dcands wrote:
They should given the lack of good technical support, their inferior focusing system, a pathetic IR on flash units, poor low light focusing on lesser expensive dslr's and I could go on. A D300 for a few hundered more makes far better sense. I know I've not been happy with Canon for some time now and am definately making the jump now. Canon has completely destroyed my faith in them as a company reguarding any kind of decent support besides "lip service" Just has been my experience with them.
Sorry to be negative but I believe others have the right to know either side ...Show more →
Well, I'm not taking it quite that badly but rather, I think it will put a rocket under Canon's throne to do something about it. Their next itteration had better be more competitive or they will lose their top spot. For the time being, Canon still has price point advantages.
It's your choice whether or not you choose to jump ship but maybe Nikon are simply looking at blurring the lines between prosumer and professional bodies whereas Canon whould prefer to have a line in the sand. Otherwise we'd all be buying 1D MkIIIs, to which Nikon currently has no answer.
To be fair, until we actually see actual pictures from production examples of both cameras, all we are doing is counting holes in the ceiling tiles. From my POV, Nikon has a FPS, AF and battery grip integration advantage over Canon but I wouldn't be interested in changing brands because of it. The only other thing I'd like to see included is micro focussing.
The 40D is not a dud but Canon have been spanked and they know it.
Ron Hew wrote:
Stan, then the result will be a smoother tonal instead of higher DR. Rite?
Correct. It's totally possible that they improved the DR, too, but that can be done without increasing the bit depth, and vice versa. You can pack an infinite dynamic range into just one bit - not to say it would be pretty
given that is supposedly a "lite" version of the 1DMK3 I predict you will be stoked with it.
My experience of the IQ out of the MK3 has been fantastic.
Pretty excited about it can't wait, I was hoping to have it on Monday but was reminded it's a holiday. Some of the people are having it overnighted so we should hear some this coming weekend
DynoMoHum wrote:
Highlight priority... You know... often I wish I had shadow priority... In fact, I personally rarely need highlights preserved.
So what? Expose to the right - and you are done.
Clear was one of the first PictureStyles available for download, and Emerald came later. Using DPP I have found uses for both.
What I wonder is this, has anyone used the new Picture Style Editor yet? It supposedly lets you create a whole new Picture Style, which means changing the actual color curves, not just changing the parameters as you can now do. Supposedly it is on the software CD in the 40D box, but I've not seen anyone claiming to use it yet.
stanj wrote:
You are missing the point. The 14 bits don't mean that you got two bits of extra dynamic range. It's more like you take a ruler that has both inches and millimeters on it; it's still the same length, but the side in millimeters will be "more accurate" if you are limited to using integer values and no fractions, simply because there are more millimeters in a meter than there are inches. The mm side of the ruler isn't any longer.
I'm not missing the point at all. There are either 12 or 14 bits between "black" on your monitor or print medium, and "white" on your monitor or print medium. I fully understand that. It's also the case that the sensor can only resolve so many stops of brightness.
Everyone endlessly complains about dynamic range in digital but guess what? In (slide) film the lightest highlight is clear and the darkest shadow is opaque - kind of like digital. The difference is in how film responds to exposure, compressing the highlights to compress more of the light variance in the scene between transparent and opaque. This is exactly what I'm talking about being implemented in newer cameras. Instead of just the post-processing tone curve against 12-bits of RAW info, the camera can apply compression DURING the exposure. Since there's more dynamic range, you need more discrete light levels to avoid posterization. Hence 14-bit RAW.
The flaw I'm pointing out is everyone arguing about 14-bit being "pointless" and suggesting that everyone chop off the bits and compare is limiting their discussion to the world of post-processing without considering new actions taken by the camera to increase dynamic range during the actual exposure process.
I can say for a fact that people would be screaming if the effective dynamic range (with HC on) went from 5 stops to 7 stops (approx numbers pulled from the air) and more bits weren't used to represent the tonal values and it created posterization. Instead now we have a few people complaining that the bits are of no use yet these same people probably obsessed over 100% zooms from 5 different RAW convertors to pick the best.
If you increase dynamic range, whether through multisampling tricks or other means, you need more bits to represent the tonalities between "white" and "black". Otherwise you risk posterization and people would add to the Canon bashing for sacrificing quality in their output.
If you generally work in 8-bit sRGB, maybe you'll never see a difference. In larger colour gamuts with difficult colours, however, the difference may become apparent more quickly. I've been shocked recently with how many of my decidedly amateur photos exceed even the AdobeRGB gamut. The wider the gamut, the greater the difference between "neighbouring" numeric colours. It's even more important in wide gamuts to have a greater number of bits to prevent posterization.
Right now we're at the beginning of whatever technology has been added to use 14-bit RAW. By the time the 5DmkII, 50D or whatever comes out, they may have honed these techniques to get another stop or two of dynamic range if enabled.
After more consideration, I have a couple more points (and then I should probably stop posting to the 40D master thread).
Many say film has more dynamic range. I'm not a big film guy, but I've read enough to understand that there are two properties - there are properties of the film, and properties of the development process, which contribute to how the picture looks and what dynamic range it captures. Like digital, film has an absolute "white" and an absolute "black". It also has a noise floor below which images are unpleasant.
I think many of us are familiar with the rationale for "shoot to the right". The sweet spot of digital sensors begins with white. Half of the tonal values are represented in the top stop, a quarter in the next stop, an eigth in the stop below that. In theory, without accounting for noise, the number of bits equals the number of stops for which data can be collected. Of course the data down that low isn't usable because the 12th stop in a normal 12-bit RAW would be 1 or 0. I've heard people quote that recent digital cameras have about 5 useful stops.
The light collection process uses cells on the sensor that are like buckets. Photons land in the buckets and are counted. I don't know how often they are counted, but for the sake of discussion sensors could count and empty the buckets any number of times per second. Up until now the buckets have simply been counted and added in a simple fashion. We also know that all digital photos are linear (no "film" properties at all) and require a tone curve to look natural (the "development" process).
What I'm suggesting may be now happening is that at the collection and count interval, simple summation may not be all that's done anymore. Decisions can be made how to count or compress based on the observed values. This is now adding a "film response" component to the digital equation.
If we simply increased the number of bits from 12 to 14 and counted linearly, we would get 4 times the number of tonal values in the first stop and we would have two stops further into the shadows without accounting for sensor noise. With sensor noise, I would wager that we may not be able to use 2 full stops deeper into the shadows with acceptable results. At this time perhaps Canon has chosen to focus on using those bits in the "useful" bright end of the spectrum. Since blowing highlights is a cardinal sin in digital, a design that uses compression can save us from blowing those highlights. With HC programming, instead of having the sensor output depict normal results (half in top stop, quarter in second, eigth in third) the algorithms and extra bits could be used to give us equal tonal values in the first n stops and then reduce by half every stop after that. Shadows are the same but highlights preserve detail.
An eight bit example. Normal is 128-64-32-16-8 values for the first five stops. 9-bit with my theories applied 128-128-128-64-32-16-8 or 7bits,7bits,7bits,6,5,4,3. I'm just pulling numbers out of the air.
Read http://www.luminous-landscape.com/tutorials/expose-right.shtml and imagine for a minute that you could remove the assumption that digital sensors are linear, and the number of tonal values per stop didn't have to reduce by half each stop. That's what I think HC + 14-bit is doing.
Aug 29, 2007 at 07:35 AM
Allen Maestas Offline • • • • • • Upload & Sell: Off
Hey, thanks for the heads up Mike! I just got mine ordered as well, should ship out Friday and in my hands Tuesday or Wednesday. I cancelled my Amazon order, so that's one notch up for everyone
Amazon.com is now listing a release date of Aug 30, 2007 on the body only & kit.. Though my pre-order is still showing a Delivery estimate of Sept 21 (I ordered on Aug 20)
"Availability: This item will be released on August 30, 2007. Pre-order now! Ships from and sold by Amazon.com. Gift-wrap available."
I was in the same boat - shipping September the 20th and arriving the 24th - changed the order to one day shipping just to see what would happen and that updated the estimated shipping date to Aug 30th, arrival 31st. We'll see.
Mike Lucas wrote:
I was in the same boat - shipping September the 20th and arriving the 24th - changed the order to one day shipping just to see what would happen and that updated the estimated shipping date to Aug 30th, arrival 31st. We'll see.
Yeah, it looks like Amazon is anticipating earlier arrivals from Canon. I just went to check my pre-order from 8/20 and after changing my shipping from Free to Standard the arrival date moved up to 9/5. ...but I moved it back to Free shipping and the arrival date is about 9/8. ....Where's the grip?
