timbop wrote:
This problem is not new. It is in the 20D and 30D as well, and is a fact of the larger AF sensor - which is much larger than the box on the screen. Think about it: it couldn't focus on the fence unless a sensor picked up the fence. Anyway, this is my major issue with the 40D, and is why the D300 or 1dm2 look so tempting.
A sound notion but I think flawed. The AF sensors do not cover a very big area - maybe up to three times the length of the indicator box but certainly nowhere near as big as the partial metering circle. Your conclusion about focusing on the fence assumes that the AF sensor covered it. An alternative conclusion is that the sensor did not cover it and that the AF just drifts.
In one series of shots that I took the focus captured a subject at the top of the frame - way past the upper AF sensor let alone the middle one that I was using.
knweiss wrote:
I was a bit surprised that exposure compensation must but enabled with the power on dial. There are the modes "Off", "On" and "/" which turns on exposure compensation. What's the point of that design decision? Who wants to use "On"? And by the way: I think I prefer the power on dial of the 300D.
The "On" position prevents the dial on the back from working and that in turn prevents unintentional exposure compensation. Sooner or later you will realise this is a useful feature
I'm not so convinced about the LCD resolution. Okay, the 3 inches are really nice compared to the tiny display of the 300D but I think the dpi could be much better.
Dpi is getting low now but the resolution exceeds the capability of the thumbnails that you can preview on the screen. They are only highly compressed 1.5 Mpx jpegs regardless of the resolution you used in the photo. That's why they can never look as sharp as the real thing even in 10x zoom. What is needed is a larger more detailed thumbnail to go with our larger image files. Only then will extra pixels be useful (except maybe in live view)
Regarding the viewfinder to be honest I didn't notice a big difference compared to the 300D.
That's a worry The 300D has half as many pixels on its LCD screen as any of the newer cameras like the 30D, 400D, 40D, 1D2, etc. It only looks good because the screen is so small, but there is actually less information there.
Side note: Yesterday evening I've also experimented a little bit with DPP 3.1. The program got little better since the last time I've tried but I still think the GUI is sub-standard and that there are still so many things missing. E.g. isn't it strange to offer a grid lines view but no function to rotate the image by an arbitrary angle!? I was also very surprised that it saves the "recipes" in the RAW file by default instead of a separate file. I.e your RAW files change! And why do they distinguish between RAW and RGB adjustments? Come on Canon, please put some more man power behind this software and let some top-notch usability people review the GUI!...Show more →
The raw file is changed but only the settings - none of the image data is changed. Nor are the as-shot settings. This approach won't work if you have write-protected the files but it's a lot more convenient than having separate .xmp files to hold the settings like other programs are forced to do because they don't have access to the proprietry part of the raw file format specification.
RAW adjustment panel settings are applied to the 12-bit/14-bit data prior to it being dumbed down to the output file resolution, whereas RGB panel settings are applied to the converted data (usually 8-bit but possibly 16-bit, and possibly also resampled to different number of pixels from the original).
Alan321 wrote:
I'm glad your copy is OK and maybe it is my curse. The photos show it is real. That it seems not to be universal might mean it's repairable. But then that applies to 1D3 too
I took several series of shots with my 1D2 today and it did not drift like the 40D did. Also using a single AF sensor.
- Alan
Yes, the smaller sensor is less likely to hit the background. The only way to focus on something is to have an AF sensor register it and the AF logic to sieze that object. Well, that or by blind chance the AF wandered randomly onto a setting that matched the distance of another object. The reason I try to make sure I fill the center circle is because you have to also take into account the variable of panning not exactly matching subject movement, particularly when shooting a burst at an erratic target. I like to fill the center circle, but that is just me because my panning technique sucks. I find if I give the camera a big enough target the shots in a burst (after the first) are in focus or close enough not to matter. When I push my FOV relative to the subject too far, then I have trouble with focusing on the background.
Can you please clarify what you meant earlier. I think some people are misinterpreting what you said earlier.
Are you saying that a low battery leading to a bad AI Servo operation occurs on 20D/30D? Or are you saying that if you don't fill the AF sensor with your subject the camera may focus on other objects?
Thanks.
The latter: the AF sensor is much larger than the center square, and in my case I am terrible at panning as well so I give myself more margin for error when i can
OK, here is a link to the 40D whitepaper. Look at pages 10 and 11 and see the "AF sensor arrangement" diagram and sensor coverage diagrams. The sensors are HUGE, and indeed the center sensor appears to cover much of the center spot meter circle, particularly since the diagonal sensors are not pictured in the second diagram.
Alan321 wrote:
I'm glad your copy is OK and maybe it is my curse. The photos show it is real. That it seems not to be universal might mean it's repairable. But then that applies to 1D3 too
I took several series of shots with my 1D2 today and it did not drift like the 40D did. Also using a single AF sensor.
- Alan
That's reassuring. If you buy the 40D some people will encounter "drifting" while in AI Servo mode. The lucky few won't. Hmmm....should I still consider the 40D? (Shrug)
Alan321 wrote:
That's a worry The 300D has half as many pixels on its LCD screen as any of the newer cameras like the 30D, 400D, 40D, 1D2, etc. It only looks good because the screen is so small, but there is actually less information there.
Even so, the 20D viewfinder was bigger than the 300D (0.9x magnification vs 0.8x) and the 40D is bigger again (0.95x a very slightly smaller sensor size). I still wonder whether that 0.95x refers to area or linear dimensions - the specs do not clarify that.
timbop wrote:
OK, here is a link to the 40D whitepaper. Look at pages 10 and 11 and see the "AF sensor arrangement" diagram and sensor coverage diagrams. The sensors are HUGE, and indeed the center sensor appears to cover much of the center spot meter circle, particularly since the diagonal sensors are not pictured in the second diagram.
The AF sensors are all "huge" but the image area they cover is not. Portions of the image have been reflected off mirrors and then put through a lens system that strectches and redirects and phase-splits it onto different areas of the AF sensor array. The corresponding area of the original undistorted image is still quite small relative to the physical size of the sensor.
You'll also notice that the cross-type sensors are more like t-types on the actual AF sensor array. The AF lens system takes that into account to make them act as if they were actually looking at small crosses of the main image.
Alan321 wrote:
Even so, the 20D viewfinder was bigger than the 300D (0.9x magnification vs 0.8x) and the 40D is bigger again (0.95x a very slightly smaller sensor size). I still wonder whether that 0.95x refers to area or linear dimensions - the specs do not clarify that.
- Alan
Yes, that's the technical side. But my subjective impression is that it isn't a big, noticable improvement (for me).
Alan321 wrote:
The "On" position prevents the dial on the back from working and that in turn prevents unintentional exposure compensation. Sooner or later you will realise this is a useful feature
On the 300D I have to press a button together with a dial to change the exposure compensation. Simple and safe. I never had a problem with it. And IMHO this solution is more elegant than an additional power on mode to prevent unintentional EC.
Alan321 wrote:
Dpi is getting low now but the resolution exceeds the capability of the thumbnails that you can preview on the screen. They are only highly compressed 1.5 Mpx jpegs regardless of the resolution you used in the photo. That's why they can never look as sharp as the real thing even in 10x zoom. What is needed is a larger more detailed thumbnail to go with our larger image files. Only then will extra pixels be useful (except maybe in live view)
Yes, that's what I want.
Alan321 wrote:
The raw file is changed but only the settings - none of the image data is changed. Nor are the as-shot settings. This approach won't work if you have write-protected the files but it's a lot more convenient than having separate .xmp files to hold the settings like other programs are forced to do because they don't have access to the proprietry part of the raw file format specification.
I prefer the separate file and expected that DPP does not rewrite the RAWs by default.
Wow, just got mine in the mail from profeel.com, they had it in stock and to me in a week while my order with amazon wasn't due to ship for another month.
Just downloaded the Lightroom 1.2 update with RAW support for the 40D, so all is good. This camera is a major step up for me, as I have been building up my lens inventory (70-200 2.8L IS, 17-55 2.8 IS) while getting by with the 300D. In comparison to that relic the 40D is
AMAZING (duh you are thinking... me too)
Went out today and shot a volleyball game. The 6.5 fps is amazing. ISO 3200 is quite grainy, but allowed me to get the fast shutter speeds I was after, and after putting them through even Lightroom 1.2's rudimentary noise reduction, the results were very usable.
My shutter sounds less treble-y than the clip posted on here several weeks ago. I really like it. The camera is great looking, solid, and so far I love it. FWIW.
Well, I jumped on the wagon and ordered it. Should be getting it next week.
Canon, be ashamed on yourself. You could have got 4500$ from me if you only you'd fix the AF issues in the 1D Mk III. Now you lost more than 3000$. That is no way to run a mega-business.
I can`t wait to see the same AF-test for the new Nikons! About half of the photos are out of focus.... even with the 1series. Now that`s not very encouriging!
Stefan
Thanks for posting the URL. Not too encouraging. Glad I haven't yet picked up my 40D. Allegedly some 40Ds are fine in AI Servo. SOme not.
I too, like Stefan, would like to see the same test done on Nikons. Particularly the D300. Hopefully we'll see this test/comparison done in December when the D300 is out.
I have read a lot of reviews on the 1D III and also the 40D and the one thing I noticed with Drew's tests, he is comparing the 1D II and the III yet he is using different backdrops. In the 1D III, the backdrop is green and I know that does affect exposure in the image when there is too much green and could this have an effect on the focus. In the 1D II tests he is using a backdrop that has no green.
If you are trying to do a comparison then you need to standardize everything. Just like the film days where is you were making test prints, you just cannot change more than one item at a time or the tests are meaningless.
Brit-007 wrote:
I have read a lot of reviews on the 1D III and also the 40D and the one thing I noticed with Drew's tests, he is comparing the 1D II and the III yet he is using different backdrops. In the 1D III, the backdrop is green and I know that does affect exposure in the image when there is too much green and could this have an effect on the focus. In the 1D II tests he is using a backdrop that has no green.
If you are trying to do a comparison then you need to standardize everything. Just like the film days where is you were making test prints, you just cannot change more than one item at a time or the tests are meaningless.
While his comparisons might not have as much weight as if he had used the same background his findings in the respective cameras still stick. Finding that in 4 of his tests there were focus problems 50% of the time. Not good.
Then you get people like MaxiKana who are getting good numbers. 50 out of a little over 1000 shots. And most of the 50 due to user error.
As I said. It looks like some 40Ds preform better than others. Less errors. Does this mean buying a 40D is a gamble? Some will get a lot of AI Servo troubles. Some next to none? (Shrug)
Oh no. My interest is in the 14 bit. This on it's own should produce way better looking images. I do not shoot sports and therefore the 1D III is way over what I need. I would dearly love a digital back for my RB but the cost is too much at this moment in time.
I actually am going to wait till the replacement 5D comes out and buy that. I want full frame. As a second body, I think my best route will be the 40D which is my interest in this thread. I just put my thoughts in to the thread because I see everyone just jumping over the so called negatives.
Yes, the 61/2 FPS would be nice for the occasional use but the image quality is what I need. What I should do is just buy the 1Ds III but I do not want to take out a mortgage to buy it. I think it is important to have 2 bodies that function and produce images pretty much the same and this is why I am going to hold off on the current 5D. To me the real change is 14 bit.
