cameras with the exception of some MF backs are seemingly all over the plavce. It's possible to bend and squash different color responses into line in post production but what if often missed are the subtle transitions and variations in hue - if these are not captured you can't invent them later. I read an interview with a guy from one of the major MF digital back companies who claimed Canon have weak colour filters over their CMOS sensors in order to improve their low light /high ISO capability. If this were true I guess it would explain the sometimes crude colour representaton. One givaway is skin - the very subtlest rendering and discrimination of tone is needed to reproduce this well and from what I've seen from M8 and DMR files leica have it cracked. For me - if I were shooting weddings - this would be a major decider.
Thanks Guy, that's interesting. Would you say that it's possible that they undercooked the IR filtration on the DMR too then? The greens in many of the shots on the DMR bible thread stopped me investigating the DMR further. The Kodak look is very appealing and when I used to shoot film I usually chose it. However, in those days I had the option to shoot Agfa or Fuji when I needed to. With digital I would rather use a camera with the most neutral colour, and play afterwards if I want to compress skin colours et cetera. The profile situation is not ideal for me as I have to shoot JPEG.
BTW, your page 50 pictures were not ignored by me, I just didn't think there was much to be said. The M shows marginally richer colour, but still seems to suffer from the various problems that we have been discussing, and you did not post crops.
Guy,
Re: "Go get the same film emulsion of Kodachrome and shoot both you will see a difference. I have done it with Nikon, Canon and leica back in my film days"
Off course films and Lenses make the color difference. More precisely the color character was given by the "Leica Lens", not by the "Leica Body".
And with the digital, the Film chacteristics are replaced by the RAW converters.
Guy Mancuso wrote:
Part of that is true but they still have a signature to them. They lenses have not changed just because we went to digital. Your also dealing with different types of sensors CCD , CMOS made by Kodak, Sony,Canon etc. etc. . AA fliters/Ir filters all of that gets thrown into a puzzle along with Profiles, Converters . You just can't say the converter is the controlling element or the only one
I didn't say it is the only one. I said "it has far greater leverage".
Now, for this issue, what RAW converters are you guys using ? The "Leica" flavor will show up if it is a Dedicated Leica RAW conversion program. But if one is using off-the-shelf generic one, all the discussion about the colors are moot.
The profile is the one that Jamie Roberts has put together. I've been using it for the past few days and am pleased with the results. It is very easy to add to C1 and really does a nice job with colors. It's not a perfect solution, but is a very good work-around.
shirozina wrote:
... what if often missed are the subtle transitions and variations in hue - if these are not captured you can't invent them later. I read an interview with a guy from one of the major MF digital back companies who claimed Canon have weak colour filters over their CMOS sensors in order to improve their low light /high ISO capability. If this were true I guess it would explain the sometimes crude colour representaton.
I hack on CRW converters, and do see a broadband response by Canon sensors. To extract the green channel, for example, a significant fraction of R and B is subtracted from G. The filters over each sensor site are more pastel, so to speak. The tradeoff is better response and spacial resolution in the processed luminousity channel, but poorer graduation in the chroma channel. The overall gamut should be unaffected.
rico wrote:
I hack on CRW converters, and do see a broadband response by Canon sensors. To extract the green channel, for example, a significant fraction of R and B is subtracted from G. The filters over each sensor site are more pastel, so to speak. The tradeoff is better response and spacial resolution in the processed luminousity channel, but poorer graduation in the chroma channel. The overall gamut should be unaffected.
Rico this is very interesting, could you expand a little more please? Is Canon the only one doing this or is it common? Other tradeoffs?
I assume it helps high ISO quality because the filters block less light - more photons get to the sensor.
The tradeoffs a company makes in teh design of digital cameras and sensors seem to be some of the most important things we talk about in the forum. It would be nice if manufacturers were honest enough to say "on this model with this sensor we made these design decisions ... to get these benefits .... ". I am not talking about simple features or the number of frame per second. There are very fundamental decisions they make that greatly affect image quality but we never hear about them. We in effect have to reverse engineer an image to figure out the choices they make.
Guy Mancuso wrote:
Thanks yes that is Jamie's hard work. They are profiles to use without the IR cut filter and some have used them in this thread and they do work pretty good and do come close. But the basic issue is it does not stop the IR light from coming in, so there is some other colors that get effected to . really the true fix is block the IR light from actually hitting the sensor.
Now the long and short story of this is the design of the camera. The lenses being so close to the sensor the IR filter had to be thin or weak to get the lenses to work as leica wanted them too. If they blocked more IR light issues would have arose like astimatism and slight softer corners. no leica shooter wants soft corners reason we buy leica is the glass and it's performance and there performance wide open. So what they did not count on is the magenta cast so either they have some ( My Theory)options recall all the M8's and put in a stronger IR filter on the sensor and we may degrade our images or casuse more reflections inside the box or make use of the extrenal filters which don't cause reflections and block the IR like all other Dig camera's. The RD-1 has somewhat the same issues although not as bad but this is a design limitations for RF as it turns out, now the question is what will leica do . Again my theory is you will use the IR filters on your lenses, changing the sensor is just not a good option. Now first thought is damn they really messed this up bigtime and I thought the same thing at first.
Than i got to thinking about this and here is my latest thoughts the IR filter also it seems increase sharpness and we have to test this more to be sure. Now if we went in and changed the sensor and made the images lees sharp than scew that I am not doing it and I will use the filters which cause no ill effects EXCEPT for 28mm and wider which causes a cyan cast in the corners. Leica with coded lenses will come with a firmware update to tell the processing engine even in raw that a certain lens is being used and to apply that correction to it, so the cast will go away or maybe select your lens in the menu for uncode non Leica lenses like a Zeiss or the C/V lenses . Not sure what there going to do on that. But the bottom line the camera as it is today is sharp as can be so i will use the IR filters to keep that quality and the firmware and new profiles will sort this all back out. How Leica handles all this is yet to be announced. There is another issue and that is streaking from light sources and that is a hardware fix that leica needs to do and more info on that will be coming soon. So yes it has a few growing pains at the moment but when all wroked out should be a great RF camera . For me i will have bothe the DMR and M8 systems and use whatever makes sense for each job at hand. So that is what it comes down to in a nutshelll. Some folks will scream about this but i understand what is happening and it is what it is, so one has to decide this for themselves what to do. Personnally it is already one of those out of my dead cold hands cameras...Show more →
Thanks...that's what I needed to know. Much appreciated.
The operative term is "spectral response", and a sample sensor with RGB filtration is shown in Figure 3. Note the overlapping color sensitivity, such that nominal red and blue light will be partly detected by the green sensor site. I imagine all sensors are designed in this way, both to increase general sensitivity (ISO) and because broadband filtration is cheaper than narrowband. I like this quote in the text:
"The quality of the final image depends in large part on which [interpolation] algorithm is used. For this reason many camera manufacturers consider their best color interpolation algorithms to be trade secrets and do not publish them."
Interesting all these technicalities, I really enjoy them.
But did anybody notice how good the M8 is at rendering the character of older Leica lenses?
Summarit 50/1.5 screw-mount+ adapter built 1952. Wide open; I think I can claim the "Leica-Glow"
That's a beautiful lens. How much is a lens like that? Presumably it is quite affordable now.
I discovered that if you go into photoshop's HSB adjuster, select 'greens' and then eyedrop anywhere green in your image, photoshop re-labels the colour segment to 'yellows 2'. Here's the correction that I think is needed. I have exaggerated it a bit so it's clear. I would do it with a layer and then paint out the number plate which has gone tonally jaggy.
The original colour was true to nature, Richard.It is autumn after all and it was late afternoon golden sunlight. You've turned the bright yellow number plate on my car into lurid green (and my wife's car)...and lost the atmosphere I captured...
Those Summarit lenses - if you can find a good one, they go for about 300$. If one stops it down to 5.6 it is as sharp as any Summicron, open it up and it produces a dreamy haze and the mother of all bokeh's..
Brainiac (desperate now): Is it the hue/saturation window you don't like? I could change it. I could make it levels... or curves... or colour replace, anything really - there are so many to choose from. Should it be a little further to the left? Don't give up on me now Jaap.
(Jaap turns away to conceal his pity, we see a close up of his eyes looking down, wise, sad, and unable to save his friend. Right of screen, very blurred, Brainiac is seen cutting off ear. Cut to closeup of ear being severed with dirty palette knife. Brown blood mixes with green paint as it gushes down like rivulets of mud over the verdant ochre pasture of Brainiac's paint-crusted goatee...)
Brainiac: Dang that hurts! ...and my art is so misunderstood!
Jaap: You've lost the plot mate. (exit)
Thanks for the tip about the summarit. Looks like good value.
Edited by brainiac on Nov 21, 2006 at 03:24 PM GMT
I got flamed in another forum for saying that colour balance was subjective..It seems like a good idea to cut off your ear. Your heirs will be thankful to make zillions of dollars from your photographs...
jaapv, post some summarit shots of people and such at wide open :P I dont know how to define the glow seeing of your last pix but it is so nice colors, indeed
. 
Off course films and Lenses make the color difference. More precisely the color character was given by the "Leica Lens", not by the "Leica Body".
