40D autofocus

[Rob C]

And, actually, we don't really let the camera decide on what to focus on. We decide with the auto focus points.
Focus problems: I messed up a few shots in the beginning, when I had IS on while shooting at higher speeds. When you shoot quickly and move the camera, the IS may not be ready for every shot and still be working while you are already pressing the release.
[a href=\"index.php?act=findpost&pid=214890\"][{POST_SNAPBACK}][/a]

Two things:

a. you decide with the focus points;
b. the IS may not be ready etc...


I don´t think telyt misuderstood a damn thing - he got it spot on. You are lumbered with the points the camera can handle, or that you have selected, or something else that is less than having total control.

IS is another mixed bag, but let´s not start another Civil War.

Rob C

[woof75]

Two things:

a. you decide with the focus points;
b. the IS may not be ready etc...
I don´t think telyt misuderstood a damn thing - he got it spot on. You are lumbered with the points the camera can handle, or that you have selected, or something else that is less than having total control.

IS is another mixed bag, but let´s not start another Civil War.

Rob C
[a href=\"index.php?act=findpost&pid=214957\"][{POST_SNAPBACK}][/a]

Point the central spot at subject, focus, recompose, it aint rocket science.

[telyt]

You are misunderstanding my post.
[a href=\"index.php?act=findpost&pid=214915\"][{POST_SNAPBACK}][/a]

Care to explain? It sounds to me like you are misunderstanding my subjects.

Point the central spot at subject, focus, recompose, it aint rocket science.
[a href=\"index.php?act=findpost&pid=214968\"][{POST_SNAPBACK}][/a]

Ah, the old Focus-Lock-Recompose kludge.  My subjects move too quickly to make that work.  Even Canon recommends avoiding this technique for accurate focus with shallow DOF.

[woof75]

Ah, the old Focus-Lock-Recompose kludge.  My subjects move too quickly to make that work.  Even Canon recommends avoiding this technique for accurate focus with shallow DOF.
[a href=\"index.php?act=findpost&pid=214972\"][{POST_SNAPBACK}][/a]

But they move slow enough to focus manually on them?

[telyt]

But they move slow enough to focus manually on them?
[a href=\"index.php?act=findpost&pid=214973\"][{POST_SNAPBACK}][/a]

My guess is you haven't use a good manual-focus camera.  With a camera actually designed for quick manual focus it's not slow at all.

[Rob C]

Point the central spot at subject, focus, recompose, it aint rocket science.
[a href=\"index.php?act=findpost&pid=214968\"][{POST_SNAPBACK}][/a]

Yes, such has been suggested, but is a flawed method. If you move the centre of the viewfinder (for example) to focus on something at the far sides, then you must realise that you are focussing beyond the point you really want to have sharp.

It is nothing more than the difference in lengths of the sides of a triangle. The plane of focus of a lens is supposed to be flat: if you focus at the greater distance (as per my example of a subject at the far edges), then, when you turn the camera back to the straight ahead position you wanted, that focussed distance no longer holds and you find you have, in fact, focussed well BEYOND your subject. Unavoidable.

Not rocket science, as you said.

Rob C

[telyt]

Yes, such has been suggested, but is a flawed method. If you move the centre of the viewfinder (for example) to focus on something at the far sides, then you must realise that you are focussing beyond the point you really want to have sharp.

And in the time it takes to re-compose, the bird has turned its head or shifted its weight from one foot to the other changing the plane of best focus, or departed.

[Ray]

And in the time it takes to re-compose, the bird has turned its head or shifted its weight from one foot to the other changing the plane of best focus, or departed.
[a href=\"index.php?act=findpost&pid=215088\"][{POST_SNAPBACK}][/a]

This is always a problem with erratically moving subjects, whether trying  to manually focus or autofocus.

I don't have any fast telephoto lenses. My Canon 100-400/5.6 is virtually equally sharp at F8 and F11. The extra DoF (and latitude for focussing) at F11 is probably worth more than the extra sharpness at F8. If I owned a 400/2.8, I can imagine I'd have a new set of problems with regard to the accuracy of autofocussing and might well prefer to manually focus when using wide apertures.

In principle, a lens that can autofocus quickly and accurately should get you more (good) shots than a lens that has to be manually focussed.

Since most of my shots are landscapes with a fairly extensive DoF, autofoussing accuracy has not been a major issue. When I take people shots, street shots or candid shots, I'll generally use the 24-105/F4 on the 5D at F5.6 or F6.3. At these apertures (and focal lengths), autofocussing appears to be sufficiently accurate (on the 5D).

I tend to avoid the full aperture of F4 because I get a sense I'm losing sharpness in a noticeable way, but it might well be the case that the perceived loss of sharpness is due more to the misfocussing which is more apparent at F4, than to the lower MTF of the lens at F4.

It would take a lot of meticulous testing to sort out which lenses on which bodies in what lighting conditions and at which apertures have reliable autofocussing.

[telyt]

In principle, a lens that can autofocus quickly and accurately should get you more (good) shots than a lens that has to be manually focussed.

Even assuming an AF system is quick enough and accurate enough, it doesn't have focus points covering the entire picture area.  That's how F-L-R came up.  The point I want to be in sharpest focus - generally the animal's eye - often isn't at a focus point, and if it is at a focus point it's not there long enough to change to that focus point.

[ErikKaffehr]

Hi,

I think the problem is the menu item "View->Actual Pixels". In film days you needed a microscope to see whether an image was absolutely sharp, in my experience very few were at 40X. I normally used a 15X peak loupe but it was not really good enough for judging sharpness. Now in the digital world we just push Actual->Pixels and can peep an image 50 inches wide at 40 centimeters, with no grain.

Joke aside I actually used a Microscope to look at some slides, very few were really sharp at 40X or 100X.

Erik

Lighter moments aside, there is something worrying about these posts concerning autofocus.  It could just be my usual old reactionary problem, of course, but I have to wonder about the current need that so many people believe that they have for autofocus in the first place. More so than the invention of digital, I feel it to have been an enornmous answer looking for a problem which perhaps existed within the world of professional sports photography, though seeing so many fantastic images over my long(ish) life that were extant prior to said development, I even doubt that. Ditto war reportage.

So what happened along the way to produce this breed of photographers that can´t use their own eyes?

This is particularly worrying when one reads these posts questioning the use of the alternative Zeiss offering within the slr world. I have lived through a full-time, life-time career in photography and have never found myself unable to operate a camera because of focussing problems - never owned an autofocus lens, even. So what´s up with the new dependants, is it too much trouble to DIY; is it perhaps lack of confidence in your own eyes? Whatever, it is bloody disappointing.

Apart from the failure of the personal input, it´s also my belief that the current problems of lens build are mainly down to the need to produce lightweight materials than can move quickly under relatively low motor power. Has nobody felt the difference in quality between the current offerings and their non-af predecessors? What a price to pay for "progress".

Rob C
[a href=\"index.php?act=findpost&pid=212940\"][{POST_SNAPBACK}][/a]

[Ray]

Even assuming an AF system is quick enough and accurate enough, it doesn't have focus points covering the entire picture area.  That's how F-L-R came up.  The point I want to be in sharpest focus - generally the animal's eye - often isn't at a focus point, and if it is at a focus point it's not there long enough to change to that focus point.
[a href=\"index.php?act=findpost&pid=215104\"][{POST_SNAPBACK}][/a]

Nothing's perfect, If you focus on the eye, then recompose the composition slightly, you should still be sufficiently in focus. If the central, single focus point is far from the centre of the composition, then that's not ideal. I think Jonathan Wienke did some tests on this.

I'm a firm believer in the Photodo type MTF charts. I find it interesting that certain lenses can be very sharp at wide apertures.

Here's a Photodo MTF chart of the Canon 400/F2.8 non-IS version, probably no longer available. It's actually sharper at F2.8 than at F8, but those who rely upon autofocussing might never realise it.

[attachment=7925:attachment]

On the other hand, perhaps this lens is generally very accurate with autofocussing. Who Knows? The 400/2.8 IS is also sharper at F4 than at F8. Photodo have provided no results for this lens at full aperture, but the lens they tested is not quite as sharp at F4 as the non-IS version at F2.8.

The failure of such MTF charts to provide accurate information for the consumer, and presumably the reason why Photodo have discontinued such testing, is due to lens quality variability.

The manufacturer sometimes provides 'theoretical' MTF charts for its lenses, but none provide real, thorough and detailed MTF charts for each lens sold.

The excuse is, it's too expensive. That's a totally invalid excuse. It's a cop-out.

The real reason why manufacturers don't provide a full MTF test of each lens they sell is because it removes their power to bull shit. It's very simple. It's easy to understand. But, alas! perhaps it's not easy for the average consumer to grasp the significance of those charts.

We need more education.

[telyt]

Nothing's perfect, If you focus on the eye, then recompose the composition slightly, you should still be sufficiently in focus.

This still gives the bird plenty of time to move its eye out of the plane of focus.

I'm a firm believer in the Photodo type MTF charts. I find it interesting that certain lenses can be very sharp at wide apertures.

Yes, and those using Leica lenses have known this for a very long time.

[Ray]

This still gives the bird plenty of time to move its eye out of the plane of focus.

Yes, and those using Leica lenses have known this for a very long time.
[a href=\"index.php?act=findpost&pid=215188\"][{POST_SNAPBACK}][/a]

There's no way anyone could manually focus as quickly as a modern autofocus lens. The only issue is, how accurate is that autofocussing? If it's not accurate, then one is forced to use manual focus.

Leica lenses do have a fine reputation, but there's no Photodo test of a Leica 400mm lens to compare with the Canon 400/2.8.

[Rob C]

Hi,

I think the problem is the menu item "View->Actual Pixels". In film days you needed a microscope to see whether an image was absolutely sharp, in my experience very few were at 40X. I normally used a 15X peak loupe but it was not really good enough for judging sharpness. Now in the digital world we just push Actual->Pixels and can peep an image 50 inches wide at 40 centimeters, with no grain.

Joke aside I actually used a Microscope to look at some slides, very few were really sharp at 40X or 100X.

Erik
[a href=\"index.php?act=findpost&pid=215133\"][{POST_SNAPBACK}][/a]

Yes, probably right, but that doesn´t mean that digitally captured images or those shot with a/f are going to be any better, as Ray has also found failures on many occassions and, I´m sure, so has everyone else willing to think about it objectively, should that ever be a possibility.

[telyt]

There's no way anyone could manually focus as quickly as a modern autofocus lens.

Try the 400m f/6.8 Telyt with the Leicaflex SL and be amazed.  I wish I could put the SL viewfinder in the R8.

Leica lenses do have a fine reputation, but there's no Photodo test of a Leica 400mm lens to compare with the Canon 400/2.8.
[a href=\"index.php?act=findpost&pid=215247\"][{POST_SNAPBACK}][/a]

I've seen test files make with one of the Canon 1D-series cameras comparing the EF 400mm f/2.8 L IS with the old non-modular Leica 400mm f/2.8 APO.  At full aperture the Leica lens has a detail advantage, but a stop or two down they were pretty much equal aside for some differences in color saturation and contrast.  As I figure it the only reason to buy a 400mm f/2.8 is to use it at full aperture; that's how the owner of the lenses figured it too.  He kept both, the Canon for when AF is an advantage, the Leica for when he wants maximum image quality.  I sold my FD 400mm f/2.8 L because the full-aperture image quality was lacking.

[Ray]

Try the 400m f/6.8 Telyt with the Leicaflex SL and be amazed.  I wish I could put the SL viewfinder in the R8.

Well, your manual focussing technique certainly seems to be working well for you. You have some fine images of birds there that all seem to be well-focussed.

I find in general that trying to focus on small birds flitting around in the foliage is very difficult. The last time I tried this was on a recent cruise on the Daintree River, North Queensland. One of the most colorful small birds one can see on the banks of the Daintree is the Azure Kingfisher. It never seems to keep still for more than a second.

The combination of poor lighting and confused foliage surrounding the bird most of the time, made it almost impossible for me to get a well-focussed shot using the Canon 100-400 IS. The autofocus was all over the place. Next time, I'll try manual focussing   .

I've seen test files make with one of the Canon 1D-series cameras comparing the EF 400mm f/2.8 L IS with the old non-modular Leica 400mm f/2.8 APO.  At full aperture the Leica lens has a detail advantage, but a stop or two down they were pretty much equal aside for some differences in color saturation and contrast.  As I figure it the only reason to buy a 400mm f/2.8 is to use it at full aperture; that's how the owner of the lenses figured it too.  He kept both, the Canon for when AF is an advantage, the Leica for when he wants maximum image quality.  I sold my FD 400mm f/2.8 L because the full-aperture image quality was lacking.

I wouldn't be surprised if that's the reason why Photodo never showed any MTF tests for the Canon 400/2.8 IS at full aperture. It's not too hot wide open. At F4 though, the IS version of this lens is at least as sharp as at F8.

However, the 400/2.8 II USM (without IS and now probably discontinued) is sharper at F2.8 than the IS version at F4. Since this lens appears to be optimised for use at full aperture, which is its sharpest aperture, the lack of IS should not be too great a disadvantage. A second hand copy in good condition would be worth getting, but such lenses are too heavy for my purposes.

[woof75]

Yes, such has been suggested, but is a flawed method. If you move the centre of the viewfinder (for example) to focus on something at the far sides, then you must realise that you are focussing beyond the point you really want to have sharp.

It is nothing more than the difference in lengths of the sides of a triangle. The plane of focus of a lens is supposed to be flat: if you focus at the greater distance (as per my example of a subject at the far edges), then, when you turn the camera back to the straight ahead position you wanted, that focussed distance no longer holds and you find you have, in fact, focussed well BEYOND your subject. Unavoidable.

Not rocket science, as you said.

Rob C
[a href=\"index.php?act=findpost&pid=214992\"][{POST_SNAPBACK}][/a]

Yes, in practice it works though. I have worked with a camera designed for manual focus and the centre prism is as it's name suggests, at the center so you still focus then recompose. I often shoot 1000 frames on a fashion shoot and not have one out of focus.

[telyt]

Yes, in practice it works though. I have worked with a camera designed for manual focus and the centre prism is as it's name suggests, at the center so you still focus then recompose.

I don't use a center focussing aid.  I use the "ground glass" (actually matte plastic) area of the viewscreen.  Where practical I have replaced my cameras' viewscreens that have a central focussing aid with a plain matte screen.