Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: Lin Evans on May 22, 2004, 01:39:32 pm
-
[font color=\'#000000\']The "disadvantages," if one could really call it that, are that the 1D Mark II has eight megapixels of photosites and the 1DS has 11. The other "disadvantage," but only to the landscape or architectural shooter is that the 1D Mark II has a 1.3x reduced field of view.
On the other hand, the Mark II shoots 8 plus frame per second, has a huge buffer, has greater dynamic range, and looking at the 1.3x reduced field of view from the wildlife or sports photographer's perspective has the 30 percent telephoto "boost" effect.
Both will autofocus reliably at F8.....
Lin[/font]
-
[font color=\'#000000\']Having spent a bit of time in India (while carrying a lot of gear) I can assure you that theft is not a big problem. And the thieves everywhere steal what is made available to them, they don't check the model number....[/font]
-
[font color=\'#000000\']This means that at least in theory a manufacturer could design a lens with a variable aperture which could shift during the exposure and tap the "potential" for much greater dynamic range latent in the sensor.[/font]
[font color=\'#000000\']Lin,
Can you elaborate on this! Seems to me an aperture that varies during the exposure will simply give you one average exposure and solve nothing with regard to dynamic range. You need at least two separate exposures (I guess this is what you meant).
Two exposures with different f stops will need software similar to Helicon Focus to combine the different DoFs and I see that as problematical since parts of the two images might consist of a well exposed out-of-focus patch which has to be merged with a severely underexposed in-focus patch also deficient in detail for different reasons.
Two exposures with different shutter speeds is the usual approach (or 3 with autobracketing), but requires the subject to be static.
Those who own good RAW conversion software, such as Capture One or Adobe Camera Raw (V.2 or later) can get (roughly) a couple of stops of extra dynamic range out of a single exposure using the following procedure.
(1) Overexpose the image by one stop.
(2) Use exposure compensation during conversion to recover highlight detail.
(3) Do a second conversion but this time move EC the opposite direction, say +3 stops, to create an image with maximum detail in the shadows but completely blown out highlights. (It's probably advisable to move the NR and Luminance Smoothing controls to their maximum also.)
(4) Combine the two images with a 'digital blending' technique. Michael has a tutorial on this.[/font]
-
[font color=\'#000000\']For example, it "could" be that the processor is dumped between exposures to a buffer then later combined but I'm suspicious that it may be done differently.[/font]
[font color=\'#000000\']Lin,
However it's done, it's clear that each photodetector can only hold so many electrons. Once the 'wells' are full, they're full. No more exposures can be made without clearing the charge.
The multiple exposures possible with your S7000Z must be transferred to some sort of buffer, I would imagine. But the question does arise, how fast can the charge be cleared and the sensor reset for another exposure? If it's virtually instantaneous, then there is a clear advantage over autobracketing. For example, if you need 1/30th sec exposure for the shadows in a contrasty scene, which represents say a 3 stop overexposure with regard to the highlights, then the other exposure need be only 1/250th (call it 1/240th for simplification). The combined exposure of a 30th plus a 240th is 9/240ths or approx 1/26th sec. With Canon's Image Stabilisation one could probably get some reasonably sharp hand-held shot at this shutter speed, and if not then just increase the ISO.
Also, combining RAW images in camera should potentially produce better results than digitally blending converted TIFs in PS. I'm all in favour of such a development. I live in a place with lots of sunshine and dark forests. I'm constantly battling with dynamic range issues.[/font]
-
[font color=\'#000000\']
But bracketing with multiple exposures and blending the exposures must, by default, mean that the sensor has the latitude of "holding" the full dynamic range you are able to achieve with the blend of two or more exposures.
Huh???? The reason you bracket is precisely because the sensor DOESN'T have the capacity to hold the full dynamic range. You overexpose and let the bright parts clip on one shot and you underexpose and let the shadows be black on the other shot and you blend to give you the best combination.
As for doing this all automatically in the camera; maybe as a gimmick for consumer cameras this would be better than nothing to increase dynamic range, but all this processing to create and blend two exposures would have to take some time and during that time camera motion would be even more critical than for normal single exposure shooting.
As for a previous comment that this automatic in camera double exposure and blending being potentially superior to bracketing and saving both raw images for later blending; NO WAY! Even without the obvious problem that this procedure would have to take some time, and thus be even more sensitive to camera motion, there's also the problem that you have no good visual feedback to see if it really worked right; just a littly tiny LCD camera display. Blending two images in Photoshop gives you lots of options for exactly how to do the blend and thus getting the best possible blend and you can see exactly what you have for each trial.
In any case 1d and 1ds are both pro cameras and fortunately not burdened with a kluge like automatic in camera double exposure and blending.
Fuji Finepix technology increases dynamic range by having a low sensitivity sensor and a high sensitivity sensor for each pixel location, but for some reason this has not yet been implemented for any pro camera. Probably doesn't actually work as well as the hype would lead you to believe.[/font]
-
[font color=\'#000000\']Hi Ray,
It "appears" to simply slow to the rate you have set manually and keep chugging. I'm not sure if there are any statistics about this available - nothing in the manual.
I suspect it does much the same if you were shooting in aperture priority mode or auto.
Best regards,[/font]
-
[font color=\'#000000\']This might be taking the topic even further afield.
If we invented a sensor with an A/D converter for each pixel, and monitored in real time the rate at which each pixel gathered light during the exposure, then it seems like we could select when to grab the intensity of each pixel during the exposure. We could grab the bright values sooner and the dark values later. Effectively, we would be exposing for the rate at which light is gathered by the sensor rather than by fixing the time and reading the absolute exposure after that fixed time. The response curve could be dynamically determined during the exposure while avoiding either overexposed highlights or underexposed shadows, as long as the exposure was long enough to gather light for the shadows. With this, you wouldn't set exposure on the camera, only aperture.
Of course, we could take this one step further. At each pixel, we could record the entire light gathering curve over time during the exposure and post process this into the image. Perhaps this is what cameras will do 10 years from now... You could think of this as watching a photo print in the developer in the darkroom emerge from blank paper and controlling during when to stop the developer by painting stop on parts of the print when you were satisfied with the luminance.
Hmmm... I wonder if I should patent this idea ... :-)[/font]
-
[font color=\'#000000\']Actually the challenge of increasing dynamic range by measuring the relative times it takes for a large number of megapixels to fill up would require a time measuring device far greater than a billionth of a second. Each extra stop of exposure is going to blow an increasing percentage of the photosites. If you want a hand-held shot with a telephoto, even with IS, you probably need 1/250. If a certain number of photodetectors are going to 'fill up' within say half the exposure time, ie 1/500th, and you can time those events, then you gain 1 stop in dynamic range. If a smaller quantity fills up whithin 1/1000th sec, you gain another stop, and if a yet smaller quantity fills up in 1/2000th sec, you gain 3 extra stops of DR, and so on.
For such a system to be worthwhile, you might need to time the relative fill-up rate of, say 3 megapixels within 1/4000th of a second, not to mention the time it takes to apply an algorithm to the different rates.
It might be possible :D .[/font]
-
[font color=\'#000000\']I was looking at getting the 10D but now am considering the 1D Mark II. Are there any disadvantages to the Mark II over the 1Ds? I know that you can't go above a f/5.6 on the 10D, how about the Mark II? Also any thing that should make me want to stick with the 10D over the Mark II? Except for the price [/font]
-
[font color=\'#000000\']
I know that you can't go above a f/5.6 on the 10D, how about the Mark II?
Are you referring to aperture for auto-focus? I don't know what the AF minimum is for the 1DII.[/font]
-
[font color=\'#000000\']"greater dynamic range"
How does one know? Seems like there should be some easy way to measure dynamic range. (Maybe there is and I'm not aware.)
When I look at a 'test strip' on my CRT vs. LCD monitor I see more 'steps' on the CRT - the LCD doesn't have the range to let me discriminate between the steps at the ends. Can't something such as this be used with digitals?[/font]
-
[font color=\'#000000\']
"greater dynamic range"
From the great amount of reading I've done about this, the bottom line is that all these sensors (with maybe Fuji FinePix excluded) have the same dynamic range. It's locked into the basic technology. If your camera can shoot and save raw, then with proper conversion and processing you'll have the maximum dynamic range possible for digital camera sensors (better than slide film; not as good as negative film).
As for 1ds vs whatever. For today's ultimate DSLR quality no one contests that 1ds has no rivals and the advantages have already been mentioned. I could mention another consideration, however. There's places I simply would not take such an expensive camera, like for example traveling extensively in third world countries using public transportation and staying in cheap hotels. I could recover from having a $2000 outfit stolen, but $6000 for a camera body alone would be a big hit. In a place like India $6000 represents several years of income for many people, and therefore a huge temptation.[/font]
-
[font color=\'#000000\']Well, I've been going to India very regularly since 1972 and for some time I've been spending 6 months out of the year in Bombay. I love the culture and basic human and family values and the people of India more deeply than I can say and the incredible experiences of generosity, hospitality and honesty and simple unassuming affection have been far more impressive than any experiences of scams and deception. Home is where the heart is and my heart is in India no less than here. However, with so many people and so many extremely poor people, even a very small minority of dishonest people still means a lot of dishonest people. Reading an Indian newspaper is as appalling and depressing as any US newspaper. I wouldn't leave something valuable exposed at the Delhi main railway station for a second and I sleep on top of any valuables on a second class train. Thieves may not know specific models of cameras, but they'd be able to see that something like an eos 1ds is not a point and shoot. I've witnessed flat out jewelry grabbing theft on trains and I know people that have been ripped off in various ways in India. In some third world countries I'd have some concern for my personal safety if I were carrying anything obviously valuable.
You only need to lose a $6000+ camera once to ruin your whole day. If I get into serious photography in India (so far my long stays have been all business or visiting a spiritual retreat) I'll take my brother's eos 10d and Sigma 24-70, not my own 1ds and a bunch of expensive lenses.
In case you want to read something totally off topic, you can check out the India page on my website, about my experiences with my Indian business partner and our employees. Nothing but positive, beautiful, incredible experiences and impressions.
http://www.didgeridoings.com/India/India.html (http://www.didgeridoings.com/India/India.html)[/font]
-
[font color=\'#000000\']
From the great amount of reading I've done about this, the bottom line is that all these sensors (with maybe Fuji FinePix excluded) have the same dynamic range. It's locked into the basic technology
This is only true of consumer fixed lens sensors. The sensors used in pro level equipment vary from a bit over 8 stops to 12 stops. Canon's Director of Technical Services - Camera Division, U.S.A., Chuck Westfall says that Canon has measured the 1DS at greater than 8 and less than 9 stops of dynamic range. The 1D Mark II was measured at > 9 stops. Digital scanning backs such as BetterLight, etc., have up to 12 stops per manufacturers.
Now - the issue is measurement and how it's done and whether one uses the capacty of the sensor or the capacity of the sensor with a single exposure (not the same thing) in measurement. Who knows.....[/font]
-
[font color=\'#000000\']Hi Gary,
Yes, it's difficult to know exactly how the measurements were made and even more difficult to set a standard for these things which in some way equates to a film base which some have a better feel for.
One thing is certain, sensors have the ability to hold more dynamic range than can be captured with a single exposure aperture. With my Kodak DCS-760 and using Kodak software, it's very easy to go two stops beyond the apparent DR in the capture. Even more important is the fact that one may expose for highlights, shoot, expose for shadows and shoot an identical frame, then combine the two frames in software producing a result which has a couple stops more than can be obtained by a single RAW exposure and combining two conversions using the extremes of exposure compensation.
This means that at least in theory a manufacturer could design a lens with a variable aperture which could shift during the exposure and tap the "potential" for much greater dynamic range latent in the sensor.
Best regards,[/font]
-
[font color=\'#000000\']Hi Ray,
Two separate exposures later combined or two frames as a "double" exposure without dumping the first before the second occurs with the "combine" operation occuring in firmware would essentially be the same. Actually, it's already been done by Fuji (allowing double exposure) except it requires two presses of the shutter. What I'm theorizing is that it would be possible to combine electronically the the steps of exposure capture, different exposure capture with a single press of the shutter, especially since the majority of digital instruments with the exception of CMOS sensors have an electronic shutter anyway. Even the Canon EOS-1D only uses the mechanical shutter essentially to protect the sensor and gets its 1/500th flash sync and 1/16,000th shutter speed via the electronic shutter.
As it is presently, it takes two exposures to really take advantage of the entire dynamic range potential of the sensor. The down side is that we must combine the two in software selecting the appropriate portions of each for the final extended dynamic range composite. By combining the steps in an electronic sensor it "seems" that it would be fairly easy to accomplish much the same thing thing with a single press of the shutter and let the electronic switching and firmware take care of the remainder.
Hewlett-Packard already has a firmware feature on their consumer digicam which essentially duplicates the PhotoShop CS "shadow/highlight" function, and this approach could be used with the double exposure feature such as Fuji offers in some models to render a greatly enhanced dynamic range photo. Just meter on the highs, meter on the lows (sort of like the 1D Mark II offers) and press the shutter once rendering a double exposure processed entirely within firmware.
I don't know enough about the internals of present sensor technology to be certain this is possible, but it certainly seems logical to me that it could be done.
Best regards,
Lin[/font]
-
[font color=\'#000000\']Hi Ray,
I'm not certain of exactly how the process works. For example, it "could" be that the processor is dumped between exposures to a buffer then later combined but I'm suspicious that it may be done differently.
Let's take an example. The image below is from one of my Fuji digicams (S7000Z) which has the "multiple exposure" feature. You shoot, then you use the toggle switch to make a decision about saving the image to the card or continuing. You press the right arrow on the toggle, expose and shoot a second time or third or Fuji says "no limit" then when you are finished with exposures you press the center "O.K." portion of the multi-switch and only then is the file written to the card.
I'm assuming that if it were being written to a buffer there would be a very small number of exposures possible before filling - especially at the 12 megapixel interpolated file size on the S7000Z. But I've made dozens of exposures just for the sake of experimentation and there seems to be no limit. I just shot three quick exposures each at a different focal length for the image below. It would seem that this type technology combined with what HP does with "shadow/light" could possibly be used to greatly expand dynamic range to the limits of the sensor's ability to hold detail at the tails of the curve???
Lin
(http://www.lin-evans.com/samples/triple.jpg)[/font]
-
[font color=\'#000000\']I don't care what amazing reports of vast latitude there are for any sort of digital sensor. I share the skepticism about these claims and how they were arrived at. I figure to always go out with a tripod and to do bracketed shots if I have the slightest doubt. Blending in Photoshop is quick and easy and bomb-proof with the $20 Miranda plug. Compact flash at $200 for 4 gb means you can "waste" lots and lots and lots of shots in the field, even if you only save raw and you don't bother to delete anything until you download.
Better way too many pictures than any great shot that doesn't have all the latitude you thought it ought to have had. Canon nerds won't sponsor your next trip to re-shoot!![/font]
-
[font color=\'#000000\']It's a natural progression from questions asked about advantages of the 1D Mark II over the 1DS. One of the advantages is increased dynamic range and questions about that have lead in a circuitous route here.[/font]
-
[font color=\'#000000\']As for a previous comment that this automatic in camera double exposure and blending being potentially superior to bracketing and saving both raw images for later blending; NO WAY! Even without the obvious problem that this procedure would have to take some time, and thus be even more sensitive to camera motion, there's also the problem that you have no good visual feedback to see if it really worked right; just a littly tiny LCD camera display. Blending two images in Photoshop gives you lots of options for exactly how to do the blend and thus getting the best possible blend and you can see exactly what you have for each trial.[/font]
[font color=\'#000000\']You might have noticed, Didger, I also wrote that such blending software in-camera would have to be very good. Everything takes time, but as processor speeds increase and buffer sizes increase, saving two RAW images and/or shunting them off to a separate buffer for blending should eventually be possible in the blink of an eye.
You don't need visual feedback if the software or firmware is working properly, considering the exceptionally high dynamic range of the combined images, and I see no reason why the camera should be more sensitive to motion. The concept is that it should be less sensitive to motion because the time interval between exposures is reduced. Such a system should allow hand-held dual shots more often. The problem with my D60 is that autobracketing takes over a full second. I don't think the 1Ds is much faster. Fortunately the two shots that are the most useful (the underexposed and the overexposed) are consecutive, so the total exposure time from the beginning of one shot to the end of the next is probably often less than a second, but not by much. Even with a tripod, that's not ideal. A lot of movement can take place within a second :) .[/font]
-
[font color=\'#000000\']
I think I understand that the 1D and 1DS are "pro" cameras (I have the 1D, 1DS, 1D Mark II, 10D, D30, DCS-760, Kodak Pro back and a couple dozen fixed lens digicams),
Lin,
Keeping the thread on track, I'd be interested to know just how great is the time interval between one shot and the next on the Mark ll in continuous mode. Clearly, you can't have 8 frames per second if each exposure is 1/8th sec or greater. There has to be some shutter speed threshold below which the camera gives less than 8 frames per second, perhaps 1/15th or 1/20th. Is this a known (or advertised) specification?
As regards autobracketing for the purpose of increasing dynamic range, the 1D Mark ll would appear to have an advantage over any other camera. Those 3 bracketed shots would likely occur within less than 1/2 sec. in reasonable lighting, and the 2 consecutive shots (underexposed and overexposed) in possibly less than1/3rd sec. This means that much of the time a tripod would not be required for this technique of blending images.[/font]
-
[font color=\'#000000\']It "appears" to simply slow to the rate you have set manually and keep chugging.[/font]
[font color=\'#000000\']Lin,
You mean in continuous mode there's a variable speed setting up to a maximum of 8 frames/sec?[/font]
-
[font color=\'#000000\']Isn't the double frame trick essentially what Fuji does with one of their new sensors? I think it's the SR that has two sets of pixels, one of which is much less sensitive than the other (tucked in the spaces between the bigger pixels). They actually work during the same exposure if I understand it correctly. Of course, you lose some resolution by doing this (you don't cut it in half, because the low sensitivity pixels are much smaller, but you DO lose 30% or so). Highlights record on the small, low sensitivity pixels, while shadows are on the larger, higher sensitivity ones. I think this is what the S3 is supposed to do. I don't know about the noise implications, because the highlights are effectively being captured with digicam sized pixels, but the small pixels are using a VERY low iso (operating two-three stops down from main exposure), and capturing highlights, not noisier shadows.
-Dan[/font]
-
[font color=\'#000000\']That's essentially the way I've understood it from the rather non-technical explanations I've read.
Fuji has pioneered several interesting features on their consumer cameras and "field tested" the public acceptance before making them available in their dSLR's.
I'm still trying to understand the process they use for multiple exposures. Logic would dictate that they save each frame to a buffer and then combine them in firmware before writing to the media, but I sincerely don't believe that's how its done. The Fuji manual for the S7000, for example, says there is "no limit" to the number of multiple exposures one can make. This seems more like they are simply letting the electron charges spill over after an exposure sort of like filling buckets of water from a waterfall and only flushing the cells after the user selects the option to write the file to the media. The trick would be how they avoid the problems of adjacent sensor current bleed (blooming) but I really haven't a clue how it's done. I know that with the very high costs of buffer RAM, I can't see how they could afford to have enough to provide "unlimited" exposures when even a 1 gigabyte card can't provide that.[/font]
-
[font color=\'#000000\']
for such a system to be worthwhile, you might need to time the relative fill-up rate of, say 3 megapixels within 1/4000th of a second, not to mention the time it takes to apply an algorithm to the different rates.
Look, I said we should leave the little trivia to the engineers, and us idea geniuses can focus on the occasional brilliant basic concept and cashing our royalty checks.
Same thing applies to dealing with LCD transparency levels and the presently way too slow LCD response times and present total impossibility of making an LCD grid anywhere near that dense. Just let the engineers take care of those details. That's what engineers are for. They don't have time for great breakthough ideas; I don't have time for trivia. It all works out.[/font]
-
[font color=\'#000000\']Interesting idea. I must admit I sometimes like to wander off into science fiction speculation as to what might be possible as processing power increases, and that's what it all boils down to - how fast can the calculations be done.
I'm not qualified to even guess if it would be practical to construct an image based upon the time it takes for individual photosites to 'fill up', but in principle it seems feasible, given enough power. The question is, could a device capable of measuring time in billionths of a second be incorporated into a small camera?[/font]
-
[font color=\'#000000\']Put an LCD element in front of each CCD sensor element. Each LCD element can be controlled to be totally transparent or quite dark. [/font]
[font color=\'#000000\']My objection to this idea would be - nothing's totally transparent, not even a Zeiss 21mm lens :D .[/font]
-
[font color=\'#000000\']
Are there any disadvantages to the Mark II over the 1Ds? I know that you can't go above a f/5.6 on the 10D, how about the Mark II?
First, in my perhaps-uninformed opinion, the 2 disadvantages of the 1D2 compared with the 1Ds is the former has a smaller sensor and is lower resolution. Of course if you shoot with long lenses a lot, the former is an advantage. But a 1D2 would make my 28mm lens have the same angle of view as a 36mm lens on a full-frame camera (almost no longer a wide angle) while a 10D would make it same as a 45mm lens on a full-frame camera, about the same as a 'standard-length' lens.
To offset the lower resolution, the 1D2 apparently has lower noise than the 1Ds.
What does 'can't go above a f/5.6 on the 10D' mean? (And do you mean smaller or larger aperture?) All the Canon digitals have no restrictions on use of apertures.[/font]
-
[font color=\'#000000\']
Canon has measured the 1DS at greater than 8 and less than 9 stops of dynamic range
I'm surprised to read that.
If I point a 1Ds and a long lens at a grey card or a small patch of evenly lit blue sky it shows up as a narrow band on the histogram. If I then dial in both plus and minus compensation, and re-shoot until that narrow band hits both ends of the histogram scale, it results in a latitude range of less than seven stops. That seems to me a practical test of useable latitude, and it fits with my real world 1Ds experience.
Interestingly the width of that histogram band isn't uniform across the range, which maybe suggests the Canon firmware is applying an exposure curve. But even so I struggle to believe there's eight to nine stops of latitude. I'm sure Canon wouldn't deliberately mislead, however perhaps the additional latitude that they measure is buried so deep in the noisy shadows that it's of little practical use?[/font]
-
[font color=\'#000000\']Two separate exposures later combined or two frames as a "double" exposure without dumping the first before the second occurs with the "combine" operation occuring in firmware would essentially be the same.[/font]
[font color=\'#000000\']Lin,
As I understand it, the electric charge in each individual photodetector has to be 'dumped' or cleared before the next exposure can occur. What you seem to be saying is that within one exposure and one press of the shutter button, there can occur two electronic exposures. This might be an advantage speed wise. With autobracketing there's probably an additional delay between one shot and the next. But essentially, it's the same process of blending 2 images with different exposures. Providing the camera with 'blending' technology could be a time saver, but the software would have to be [/I]very good.[/font]
-
[font color=\'#000000\']
I figure to always go out with a tripod and to do bracketed shots if I have the slightest doubt. Blending in Photoshop is quick and easy and bomb-proof with the $20 Miranda plug.
But bracketing with multiple exposures and blending the exposures must, by default, mean that the sensor has the latitude of "holding" the full dynamic range you are able to achieve with the blend of two or more exposures.
The point we are discussing concerns theoretical ways of doing exactly what you achieve with the PhotoShop action with a single press of the shutter rather than multiple presses. Also, it would elilminate the absolute necessity of using a tripod and the care necessary in seeing that there is no mirror slap or even the slightest movement of the camera to avoid even minute motion blur.[/font]
-
[font color=\'#000000\']
The point we are discussing concerns theoretical ways of doing exactly what you achieve with the PhotoShop action with a single press of the shutter rather than multiple presses.
But the biggest question about that is what it has to do with the original intent of the thread?
Alain Briot's latest addition to the Aesthetics and Photography series, Determining The Best Exposure (http://luminous-landscape.com/columns/determining-exposure.shtml) goes into determinging a film/sensors DR. Anyone curious how their camera stacks up can try that. I havn't tried it with my 300D, but in experience it seems there is around 5 stops of light with an extra stop of info at either end of the histogram (when in RAW).
audibeara6, as to your question and asuming, based off the fact that your here, you will be doing nature photography:
The 1Ds obviously has the advantage in resolution and the fact that it's full-frame wich comes in handy for wide angle lenses.
The 1D MkII has a 1.3 crop so corner sharpness/etc isn't a great concern with that camera. It also helps with zooms if that's your thing. The 1DMkII is much faster. Not a great concern when shooting a mountain, but can be a very nice thing when shooting birds and animals or that UFO that just sucked up your photo assistant. From what I've read and seen of samples, it also is the best camera at high ISO shooting with the least amount of noise I've seen in a digicam (agian helpfull for that UFO since they usually come at night).
There's been alot of squaking at the DPreview forums of the 1DMkII being soft. It uses a stronger AA filter but with some strong sharpening you can pull the detail back out (http://www.outbackphoto.com/reviews/equipment/Canon_1d_mkII/canon_1d_mkII.html).
I don't know what the smallest aperture AF will work at on either camera but the 10D is f/8.[/font]
-
[font color=\'#000000\']
In any case 1d and 1ds are both pro cameras and fortunately not burdened with a kluge like automatic in camera double exposure and blending.
I think I understand that the 1D and 1DS are "pro" cameras (I have the 1D, 1DS, 1D Mark II, 10D, D30, DCS-760, Kodak Pro back and a couple dozen fixed lens digicams), I make my living with the pro models and collecting and experimenting with digicams is a long standing hobby. The process may be "kluge" to you, but anything which improves my ability to get a better image is important to me.
I'm sorry if I have confused you. I'm using the Fuji and Hewlett-Packard models as examples only because they happen to have the features and are convenient to discuss. The theoretical use model I'm presenting is intended for the pro sensor. The pro sensor does indeed have the capacity to "hold" the full dynamic range - you're not fully understanding the process here or we have a different understanding of the term "hold". Whether this type feature is desirable or not is a marketing issue. I like the idea and apparently you don't and that's fine - that's why there are features on cameras which I don't care for and features which you don't care for - there is no perfect solution to everyone's satisfaction.
As for your concerns with camera shake etc., and the speed of the capture, that's an engineering issue and not of any serious importance. I can hand hold and get perfectly focused images at 1/30th second and even slower with my better stabilized lenses. For landscape work This is more than ample time for processing.
It's obvious that you have no interest in this theoretical approach and see no utility in it and you've made that perfectly clear. I do see promise in it and I've made that perfectly clear and since the thread has migrated far from the original question and the tone is changing from a sharing of ideas and potential to argumentative, I'll stop here.
Regards,[/font]
-
[font color=\'#000000\']No, if you are in manual mode and set the shutter speed lower than the selected burst speed, such as 8 frames per second, it seems to simply use the speed you have set. It appears in that sense to work exactly like the 1D.[/font]
-
[font color=\'#000000\']
The question is, could a device capable of measuring time in billionths of a second be incorporated into a small camera?
Faster and smaller go together; fast has to be small. As for billionth of a second, that's a pretty long electronic time interval nowadays. Consider that a $200 Sony Playstation 2 at about 7 gigaflops is 7 times faster than the first generation of $10 million+ Cray supercomputers. I wouldn't worry about speed being a bottle neck.
I had an equally crazy idea, but even more useful for a landscape photographer. Put an LCD element in front of each CCD sensor element. Each LCD element can be controlled to be totally transparent or quite dark. Do a test exposure with full transparency of all LCD elements. Then you can instruct your camera to darken the LCD elements where the test exposure CCD elements showed the most brightness. By controlling the selective darkening of bright areas, you could end up with exactly the final contrast you want, with no black shadows and no clipped highlights, all in a single final exposure. Only for tripod exposures, obviously. I suppose the test exposure and then adjusting the LCD pattern to exactly compensate for no highlight clipping before doing the real exposure could all be done automatically and fast enough so that you could do perfect contrast shots even without a tripod. High end cameras could have different LCD response curves for different latitude control effects and you could do something like autobracketing, but where the camera takes several quick shots with different latitude compensation algorithms. The key is the LCD grid, the rest is trivial software issues.
Let's patent both ideas, but the clever Japanese engineers can work out all the minor technical aspects and just send us frequent big royalty checks for the vital genius of the ideas.[/font]
-
[font color=\'#000000\']I just received my new Canon 1D Mk2 as an upgrade from 1 yr old Canon 10D. I cannot speak for the 1Ds as I have never used one, but the 1D Mk2 is a truely amazing camera.
The high speed motor drive and buffer allows my to shoot Raw and not missing any action shots (mostly critters in the woods). The shutter is noiser than my 10D (which is the quietest SLR I've ever owned), but will all of the other features, this is not a complaint.
I've only had the 1D Mk2 for 2 days so I'm still reviewing the images and learning how to use all of the features of this marvelous camera. So far, the images are looking real fine. However, I can see that the various 1GB CF cards for my 10D get filled up much faster with the 1D Mk2's larger file size. It looks like I'll be ordering a couple of 2GB CF cards soon.
The real test is when I have a few nice shots to print to 13 x 19 on my Epson 2200 printer using Hahnemuhler papers. I'm looking forward to this.
Cheers.
Bud James[/font]