40D vs 1D MkIII for BIF

[Leonard Malkin]

I have focusing problems with my Canon 40D and 100-400 lens when trying to photograph flying birds (except the large slow ones). The 1D Canon series supposedly has better AF. Is there anybody out there with specific experience with the 1D MkIII and 20,30,40,50D who can tell me if the 1D really makes a big difference?

[Doug West]

Leonard,

What are some of your techniques that you use when it comes to flight photography?

For example, are you using AI Servo mode? Do you try to zoom in as the bird gets
closer vs lets say starting off at 400 (which makes it a little harder to track)? What
AF points are you using?

These are just some of the items that could contribute to trying to capture a bird in
flight.

Doug

[Axel Hildebrandt]

Welcome to BPN, Leonard! The Mark III has two processors and can therefore focus faster and - ideally - more accurately. One problem is the 100-400, it focuses noticeably slower than the 400f/5.6. For BIF many people like the 50D.

[Arthur Morris]

Most BIF problems are due to operator error. Once we hear back from you, I can suggest several sources for info.

[Leonard Malkin]

[QUOTE=Doug West;234967]Leonard,

What are some of your techniques that you use when it comes to flight photography?

For example, are you using AI Servo mode? Do you try to zoom in as the bird gets
closer vs lets say starting off at 400 (which makes it a little harder to track)? What
AF points are you using?

These are just some of the items that could contribute to trying to capture a bird in
flight.

AI Servo, continuous 6.5 fps, either center or all points (doesn't seem to make any difference), bird is located and centered at 100, then zoomed to 300-400 keeping bird centered (if I can).

[Doug Brown]

Hi Leonard and welcome to BPN! BIF are challenging, and as Artie says, operator error is a very real problem. The Mark III is better than the 40D but not as good as the 50D at tracking BIF in my opinion. Perhaps if you'd post some sample photos along with tech specs, we could be of more assistance to you.

[Tim Dodd]

When aimed accurately both the 40D and 50D are good at tracking BIF. IMHO the problems arise when the bird is difficult to follow accurately and it is too small in the frame to cover multiple focus points at once. At these times it can fall through the gaps between points and focus gets lost. With the 40D and 50D as soon as the focus point is misplaced the camera will almost instantly seize upon whatever then falls under the focus point. In order for the camera not to suddenly go off into a world of its own you must have very quick reactions to release the AF-On button, reacquire focus and then resume tracking. With smaller targets you cannot use all points successfully, because they are spaced too far apart, so you are compelled to use only a single focus point. That leaves little margin for human error in aiming the camera at the subject.

(Disclaimer - since getting my 1D3 in January I have not yet shot BIF with it. However, I have had plenty of practice shooting my running black dog with the 1D3, 40D and 50D. The comments below are based on my experience tracking my dog, often coming towards me at full tilt, not BIF.)

Where the 1D3 shines, by comparison, is in being able to customise how quickly the camera searches for a new target, and in having AF expansion/assist points that closely surround the chosen focus point. I may be going out on a limb here but I also find the 1D3 focus to be snappier than the 40D/50D. In other words, the 1D3 AF is superior to the 40D/50D for fast moving subjects coming towards the camera on a somewhat unpredictable path - such as running dogs and, I suspect, BIF.

I have uploaded an album of running dog examples. Those from my 40D and 50D are so marked. The remaining 133 unmarked images are from my 1D3. A few are static shots but were taken in AI Servo mode - it's pretty much all I ever use. EXIF is available for all the images by clicking "more info" on the right when viewing an individual image. As far as I know, the only edits there may be in these examples are WB and levels, but many have no edits at all. Sharpening and NR will be at defaults as spurted out of Lightroom. There are no crops that I am aware of. Here is the album....

http://picasaweb.google.com/EezyTige...KLlgIbEyITQxgE#

I find the 1D3 beats my other cameras not just for AF ease of use and performance but also in IQ. I can use higher ISOs with less noise penalty and thus maintain higher shutter speeds and/or buy a little extra DOF or hit the lens sweet spot more easily. The larger, brighter viewfinder and greater field of view also make tracking easier.

All that said, I still have a long way to go in improving my own tracking skills and getting the most from my gear. Practice, practice, practice - that's what I need.

[Leonard Malkin]

Hi Leonard and welcome to BPN! BIF are challenging, and as Artie says, operator error is a very real problem. The Mark III is better than the 40D but not as good as the 50D at tracking BIF in my opinion. Perhaps if you'd post some sample photos along with tech specs, we could be of more assistance to you.

This is the first time I ever heard anybody claim the 50D is better than the MkIII for BIF (and also better than the 40D). Why should this be so? The 50D I understand uses the same 9 point AF as the 40D while the MkIII has many more points. In reading the Canon manual, I think it says that with multiple focus points, initial focus is on the nearest object and as long as the object is covered by one of the other focus points (in AI Servo) it will stay in focus. This has led me to believe 2 things - 1. use all the focus points and 2. the more focus points the better.

Also, I'm sure operator error is a problem although I wouldn't call it an error, it's just that at my age (72) I'm just not fast enough to keep the fast fliers in place. This at least seems to be one of my problems anyway. If you or Artie can solve that problem, I would be so grateful.

I will post photos as soon as I can figure out how to do so.

[Arthur Morris]

Hi Leonard,

re:

This is the first time I ever heard anybody claim the 50D is better than the MkIII for BIF (and also better than the 40D).

Obviously you have not been reading BIRDS AS ART Bulletins; they are free and packed with great info, as is the blog.

Why should this be so?

Don't know.

The 50D I understand uses the same 9 point AF as the 40D while the MkIII has many more points. In reading the Canon manual, I think it says that with multiple focus points, initial focus is on the nearest object and as long as the object is covered by one of the other focus points (in AI Servo) it will stay in focus. This has led me to believe 2 things - 1. use all the focus points and 2. the more focus points the better.

Wrong, wrong, and wrong. For most flight photography, using only the central sensor is best. The latest blog post deals with when to use AAFPS for flight photography: http://www.birdsasart-blog.com/

Also, I'm sure operator error is a problem although I wouldn't call it an error, it's just that at my age (72) I'm just not fast enough to keep the fast fliers in place.

I am 62 and I am in the same boat. I ain't as fast as I used t be. The great news is that I am still breathing.

This at least seems to be one of my problems anyway. If you or Artie can solve that problem, I would be so grateful.

:) :) :)

You can find some good stuff on flight photography here: http://www.birdsasart.com/bn260.htm There is lots more in the Bulletin Archives which has a great Google Search feature. There is lots there in recent Bulletins on the 50D, and lots more on flight photography in my CD book, ABP II.

[Jackie Schuknecht]

Hi Leonard,
I have the same combo as you and I used the back button to focus in AI Servo with all nine points activated and then pressed the shutter button at the front. My first time out I had pretty good success. HH too. The only thing I can think of is that maybe your bird(s) was/were quite small in the frame to have been locked on by any sensor point?

[Roger Clark]

Leonard,

While I don't have nor have I used a 1DIII or 50D, I have/had D60, 10D, 20D, 30D, 40D, 1DII and 5DII. I also have the 100-400, although it has sat on the shelf for years as I found it both slow and soft. I replaced it with a 300 f/4 L IS. I also use a 500 f/4 L IS and 300 f/2.8 L IS.

I suggest you get a faster AF lens as I feel you are being limited by the 100-400.

In my experience, there is no comparison on AF speed and tracking of the 1DII versus the 10D, 20D, 30D, or 40D. Like others have noted, one major factor I have found with the 1DII is that if I slip my single focus point off the bird, I can quickly get it back in focus, even after locking onto something in a complex background. On 10D, 20D, 30D, 40D, if I slip off with a complex background, it is rare that I can ever get the bird back in focus before it is gone or landed. Plus the 45 1D series AF points are wonderful and I like the user interface of the 1D bodies better. But I do like the 5DII and it became my prime camera on a recent trip to Tanzania, getting better focus on animals in action, though I did not do too much BIF (encountered too many other animals, like lions and cheetahs and leopards) ;-) .

Here is my technique for BIF which does differ from what has been posted so far:
http://www.clarkvision.com/photoinfo...with.autofocus

[Arthur Morris]

Hi Leonard,
I have the same combo as you and I used the back button to focus in AI Servo with all nine points activated and then pressed the shutter button at the front. My first time out I had pretty good success. HH too. The only thing I can think of is that maybe your bird(s) was/were quite small in the frame to have been locked on by any sensor point?

If the birds are so small in the frame that they are between the AF sensors you should not be pressing the shutter button!

[Jay Gould]

Artie, unfortunately some of us are very linear and also unfortunately need it spelled out.

You wrote, and have written numerous times:

"For most flight photography, using only the central sensor is best. The latest blog post deals with when to use AAFPS for flight photography: http://www.birdsasart-blog.com/"

Roger, on his website in a wonderful description of Autofocus wrote:

"Use one focus point: A common mistake in action shots is to use multiple Auto Focus (AF) points, then the camera confuses the subject and often locks on to something you don't want. Use one focus point and keep that AF point on the moving subject so the AI servo can track it. The AF system has trouble with contrasty back lit subjects, so try and choose well lit subjects and/or backgrounds that are smooth (like birds with blue sky background). " http://www.clarkvision.com/photoinfo...cus/index.html

Leonard, as an aside, trust me (;)), buy Artie's BAA1 and BAA 2, and study Roger's website.

Continuing,

Artie, you wrote on the blog:

"When attempting to photograph erratic flight or action, I will often choose AAFPS, All AutoFocus Points Selected as in most cases this make it easier to track the bird or to maintain focus on two birds that are interacting."

I will wear the dunce cap: please expand the differences between "flight" and "erratic flight or action" so that us dunces will have some idea when to use a single sensor and when to use AAFPS.

Everyone else can jump in regarding this distinction too!

Thanks,

[Ákos Lumnitzer]

please expand the differences between "flight" and "erratic flight or action" so that us dunces will have some idea when to use a single sensor and when to use AAFPS.

Everyone else can jump in regarding this distinction too!

Hi Jay,
I am far from being an expert, but I would read it as:

Flight: a straight, predictable path which your subject takes making panning relatively straightforward. (then why do I get so many crap images?) :D

Erratic/Action: This is when you cannot comfidently predict that the path will remain the same, as in say a bird like a Cockatoo zig-zagging above the field as an example, at high speed? I guess that having multiple points would at least give you half a chance of getting something sharp as it would be hard enough to keep the bird in the frame with the center AF point locked on in the SAME spot as say in Flight.

At least that is how I would interpret it. Surely the Maestro will chip in his wisdom soon enough. :)

[Tim Dodd]

"For most flight photography, using only the central sensor is best. The latest blog post deals with when to use AAFPS for flight photography: http://www.birdsasart-blog.com/"

I just had a quick read of the Blog regarding settings for the 50D. One piece of advice I would take issue with is that of leaving HTP enabled. IMO this is poor advice. If you shoot raw it is very questionable whether you gain any useful benefit, but you surely cause some detriment to IQ. If you shoot JPEG then I would consider it reasonable to use the function, BUT, only in conditions where you have critical highlight detail that you want preserved, in a scene of high dynamic range.

HTP works by using an ISO value within the camera that is 1 stop below the ISO value you "think" you are using. This is why, for example, the lowest ISO value you can use with HTP enabled is 2oo. The camera tells you that you are using 2oo (not 200) but it is actually using 100 ISO. This means that the capture at the sensor (the raw capture) is underexposed by 1 stop compared to a "correct" exposure. This underexposure is what preserves your highlight details and greatly reduces the risk of clipping. If you shoot JPEG, or review the preview image/histogram when shooting raw, the camera will apply a special tone curve to brighten the shadows and mid tones, bringing them up to "correct" exposure levels, while leaving the highlight tones far less pushed. The obvious problem this generates is that your "deliberate" underexposure and subsequent brightening raises noise levels in the shadow and mid tone areas. This is the very antithesis of good ETTR practice. Canon themselves state in the manuals that HTP "may" increase shadow noise. Now you know why. There is no "may" about it.

If you insist on shooting raw with HTP enabled then the benefits to be realised, if any, really only come if you use DPP to process your files. DPP knows about HTP and its intentions and has the "special" tone curve built in to "develop" the raw image correctly. Other software, such as Lightroom, is aware of the HTP flag but does not have the special tone curve built in. Thus, all it does is to note the 1 stop underexposure and double all raw pixel values, getting you back to where you would have started but with all the extra noise. It may also send the highlights into clipping, which of course you can pull back for yourself, but why bother? Why not expose properly in the first place and avoid unnecessary noise?

If your subject and scene has low dynamic range, or is generally dark or mid toned in nature (no obvious highlights), such as a thrush standing on grass, then you really really really do not want to be using HTP, whether you shoot raw or JPEG. My strong advice is to have HTP disabled as a default, never use it if you shoot raw, and only use it for JPEG shooting when the conditions warrant it.

If you want to check this out for yourself then, using raw (raw + JPEG if you like), shoot a controlled scene under controlled lighting, both with and without HTP enabled. Review your raw histograms using the free software available here - http://www.cryptobola.com/PhotoBola/Rawnalyze.htm - page to the bottom for the download link. I'm sure you'll find that all HTP does is underexpose your raw capture by 1 stop.

p.s. I haven't checked this out on the 50D, because I no longer use HTP, ever, but on the 40D you were unable to use 3200 ISO with HTP enabled. This is because 3200 ISO is not a real ISO. A shot taken at 3200 ISO is actually shot by the camera at 1600 ISO (thus underexposing by 1 stop at the sensor) and then the camera doubles the pixel values digitally to simulate an exposure 1 stop brighter. That isn't far off what HTP does anyway, except you have a different tone curve applied in each case. It would make no sense to try to use HTP at 3200 ISO because you can't apply both tone curves simultaneously and you would not want to be capturing at the sensor at 800 ISO, digitally doubling the pixel values and then adding an HTP tone curve on top of that. That would be madness.

This actually brings me on to another point. Why use expanded ISOs, if you shoot raw? All you are doing is recording mathematically faked pixel values instead of the values actually captured by the sensor. The mathematical doubling of pixel values might push pixels into clipping that were actually just fine when captured. IMO it would be better to exposue at a "real" ISO value, underexposing if need be to maintain shutter speed, and then fine tune the image in your raw processing software. Then you have more control over whether or not clipping occurs. FWIW, I have ISO expansion disabled on all my cameras. I have no use for faked ISOs.

While we're on the subject of faked ISOs, my understanding is that for all bodies below the 1 series each of the ISOs between the full stop ISOs - e.g. 125, 160, 250, 320 - are actually shot within the camera at the next nearest true ISO either above or below and then the camera pulls or pushes the exposure to fake the actual ISO you thought you were using. i.e. 125 ISO exposures are captured at 100 ISO and then the pixel values are mathematically raised by a factor of 5/4 to simulate the higher ISO. Personally I'd rather shoot at 100 ISO and find that 1/3 stop extra exposure somewhere else. Alternatively, as a raw shooter I'd rather shoot at 100 ISO, underexposing by 1/3 stop (just as the camera does anyway) and then sort out the final image in my raw software. This approach on its own could buy me 1/3 stop extra highlight headroom without the full stop loss that HTP brings about.

With my 1D3 I am happy to use any ISO necessary, other than the expanded ISOs.

[Grady Weed]

This actually brings me on to another point. Why use expanded ISOs, if you shoot raw? All you are doing is recording mathematically faked pixel values instead of the values actually captured by the sensor. The mathematical doubling of pixel values might push pixels into clipping that were actually just fine when captured. IMO it would be better to exposue at a "real" ISO value, underexposing if need be to maintain shutter speed, and then fine tune the image in your raw processing software. Then you have more control over whether or not clipping occurs. FWIW, I have ISO expansion disabled on all my cameras. I have no use for faked ISOs.

While we're on the subject of faked ISOs, my understanding is that for all bodies below the 1 series each of the ISOs between the full stop ISOs - e.g. 125, 160, 250, 320 - are actually shot within the camera at the next nearest true ISO either above or below and then the camera pulls or pushes the exposure to fake the actual ISO you thought you were using. i.e. 125 ISO exposures are captured at 100 ISO and then the pixel values are mathematically raised by a factor of 5/4 to simulate the higher ISO. Personally I'd rather shoot at 100 ISO and find that 1/3 stop extra exposure somewhere else. Alternatively, as a raw shooter I'd rather shoot at 100 ISO, underexposing by 1/3 stop (just as the camera does anyway) and then sort out the final image in my raw software. This approach on its own could buy me 1/3 stop extra highlight headroom without the full stop loss that HTP brings about.

With my 1D3 I am happy to use any ISO necessary, other than the expanded ISOs.

I did not know the above. Very interesting if so. I always shoot RAW and I do use DPP to make initial adjustments to the RAW file before converting to the tiff and making final adjustments in PS. A very interesting read Tim.

[Tim Dodd]

I did not know the above. Very interesting if so. I always shoot RAW and I do use DPP to make initial adjustments to the RAW file before converting to the tiff and making final adjustments in PS. A very interesting read Tim.

Have a look at this article....

http://forums.whirlpool.net.au/forum...fm/999699.html

I am not the author but I might well have been. I turned it up fairly quickly with a search for "fake ISO 100 200 400" in Google and then selecting a likely looking candidate at random from the returned results.

There are various discussions about fake ISOs on forums, such as DPReview and POTN. There is a guy who goes by the nickname of Gaborscsh or Panopeeper on different forums who has studied this phenomenon very carefully and presented raw file analyses to back up his claims. I trust what he says.

[Roger Clark]

Have a look at this article....

http://forums.whirlpool.net.au/forum...fm/999699.html

I am not the author but I might well have been. I turned it up fairly quickly with a search for "fake ISO 100 200 400" in Google and then selecting a likely looking candidate at random from the returned results.

There are various discussions about fake ISOs on forums, such as DPReview and POTN. There is a guy who goes by the nickname of Gaborscsh or Panopeeper on different forums who has studied this phenomenon very carefully and presented raw file analyses to back up his claims. I trust what he says.

Except that significant parts of what he says is wrong. For example, the dominant noise source in people's images is not due to the camera electronics at all. It is photon noise, the inherent noise in the photon signal itself. Then there are subtle things (that lead to basic errors), like noise is actually less with smaller pixels, contrary to what people say all the time. The difference is that with small pixels, the signal is even lower, so what we see as noisier images is signal-to-noise ratio that is lower (both signal and noise are decreasing with smaller pixels, just signal is decreasing faster). Then the idea that intermediate ISOs are simply digital multiplications is not correct for all cameras. See for example, Emil's (Emil is a BPN member) very technical write-up: http://theory.uchicago.edu/~ejm/pix/.../noise-p2.html

I suggest if you want to discuss these issues more, we start a new thread.

Some additional references if you want to delve into it (also check out Emil's other pages from the above link):

Does Pixel Size Matter:
http://www.clarkvision.com/imagedeta...el.size.matter

Sensor performance data:
http://www.clarkvision.com/imagedeta...rmance.summary

Roger

[Tim Dodd]

Roger,

This thread is about the Canon 40D and 1D3. The remarks I've made should be taken specifically within the context of this thread and hence those two cameras. My comments are not intended to apply to cameras outwith the Canon brand. I should also make it doubly clear that unless otherwise stated, my remarks are intended to apply to raw shooters, not JPEG shooters.

With that out of the way, are you saying that the things I have said and the article I have referenced are invalid, even if only in parts, for the two cameras under discussion here? If I have made an error can you please help me out and explain in lay terms whereabouts I have made a factually incorrect statement with respect to the 40D and 1D3? I'm more than happy to be corrected, and to learn, if I am wrong. I just happen to think that I am not wrong.

If I can try to help things along, which of the following assertions do you disagree with....?

1. Expanded ISO settings are nothing more than digitally multiplied derivations of data captured at lower ISOs (for all Canon cameras released to date, including both H and L expansions).

2. Intermediate ISO settings between the full stop ISO settings of 100, 200, 400 etc. are nothing more that digitally multiplied/divided derivations of data captured at the nearest one of the "full stop" ISOs (for the 40D).

3. HTP effects an underexposure of 1 stop at the sensor, via a deliberate reduction in actual ISO and than digitally manipulates the JPEG data (not the raw) to generate an image file which captures increased dynamic range in the highlights.

4. If you shoot to raw only, HTP accomplishes little or nothing that you could not do for yourself by underexposing by one stop with your ISO setting and then fixing with your own custom tone curve in editing.

5. Processed within DPP on default settings, a shot taken with HTP enabled will result in higher shadow noise than an image shot with the same nominal exposure settings but with HTP disabled.

If I've missed a key point then please feel free to comment on that as well.

Thanks.

[Doug Brown]

There's a lot of misinformation floating around the internet on the 50D. Don't mistake group-think for fact. Just because Canon doesn't advertise the fact that AF is improved on the 50D doesn't mean that it hasn't been improved. A growing number of pros, including Artie, Brutus Ostling and I, have stated that the 50D is Canon's best autofocusing camera.

[Jay Gould]

There's a lot of misinformation floating around the internet on the 50D. Don't mistake group-think for fact. Just because Canon doesn't advertise the fact that AF is improved on the 50D doesn't mean that it hasn't been improved. A growing number of pros, including Artie, Brutus Ostling and I, have stated that the 50D is Canon's best autofocusing camera.

Doug, from an amateur (ME) to an acknowledged professional (YOU):

Earlier in this thread Artie honestly stated he did not know why the 50D is focusing better than the 1DIII; do you? If you don't that is OK too.

To us amateurs, on its face it doesn't make sense because the 1DIII has substantially more points of focus than the 50D, and while the 1DIII is using Dual DIGIC 3 sensors, the 50D is using a single - supposedly upgraded - DIGIC 4 sensor.

You are a professional photographer; I am a professional "word person" (a retired trial lawyer) and an amateur photographer.

When you say "just because Canon doesn't advertise the fact that AF is improved on the 50D doesn't mean that it hasn't been improved", that statement potentially contains both a statement of fact and a statement of opinion.

As an amateur with an open mind, when many professionals say that the 50D is the best BIF camera they have ever used, I accept that without further question, and for me it becomes a statement of fact upon which I am going to act. More than that, my amateur hopefully logical mind says that any activity similar to BIF, e.g., fast moving sports, will also be photographed better by the 50D than the 1DIII. Fortunately, both bodies top out at ISO 3200, so we do not have to bring in additional variables.

However, you didn't make a comparison in your statement to either of the cameras which are the subject of this thread as pointed out by Tim to Roger; you did not compare the 50D to any other camera in a thread having nothing to do with the 50D.

Therefore, can it be reasonably assumed by some readers of your statement that you are saying that the 50D AF is better than the 40D AF - one would hope so since it is the next generation (the 40D is a DIGIC 3) - and that the 50D AF might not be as good as the 1DIII (that is what an unknowing amateur would assume)?

There is a saying that "understanding is the booby prize of life"; you don't have to understand everything - sometimes just accept it.

I have and I thank all of the professionals here - I will use the 50 or the 60D (hopefully the AF will be even better with the 60) - however; I don't understand why it is so.

Obviously, some of the other readers/members still have not accepted the professionals' statements and they need further analysis.

Sooner or later you have to decide upon whom you are going to rely if you are not going to do your own in depth research. I have decided to rely heavily on some of the experts on BPN without doing further in depth research.

Leonard said: "This is the first time I ever heard anybody claim the 50D is better than the MkIII for BIF (and also better than the 40D)."

Artie responded: "Obviously you have not been reading BIRDS AS ART Bulletins; they are free and packed with great info, as is the blog."

Artie's bulletins and Roger's website are becoming two of my bibles; my choice.

However, as soon as I am comfortable with the decision regarding the 50D being the holy grail of BIF, I look further down the Photography Gear Discussions - and lo and behold - ;) - just 5 days earlier there is a thread pertaining to "1DIII Autofocus settings for BIF" and some of the experts that are saying that the 50D is the holy grail are also indicating their settings for the 1DIII.

Doug, while it is expected that professionals like yourself would have all of the bodies - 50D, 5DII, 1DIII - and I do not know how else to say this, when you respond to the 1DIII BIF thread by providing the settings without more, sorry but it does confuse us amateurs.

It suggests that you use both cameras for BIF; however, if the 50D is the holy grail why do you use the 1DIII for BIF? And, if you do not use the 1DIII any longer for BIF and you are simply providing information to the initiator of the thread, please, clarify your information for us dunce amateurs.

I would think that is why someone like Leonard - or me! - might say

"This is the first time I ever heard anybody claim the 50D is better than the MkIII for BIF (and also better than the 40D). Why should this be so? The 50D I understand uses the same 9 point AF as the 40D while the MkIII has many more points. In reading the Canon manual, I think it says that with multiple focus points, initial focus is on the nearest object and as long as the object is covered by one of the other focus points (in AI Servo) it will stay in focus. This has led me to believe 2 things - 1. use all the focus points and 2. the more focus points the better."

I have accepted without technical understanding that the 50D AF is better than the 1DIII, and I have accepted it as fact because you and others have said so; is it a fact? :confused:

Sorry for the ramblings.

[Roger Clark]

I have accepted without technical understanding that the 50D AF is better than the 1DIII, and I have accepted it as fact because you and others have said so; is it a fact? :confused:

Jay, I can't comment on the 50D because I've never even held one. But what I can say is that Canon has a history of bringing out better stuff on low end cameras before bringing it out in higher end cameras. This seems astonishing to me, but it has happened several times. One would think companies would bring out better stuff on the high end to entice people to move up. But then maybe they use the lower end to "beta" test state of the art features to later put, in more refined versions, in higher end cameras. Who knows? So it doesn't surprise me that people say the 50D is the best. That means (we hope) that the 1D Mark IV will be at least that good too and probably better.

[Roger Clark]

Roger,

This thread is about the Canon 40D and 1D3. The remarks I've made should be taken specifically within the context of this thread and hence those two cameras. My comments are not intended to apply to cameras outwith the Canon brand. I should also make it doubly clear that unless otherwise stated, my remarks are intended to apply to raw shooters, not JPEG shooters.

Tim, you referenced a specific web site,
http://forums.whirlpool.net.au/forum...fm/999699.html
and I was referring to the information in that web site, not anything specific about what you said.
But let's explore the implication below.

With that out of the way, are you saying that the things I have said and the article I have referenced are invalid, even if only in parts, for the two cameras under discussion here? If I have made an error can you please help me out and explain in lay terms whereabouts I have made a factually incorrect statement with respect to the 40D and 1D3? I'm more than happy to be corrected, and to learn, if I am wrong. I just happen to think that I am not wrong.

The referenced article has numerous wrong concepts, and that leads to additional invalid concepts, or the magnitude of effects, and this includes the 40D and 1D3.

If I can try to help things along, which of the following assertions do you disagree with....?

1. Expanded ISO settings are nothing more than digitally multiplied derivations of data captured at lower ISOs (for all Canon cameras released to date, including both H and L expansions).

It is not clear that all Canon cameras have this method of expanded ISOs. One needs to actually test each
camera. I have not seen data on all cameras that prove this. And we can discuss the implications of
using high ISOs. I'll do that below.

2. Intermediate ISO settings between the full stop ISO settings of 100, 200, 400 etc. are nothing more that digitally multiplied/divided derivations of data captured at the nearest one of the "full stop" ISOs (for the 40D).

Again we do not have info on all cameras (at least I have not seen it). But below I'll discuss
the implications.

3. HTP effects an underexposure of 1 stop at the sensor, via a deliberate reduction in actual ISO and than digitally manipulates the JPEG data (not the raw) to generate an image file which captures increased dynamic range in the highlights.

I agree.

[quote=Tim Dodd;235937]
4. If you shoot to raw only, HTP accomplishes little or nothing that you could not do for yourself by underexposing by one stop with your ISO setting and then fixing with your own custom tone curve in editing.

Yes, but with a caveat that I'll discuss below.

5. Processed within DPP on default settings, a shot taken with HTP enabled will result in higher shadow noise than an image shot with the same nominal exposure settings but with HTP disabled.

I agree, but let's look at the implications.

We start with the sources of noise. The article you referenced said electronics. But that is incorrect. The dominant noise source in most images (pretty much everything but low level astrophotography) is photon noise. With photon noise, the noise scales as the square root of the signal. So when you change ISO by a factor of two, noise only changes by square root 2 to 1.4x. So in the case of ISOs that are "fractional" the increase in noise is even less. For example, if you use 1/3 f/stops, and you go from ISO 400 to ISO 320, the noise changes only 11%. You would be hard pressed to see the difference. But if you want to use that intermediate ISO for a particular reason, go ahead because then you can evaluate the camera data (e.g. the histogram and the look of the image). If instead you use exposure compensation, you may see an overly dark (or bright) image, impacting your judgment of the image data.

Similar with the HTP mode. Yes it is true the noise in the shadows would be increased, but only 40%. If you are concerned about highlights, that might be a compromise you want, otherwise shoot HDR. If you "do it yourself" you underexpose one stop, so the on screen review will be dark and your histogram shifted. That will make evaluating the image more difficult. So if you underexpose one stop, you still have the issue of more noise in the shadows. So there are pluses and minuses to each method and people need to understand there are pros and cons to each, but in any case they are minor in the big scheme. If noise scaled linearly, it would be a greater factor, but it is not.

High ISOs. Eve if high if ISOs above 1600 are just digitally multiplied data, it would produce the same result as with analog gain before the A/D converter. At these ISOs. the photon count is very low and noise is dominated by 2 sources: photon noise in mid-tones and highlights and read noise in the shadows. Read noise does not depend on ISO, so the resulting noise in the image has little contribution from other camera electronics. Thus it is effectively the same, whether a digital multiply or not. So, if you are in a situation where you need high ISO and you want to review your images on the camera, go ahead and use the high ISO setting.

If, however, you need high ISO and want to keep as much dynamic range as possible, do not go above 1600 ISO, and boost in post processing. Those doing astrophotography, generally do not use ISO above 1600, as there is no real advantage.

In summary, photon noise is greater than and therefore masks a lot of the lower level noise from camera electronics. I say go ahead and use the features of the camera, whether intermediate ISOs, HTP, ot expanded ISOs if you feel there is a benefit, especially in evaluating results in the field. If you are doing the highest precision work that requires the lowest possible noise, then you might want to avoid these things, but I bet most people will not be able to tell the difference with most cameras unless you are doing astrophotography that includes stacking and averaging many images.

[Doug Brown]

Hi Jay! I'm not privy to any inside info regarding the 50D. However I've used the 50D to the tune of 50,000 to 100,000 flight frames and the 1D Mark III for over 300,000 flight frames. It's that experience that helps me form my opinion about the 50D.

I did make a statement comparing three bodies in my original comment on this thread (you're reading my second post), and here's what I said: "The Mark III is better than the 40D but not as good as the 50D at tracking BIF in my opinion."

As you point out, I did post a comment about my settings for the Mark III in an earlier thread. I love the Mark III; it is my primary body. This current thread is about comparing the AF of the Mark III with the AF of the XXD bodies, and I answered the OP's question. You are asking an entirely different question that has more to do with why I use other bodies if I think the 50D has such great AF. There is more to the decision of which body to use than how the body AFs on BIF. The Mark III does an entirely satisfactory job in the AF department, but it's not as good as the 50D in my opinion. Where the Mark III excels is in the pixel quality of the images I get; the images simply look better to my eye. It's also built like a tank, and I can be pretty hard on my gear. The Mark III will AF at f/8, another plus when you're using a 2x on a big lens. I've never had a problem with the Mark III's AF; I get a high percentage of keepers when photographing BIF. I'm willing to live with a slightly lower keeper rate in most circumstances because of all the other things I like about the Mark III.

I have posted extensively on both the Mark III and the 50D, including lots of flight images using both bodies. And I have favorable opinions of both bodies. The 50D excels at AF for BIF but the Mark III takes the overall crown.

I hope this helps clarify my opinions for you.

[Grady Weed]

Doug.

I have read this thread from the beginning. I have to say I appreciate your clarification here. To me experience beats a study every time. Your sound logical explanation cuts through the mumbo jumbo. After all if you have a study but in the field experience makes it no good, then....My wife wants the 50d. She does a lot of research and is pretty careful with money. This type of post and your response is a big deciding factor in making the purchase.

[Jay Gould]

Hi Doug, totally agree with Grady - thanks for the clarification.

Now I am even more confused which two of the three to purchase "today"! :confused:

The landscape/low light side of me says to definitely purchase the 5DII.

The "how much can you carry" side - we are on the road 50% of the time living out of a suitcase and in addition me living out of a camera bag - kinda leans to the 50D (because of the weight) although it is now clear to me from your clarification and the posts of the other professionals that other than AF the 1DIII is a much better overall body compared to the 50D.

Thanks again; more to ponder.

[William Malacarne]

1D series cameras are moisture proof. the others are not....some lenses are but not all.

Bill

added

5D MKII is 28.6oz. (810g) body only
50D 26.1 oz. (740g) body only
1D MKIII 40.7oz. (1155g) body only

[Grady Weed]

Jay

If money is not an issue, then go for the MKlll, you can't beat the 10fps and the high pixel count. Then get the 5Dll. You get full frame and the video is to die for. I own the 5D, original, and it is great. It only has 3fps but I used it for birds and landscapes in the beginning, never regretted it. I also use the MK2n. It is 8.5 fps and only 8.5 megapixels, but it rocks for speed. I use the 300 2.8, 28-135, 100-400 and the 100mm macro. I can put it all in a Tamrac expedition 8 series back pack. I have taken it on airplanes, throw it in my truck's back seat, leave it on the ground as I photograph etc.

You wont find better pro body cameras by Canon, the 50d, MK2n or the 5D's. They are all 4 great. Just a matter of money or preference. At least in my opinion. Remember, everybody has their own.

[Ákos Lumnitzer]

In my humble opinion it matters not whether one picks a 50D or 1DMkX (with 100 AF points - sometime soon :D ) the crux is that in the right hands, both will take exceptional "in flight" images regardless of AF performance. Heck, my long outdated 30D is still coping at times.

Though the 1D's higher frame rate will provide a better chance of getting a good wing position in frame. :) I'd go for an older 1DMkIIn or MkIII over a newer 50D or 60D anyday.

I'll shut up now. LOL.

[Peter Hawrylyshyn]

Leonard/Jay -
If you read some of the earlier post here at BPN and the Canon white papers, there are significant differences in the AF sensors on the cameras which might help explain some of the differences people are describing.

FROM CANON: The EOS 40D camera's completely redesigned nine-point AF sensor provides cross-type AF measurement at all nine focusing points for maximum apertures up to f/5.6, and for the first time in any EOS camera, the central AF point offers enhanced precision for both vertical and horizontal subject contrast when using EF or EF-S lenses featuring maximum apertures of f/2.8 or faster.

This is what's in the 50D except that it's using "faster" Digic 4 vs Digic3 in the 40D. This AF system is NOT in the 1d Mk 3 or 5D Mk2. What becomes important is the maximum apeture at which cross-type AF sensors are still active/functional. Although the 1 D Mk3 has 45 sensors and slower Digic 3, the 19 more sensitive cross-type AF sensors are only active with lenses faster than f2.8 (which excludes most telephoto lenses used for BIF). WHen using an f4 lens on a 1D Mk3, only the central AF point is the senstive "cross-type" sensor, all others are less sensitive horizontal only sensors.

But then you have to add in : using central sensor only (difference will be Digic 4 vs 3) or all sensors - "ring of fire" (when sensor type would become important) and finally your technique/BIF experience. Having stood beside a BPN pro/moderator and having swapped back and forth between 50D vs 1D Mk3 trying to photograph purple martins/grackles- it ain't the camera, it's the shooter.

PH

[Jay Gould]

Leonard/Jay -
If you read some of the earlier post here at BPN and the Canon white papers, there are significant differences in the AF sensors on the cameras which might help explain some of the differences people are describing.

FROM CANON: The EOS 40D camera's completely redesigned nine-point AF sensor provides cross-type AF measurement at all nine focusing points for maximum apertures up to f/5.6, and for the first time in any EOS camera, the central AF point offers enhanced precision for both vertical and horizontal subject contrast when using EF or EF-S lenses featuring maximum apertures of f/2.8 or faster.

This is what's in the 50D except that it's using "faster" Digic 4 vs Digic3 in the 40D. This AF system is NOT in the 1d Mk 3 or 5D Mk2. What becomes important is the maximum apeture at which cross-type AF sensors are still active/functional. Although the 1 D Mk3 has 45 sensors and slower Digic 3, the 19 more sensitive cross-type AF sensors are only active with lenses faster than f2.8 (which excludes most telephoto lenses used for BIF). WHen using an f4 lens on a 1D Mk3, only the central AF point is the senstive "cross-type" sensor, all others are less sensitive horizontal only sensors.

But then you have to add in : using central sensor only (difference will be Digic 4 vs 3) or all sensors - "ring of fire" (when sensor type would become important) and finally your technique/BIF experience. Having stood beside a BPN pro/moderator and having swapped back and forth between 50D vs 1D Mk3 trying to photograph purple martins/grackles- it ain't the camera, it's the shooter. PH

Thanks Peter, I am starting to move from dunce to grade school. :)

Faster means wider aperture/more light/faster shutter speed. Please, lets apply to my lenses: 16-35 f/2.8; 24 - 105 f/4; 70 - 200 f/4; 180 Macro f/3.5; 100 - 400 f/4.5-5.6. I may trade the 100-400 for a 400 f/5.6; that decision would be based on the f/5.6 being a better lens for my mix since the weight/price are relatively similar.

Of course, its the shooter and not the camera; we can all agree that there is no substitute for experience. Having said that, if you have the opportunity to purchase good gear, you may as well purchase the best/right gear for the situation.

Taking your information one point at a time, please try to expand in lay terms; let me apologize in advance because I am going to ask some very very basic questions:

"The EOS 40D camera's completely redesigned nine-point AF sensor provides cross-type AF measurement at all nine focusing points for maximum apertures up to f/5.6, and for the first time in any EOS camera, the central AF point offers enhanced precision for both vertical and horizontal subject contrast when using EF or EF-S lenses featuring maximum apertures of f/2.8 or faster. This is what's in the 50D except that it's using "faster" Digic 4 vs Digic3 in the 40D."

Both the 40/50Ds have eight outside point and one central point.

You have said "cross-type AF" when referring to all nine point, and "vertical and horizontal subject contrast" when referring to the central point.

1. What is the difference if any without reference to f/ stop?

You have also said that the outside eight provide AF for maximum aperture up to f/5.6, and the central point provides enhanced focusing for maximum apertures of f/2.8 or faster.

2. Does that mean that the eight outside point do not AF a lens faster than f/5.6 which means that only the 100 - 400 at f/5.6 is AF by the outside eight? In other words, I must focus all of my lenses since they are faster than f/5.6 by the center sensor. Do I have it backwards?

3. The central sensor: it does provide AF at f/5.6, f/4, f/3.5; it provides enhanced focus for f/2.8, f/1.4 etc. What is the difference between the AF provided between f'/5.6 - f/3.5 and the AF provided at f/2.8 and faster? I thing the answer is the same as the answer to #1, correct?

"Although the 1 D Mk3 has 45 sensors and slower Digic 3, the 19 more sensitive cross-type AF sensors are only active with lenses faster than f2.8 (which excludes most telephoto lenses used for BIF). WHen using an f4 lens on a 1D Mk3, only the central AF point is the senstive "cross-type" sensor, all others are less sensitive horizontal only sensors."

4. 45 - 19 = 26; with an f/4 lens as you have explained, are 44/45 sensors effectively the less sensitive horizontal only sensors?

5. When you say "faster than f/2.8" does that mean that any f/2.8 (my 16 - 35) is treated the same as an f/4 or slower lens? To say it another way, unless you have an f/1.4 lens, all of the sensors in the 1DIII are slower than the 50D sensors. Perhaps this explains why the 50D AF is better for BIF in as much as the both the central sensor and the eight outside sensors will operate with slower lenses than the 1DIII's 45 sensors. If I have said this correctly it would certainly argue for the 50D in combination with the 5DII.

6. "But then you have to add in : using central sensor only (difference will be Digic 4 vs 3) or all sensors - "ring of fire" (when sensor type would become important)": When I read this line I feel like I must have missed something/incorrectly analyzed to this point. Haven't you said that both the central sensor, and the "ring of fire" in the 50D permits the use of a wider range of lenses, and also slower lenses and compared to the central sensor/"ring of fire" in the1DIII?

Finally, while camera build, waterproofness, and other ergonomic factors are important, my observation is that cameras are not being kept long enough by the general public/high use amateurs/ME to worry about wearing them out. Your group of shooters, consumate professionals shooting 100,000s of images each year will wear out the best cameras through day in and day out use; my group most likely will upgrade before the camera is all used up.

Thanks for taking the time to help me work through these issues.

Cheers, Jay

[Alfred Forns]

... with regards to some statements regarding AF and some lenses I've been liking the versatility of the 100-400 and being more impress lately. Will post some images of a speeding Pileated woodpecker in not so good light speeding by and nailed with the Mk3 with 100-400 lens.

I think the key is being able to get the center point on the bird and keep it there.

[Jay Gould]

... with regards to some statements regarding AF and some lenses I've been liking the versatility of the 100-400 and being more impress lately. Will post some images of a speeding Pileated woodpecker in not so good light speeding by and nailed with the Mk3 with 100-400 lens.

I think the key is being able to get the center point on the bird and keep it there.

:p Shame on you :D Just when I was accepting that the 100-400 was "perhaps" to slow with its AF and I should switch to the 400, even though the 100-400 is a much more versatile lens, along you come and stir the brew the opposite way! Thanks; actually it is what I want to hear as thus far I have loved my 100 - 400.

I assume that you were using the lens at the 400 range which means f/5.6; you had a burst speed of 10fps vs 6.5fps; and the central sensor was more than sensitive enough for your situation.

And, as everyone is quick to point out; it is the shooter and not the gun that at the end of the day makes the difference. :D:D

[Tim Dodd]

When people say the 50D AF is as good or better than the 1D3, can someone please describe the shooting conditions in which they find this to be the case? The reason I ask is that because in my experience I find the six focus assist points of the 1D3 to be a tremendous asset and their absence from the 50D makes it the weaker camera, AF -wise, for what I shoot - running dogs. I have no doubt that aimed accurately, and with each camera only using one focus point the 50D will be at least the equal of the 1D3, but when it is tough to keep the focus point aimed with perfect accuracy (dog's head bobbing up and down and only the teeth/tongue offer sufficient contrast against black fur) that is when the power of the 1D3 shines.

Maybe BIF, which I do try to shoot as well, is a different challenge, so....

- Is the 50D as good or better than the 1D3 for shooting birds coming head on towards the shooter?
- Is the 50D as good or better than the 1D3 for shooting birds moving across the field of view?
- Is the 50D as good or better than the 1D3 for shooting birds moving erratically - ducking and diving and looping and swooping?
- Is the 50D as good or better than the 1D3 for shooting birds that are somewhat small in the frame?
- Is the 50D as good or better than the 1D3 for shooting birds that are moving quickly at close distance?
- Is the 50D as good or better than the 1D3 for shooting birds when light levels are low?
- Is the 50D as good or better than the 1D3 for shooting birds when light levels are high?
- Is the 50D as good or better than the 1D3 for shooting birds with strong backlighting?
- Is the 50D as good or better than the 1D3 for shooting brightly lit birds against dark backgrounds?
- Is the 50D as good or better than the 1D3 for shooting birds when subject contrast is low?
- Is the 50D as good or better than the 1D3 for shooting birds when subject contrast is high?
- Is the 50D as good or better than the 1D3 for shooting birds against a clear sky?
- Is the 50D as good or better than the 1D3 for shooting birds against a busy background?
- In conditions where you would use automatic AF point selection are the 9 points of the 50D as good or better than the 45 point ring of fire on the 1D3?

I'm not trying to stir the pot here. I am genuinely looking for answers from those more experienced than me. I have both cameras and mostly use my 100-400 for BIF. In my limited experience I do not find the 50D to be the better BIF camera. Maybe it can equal the 1D3 in the right conditions, but I believe there are a number of situations where the 1D3 will make it easier to go home with more keepers. I'd like to ignore price and frame rate and other things from the discussion - just concentrate on AF performance.

One more point - for those who feel the 50D bests the 1D3 for BIF, have those people had the latest "recall" fixes applied to their 1D3 bodies and, if so, has that altered the perception at all?

Thanks for your input.

[Jay Gould]

Tim, I just looked on the Canon site at the 1DIII sensor: 5 lines of points - 7, 10, 11, 10, 7. What are you referring to by "I find the six focus assist points of the 1D3 to be a tremendous asset"?

I hope everyone finds this as interesting as me!:D

BTW, you have both the 50D and the 1DIII; under what circumstances do you choose between the two cameras since they are both cropped sensors?

I will ask you a lens questions in another thread.

[Tim Dodd]

The 1D3 has a total of 45 AF points. You can select any one of 19 of those points as your focus point of choice. The other 26 points are not user selectable. They are "assist" points or "expanded focus points" that you can enable so that when you are shooting with a single chosen focus point you can also have either two (one either side) or six (completely surrounding, where possible) assist points also operating. It's basically like being able to make your chosen focus point larger than just the one point.

If you let the camera pick the focus point then it can use any of the 45 points it likes. For practical purposes there are no gaps between the 45 points, unlike the 50D which has quite large gaps in between the points.

See this extract from a guide to using the 1D3 AF system....



When I'm shooting my running black dog his fur does not make a suitable target for the AF to pick up on, at least not at a distance. There is no way I can reliably track an eye, or his teeth/tongue, with a single focus point. Using all points (ring of fire) would risk the AF picking up something else in the scene. The expanded focus points make it far easier, though still difficult, to track the dog successfully.

For completeness, here are the 19 selectable points highlighted within the total of 45 points. You can optionally limit the camera to only offer 9 points, either the centre point plus an inner ring of 8 points, or the centre point and an outer ring of 8 points. This can speed navigation to the point you wish to choose.



p.s. The 5D and 5D2 each have six AF assist points clustered around the centre point. These can't be selected individually, just turned on or off. There are no AF assist points for the other eight focus points.

[Tim Dodd]

BTW, you have both the 50D and the 1DIII; under what circumstances do you choose between the two cameras since they are both cropped sensors?

What counts for me for birding is not so much the sensor size as the pixel density and pixel quality. While they are both crop sensors the 1D3 does have a larger sensor (1.3X crop vs 1.6X crop), but far fewer pixels (10MP vs 15MP). Thus the pixel density of the 50D is significantly greater than that of the 1D3.

If we do some basic maths let's look at the figures....

50D sensor size = 22.3 x 14.9 = 332 square mm. At 15MP that's 45k pixels per square mm.
1D3 sensor size = 28.1 x 18.7 = 525 square mm. At 10MP that's 19k pixels per square mm.
5D2 sensor size = 36.0 x 24.0 = 864 square mm. At 21MP that's 24k pixels per square mm.

So, for any given focal length the image projected onto the sensor will be the same physical size in mm regardless of which camera you use. Assuming you can use a low ISO and noise is not an issue, nor dynamic range, and your lens is sharp and focused accurately, you will be able to capture far more detail with the 50D than either of the other two cameras. You will have far more pixels covering your subject with the 50D. The 5D2 will capture a bit more detail than the 1D3 but is still a long way short of the 50D.

So, in ideal shooting conditions, where AF performance is no concern, the 50D will get you more detail than either of the other two cameras, for a given focal length. Now, if your glass is long enough to fill the frame on the 5D2 then the tables will be turned because quite simply 21MP is clearly more than 15MP or 10MP. But the price you pay is in needing an 800mm lens on the 5D2 to match the reach of a 500mm lens on the 50D. That's a lot of money and a lot of weight.

Another way to look at the issue is to consider, for example, that your lens is just long enough (let's say it's a 100-400 for the sake of argument) to fill the frame on the 50D with your subject. You end up with all 15MP being useful and contributing towards the final image. Now suppose you replace the 50D with the 1D3 under identical conditions. You would need to crop away the excess from the scene, reducing your 10MP down to something even less. You would actually end up with only 6.6MP for your scene vs 15MP from the 50D. The 5D2 would be somewhere in between, but nearer the 1D3 end of things.

Depending upon what your final purpose is for the image - printing, to what size?, at what quality? or online viewing only, to what size? - you may find 6.6MP quite good enough. But what if you actually had to crop that 50D image, to say 1/4 of the whole frame. Suddenly your 50D image only has 3.75MP and your 1D3 image would have less than 2MP. Well for printing at 300ppi a 2MP file will only get you pixel dimensions of approx 1730 x 1150 so you'd just about squeeze a 6 x 4 print from that file. That would also only just fill an HDTV or WUXGA disply, when viewed at 100%. I'd suggest you really want to end up with nearer 8MP after cropping if you want to make a worthwhile print or fill a large, high resolution display, which you can then resize to 50% and sharpen. Should look nice :)

Things are also more complicated because there is the issue of noise and DR, and 1D3 pixels and 5D2 pixels will individually look better than 50D pixels, but taken as a whole the image from a 50D can look just fine and dandy.

So, how do I choose what to shoot with? Well, if the action is lively (needs competent AF), or the light is poor (high ISO), and my glass has adequate reach (no shortage of pixels) then the 1D3 is my tool of choice. If the action is moderate to nil (perched or walking birds) and the light is good (low ISO) and I'm struggling for reach with my lens then I'll use the 50D. Of course, there are no hard and fast rules. It's all about judgement and trade-offs and what I'm likely to be shooting mostly at the time.

e.g. I visited a bird reserve recently with my 1D3, 50D, 100-400 and 1.4X teleconverter. I arrived with the lens on the 1D3 but saw there was nothing either interesting enough or close enough in the air to bother with BIF. So I turned my attention to perched birds, and we're talking finch sized birds here. Without the benefit of a cover or hide the 1D3 was pointless so I swapped to the 50D. Even that was still not long enough to get a worthwhile capture so on went the teleconverter. Here's an example from the day, shot with my 50D at 560mm and f/8 (wide open). This is cropped to 1/4 of the whole frame and then resized to 50%. It isn't a great shot, and I'm sure a prime lens wold have netted me improved sharpness, but it is better than anything I would have got with the 1D3, I think....

http://www.flickr.com/photos/tdodd/3373463720/sizes/o/

[Alfred Forns]

Hi Jay If I was using at the 400 range only would go for the 400 5.6

btw sometime back I gave up on the 100-400 since it had to be rebuilt often from the barrel being frozen. Not good for crawling on the beach so I replaced it with the 70-200 and 400 5.6. Now I find it so versatile I don't want to give it up.... and again think its a good performer AF wise. I do not argue that a fixed 300 2.8 will be faster but might be good looking at the overall picture.

Tim for comparing lenses and cameras I like doing it in a different way. Just pick one up and use it then check the results. Megapixel wise don't get caught up in the numbers. Remember talking to John Shaw and hearing say how he made "very large" prints out of his D1 camera ... had 2.7 MP Also for testing AF do try birds for coming up to a conclusion. Will find both 50D and Mk3 very capable.

[Jim Neiger]

I have focusing problems with my Canon 40D and 100-400 lens when trying to photograph flying birds (except the large slow ones). The 1D Canon series supposedly has better AF. Is there anybody out there with specific experience with the 1D MkIII and 20,30,40,50D who can tell me if the 1D really makes a big difference?

Leonard,

I specialize in BIF images handheld. My lens of choice is usualy the 500mm. I have used all of the camera bodies that you asked about. I currently own 2 1D3 camera bodies, a 1D2N body, a 20D body, and a 10D body. The 1D3 is by FAR the best camera I have used for BIF. The 1D2N is second, with the 50D a close third. The image quality, ISO capability, buffer size, and frame rate are all clearly best on the 1D3. All of these features are very important for BIF images. IMO, the 1D3 is the clear choice.

[Alfred Forns]

Hi Leonard

Fully agree with comments by Jim and you won't get a better reference source for bif ... is about all he does :)

[Arthur Morris]

While this is interesting reading, much of it is above my head technically. I would humbly suggest that folks would be best off getting in the field and practice making images of birds in flight!

As far as comparitive tests for AI Servo AF, it would be impossible to create a test that would prove anything--there are just too many variables.

[Arthur Morris]

Roger,

This thread is about the Canon 40D and 1D3. The remarks I've made should be taken specifically within the context of this thread and hence those two cameras. My comments are not intended to apply to cameras outwith the Canon brand. I should also make it doubly clear that unless otherwise stated, my remarks are intended to apply to raw shooters, not JPEG shooters.

With that out of the way, are you saying that the things I have said and the article I have referenced are invalid, even if only in parts, for the two cameras under discussion here? If I have made an error can you please help me out and explain in lay terms whereabouts I have made a factually incorrect statement with respect to the 40D and 1D3? I'm more than happy to be corrected, and to learn, if I am wrong. I just happen to think that I am not wrong.

If I can try to help things along, which of the following assertions do you disagree with....?

1. Expanded ISO settings are nothing more than digitally multiplied derivations of data captured at lower ISOs (for all Canon cameras released to date, including both H and L expansions).

2. Intermediate ISO settings between the full stop ISO settings of 100, 200, 400 etc. are nothing more that digitally multiplied/divided derivations of data captured at the nearest one of the "full stop" ISOs (for the 40D).

3. HTP effects an underexposure of 1 stop at the sensor, via a deliberate reduction in actual ISO and than digitally manipulates the JPEG data (not the raw) to generate an image file which captures increased dynamic range in the highlights.

4. If you shoot to raw only, HTP accomplishes little or nothing that you could not do for yourself by underexposing by one stop with your ISO setting and then fixing with your own custom tone curve in editing.

5. Processed within DPP on default settings, a shot taken with HTP enabled will result in higher shadow noise than an image shot with the same nominal exposure settings but with HTP disabled.

If I've missed a key point then please feel free to comment on that as well.

Thanks.

Hi Tim, I read your post with interest. Sent your comments to Chuck Westfall at Canon. Here is his response:

"Canon has not publicly disclosed all the details of the method used to achieve Highlight Tone Priority. I can confirm that Canon's method involves lowering the gain on the image sensor and modifying the image processing algorithm to simulate the ISO speed set by the user, but the details of the algorithm are confidential. Since HTP does involve lower gain than standard camera settings, it affects both RAW data and JPEG images. Although it would be possible for users to come close to simulating the effects of HTP for RAW images by shooting at a lower ISO speed and altering the tone curve of the image during post-processing, this method requires advance planning and furthermore offers no advantages in terms of noise reduction compared to setting HTP on the camera. Finally, as stated in the camera instruction manuals, "noise in the shadow areas may be slightly more than usual" when HTP is activated. Therefore, it's basically up to users to decide if the benefits of HTP outweigh the potential drawbacks."

Best Regards,

Chuck Westfall
Technical Advisor/Professional Products Marketing Division
Consumer Imaging Group/Canon U.S.A., Inc

[Tim Dodd]

Hi, Artie.

Thanks for following up on this. Although Canon does not care to disclose the algorithm for the HTP adjustment, please note that Chuck confirms that HTP is effectively a reduced exposure through using lower gain (lower ISO) than the user believes they are using.

For somebody keen to "Expose To The Right" when shooting raw, this deliberate "underexposure" at the point of capture, which the camera conceals, can not be regarded as a good thing. On the one hand you are trying to just reach clipping point while the camera is purposely underexposing. The histogram you see, which is derived from the embedded preview image, is generated after the HTP tone curve has been applied. Thus, unless you know the truth of the matter, as I feel I do, you will be led falsely to believe that you are exposing to the right correctly when in fact you are doing nothing of the sort.

I have no objection to HTP, used when the conditions warrant it, but when you shoot to raw I do not believe its use is warranted and when you shoot a scene without highlights (of importance) it also serves no value. It is fundamentally an underexposure and I have rarely seen that described as a good thing.

My only concern is that in reading your blog you suggest that HTP should be left enabled as a default....

C Fn II-3/Highlight Tone Priority: Enable. As far as I know, there is no reason to leave this off as enabling it increases both highlight detail and dynamic range for tones lighter than a middle tone.

Take this image as an example....



The brightest thing in the scene is the sky. There are no highlights as such, certainly none in the subject, and certainly none that really require any detail to be preserved, just a bright area of the sky which I chose to preserve by metering carefully and exposing to the right, right on the limit. You will note that, with no edits here, nothing is clipped. Of course, with my philosophy I had HTP disabled. If I had enabled HTP this whole scene would have been captured one stop dimmer, and the raw pixel data would reflect that. What good would have been achieved? My subject is in the shadows, or the mid tones at least. It would have suffered far more than the sky from the effects of underexposing the capture and then pushing the pixels later. There is not only noise to consider but also tonal subtlety as well. Of course, the argument could be made that the photo was not worth taking, but that's a separate discussion. The point is that a subject/scene need not always contain detail in the highlights, and certainly none that needs protecting to the extent of one full stop of underexposure which puts at risk the IQ of the rest of the scene.

The first image in your blog currently - http://www.birdsasart-blog.com/wp-co...pa-bay-fl4.jpg - shares something in common with my example as far as light conditions are concerned, and even subject tone to some extent. By metering off the sky (the brightest thing in the scene) at +2 2/3 You have already secured your perfect ETTR exposure. You did not need the "protection" of HTP. Assuming you had HTP enabled you really only shot this at +1 1/3 stops and the camera, or your raw processing took care of the rest. That's basically one whole wasted stop of fine tonal detail, and an extra stop of noise. Now, at this size I couldn't tell whether you had HTP enabled or not, but if you were to enlarge this significantly then such weaknesses might become apparent.

In my opinion, HTP is a great "get out of jail card" for JPEG shooters when there are important highlight details to be preserved. For raw shooters there is no substitute for good ETTR exposure technique. If you want a safety net then don't push the envelope on ETTR, just ease back a bit. You do not need HTP enabled permanently.

Hopefully, if you understand and agree with what I am saying, and read between the lines of Chuck's reply, you now do know of a reason not to leave HTP enabled at all times.

Regards,
Tim.

[Arthur Morris]

Hi Tim, I have read what you have written and appreciate your time and efforts. As I have stated many times, I am not at all a technical person. I never do any testing. I have pretty much left HTP on with all of my bodies since the feature was introduced. My only defense is that my images from the past two years have not exactly sucked. Would they--if viewed under a microscope--have been of lightly higher quality? Perhaps. Or maybe not. Maybe your posts will get me to experiment with turning HTP off. Maybe not.

The image above was created at ISO 800 with HTP enabled with the 50D, the "very noisy" camera. There were a few flashing highlights in the sky. I am much more of a results person than a technical person.

[Tim Dodd]

I am much more of a results person than a technical person.

I'm quite up on the technical side of things. Now, if only I could get results like you do :)

Regards,
Tim.

[Arthur Morris]

I'm quite up on the technical side of things. Now, if only I could get results like you do :) Regards, Tim.

:) :D :)

I really feel that I do not have any great talent other than the willingness to work hard. And long.

[Roger Clark]

Hi Tim, I have read what you have written and appreciate your time and efforts. As I have stated many times, I am not at all a technical person. I never do any testing.

Hi Tim. I'll speak again from the technical side. I am a "technical person" and do a lot of testing. But I really only got into such testing when I wanted digital cameras for astrophotography where noise is a critical factor in getting a good image. But like I tried to show earlier, from a technical standpoint the noise differences from in between ISOs and with HTP use are small. I think you are trying to split too many hairs even though technically you are right. The large pixel DSLRs have huge dynamic range and take wonderful images. Such small noise differences are small in the scheme of trying to get good images in my opinion. So I agree with Artie, get out and make images. (Ha Ha, I say this as I'm chained to my computer, but at least I am working on some images that will make a big splash soon.)

Roger

[Arthur Morris]

Thanks Roger. We agree! And we are looking forward to seeing some of those big splash images.

[Roger Clark]

Thanks Roger. We agree! And we are looking forward to seeing some of those big splash images.

Thanks Artie. But don't hold your breath. They have to get published first which will take a couple of months.

[Arthur Morris]

No sweat. I have taken a very deep breath. Good luck!