Canon TS-E 24 II v 17mm

thanks...
plenty sharp ...
interesting settings...

appreciate you posting your samples...
you seem to be making use of the features...

I am leaning towards the 17 right now..

if they have a new 90mm version that would be cool too..
current one is supposed to be pretty good already

after I try the 17
I may get more into it all..lots of learning and discovering new apps for it

thanks
TOM

steve...what did you decide..
which lens?

or are you in the middle of rental

TOM

Tom

The good news is that there isn't a bad choice here: both lenses are exceptional and represent an entirely new level of optical performance for Canon manufactured wide angle. I bought the 17 first because I didn't have anything that wide that was any good, but found myself loving it so much that I got the 24 too. I was torn for a moment between the TSE 24 and the 24 f1.4 II, but I've already got the first generation 24 f1.4 which is pretty good and so decided to go with the extra flexibility of the TSE.

Once you've shot with a TSE lens for landscapes and architecture it's really hard to believe you survived so long without one. It's an entirely new approach and very reminiscent of large format days.

Kevin

Have not had chance to rent these babies yet, but have a few days off coming off soon. Thanks for the info everyone, I am shooting FF and as mentioned there is no loser due to the high quality. I would love to be able to have both but cost is a factor in that department. Thanks for posts and examples

Chez Wimpy said:

...plus the ability to match the 17mm FOV with a three shot stitch on the 5D2, yet with 40MP, is not something to overlook.

Click to expand...

Has anybody directly compared edge/corner resolution in a 5D2 shot from the 17mm TSE with a three shot shift-stitch from the 24mm TSE (downsampled to 21 MP)?

According to the testresults here on dpreview resolution of the 24TSE drops quite a bit in the extended image circle, not to mention vignetting, even stopped down. I would expect this to be a problem in the three shot stitch.

I have heard a few people say the 24 suffers a bit fully shifted in the corners - but even then I expect it's relative to its excellence overall. The 17TSE doesn't seem to degrade too much though, when fully shifted, so I'd be very surprised if the 24 was worse. Corner performance of the 17 is fabulous also btw....
--
Lizzie
----------------------------------
http://www.lizzieshepherd.com

I just rented the 24 and it is a superb lens. 24mm on a FF body is very wide for landscapes. I always seemed to have stuff in the foreground I did not want. For architecture the 17 would work well.

--
The solution is always simple. Getting there is the hard part.

mattr said:

Has anybody directly compared edge/corner resolution in a 5D2 shot from the 17mm TSE with a three shot shift-stitch from the 24mm TSE (downsampled to 21 MP)?

Click to expand...

WRT corner performance, it is a good idea for a comparison... but so few have both lenses that it would be hard to get a proper test done. For sure the 17 probably has the sharpness / light falloff advantage, but the 24 gets much greater resolution out of the center of any stitched image.

I did a test a week or so ago to see just how bad the falloff might be. I mostly use the 24 on the 550D because I don't need the wide FOV (why I didn't get the 17 to begin with!) but I will need the pixels for poster-sized prints (the 550D gets about 50MP over the same FOV as the 21MP 5D2... so when I don't need movements, I do it in three shots). f11 here, and vignetting was not so obvious:



slight downward tilt to get the rocks / grasses in focus.

Here is the bottom right corner at 100% with 300 0.3 USM:



bottom edge (center):



upper center of image:



upper right of image:



There is a slight curve to the horizon since I couldn't shift for perspective, and the lens shows some barrel distortion in its "native" MF projection. Another thing to watch out for (besides 5D2 banding!) is that DOF is placed precisely and that field curvature hasn't hurt the sharpness. There is a strange effect near the edges of images with strong tilt wherein objects (logically within DOF) become mush. I notice this more with the 90 than the 24, but I suspect we are seeing it here where the upper right side of the image should be (by my estimation) a bit sharper to match the lower right edge's (not pictured, but about the point where the "moss" kicks in) extreme sharpness.

mattr said:

I would expect this to be a problem in the three shot stitch.

Click to expand...

Haven't seen it as a problem so far - this is not obvious at all in A4 sizes (the biggest I can print locally... for keeper images I am sending away for A1, so I will comment more in the future)

Stitched with the 550D, the 24 looks very good out to the corners in almost every respect. Here there is no sharpness falloff at all.



I also posted some impressions here:
http://forums.dpreview.com/forums/read.asp?forum=1029&message=35923039

--
-CW

よしよし、今日も生きのいい魂が手に入ったな

From memory this one was not shifted at all (nor tilted I think) and so is perhaps not a fair comparison but gives some kind of reference point I guess.
Whole image:



Bottom right corner at 100% and sharpened - possibly a little too much but I can't upload anything to my personal webspace at present - but you can see how detailed it is. The far distant detail is also really good but again I can't upload an example at present:



--
Lizzie
----------------------------------
http://www.lizzieshepherd.com

The issue that I think you run into with stitches and depth of field is due to these being curved field lenses. The point of focus, a precise distance upon which the lens is focused, describes an arc with the optical center of the lens as it's center point. When you look at a distant scene the objects to the left and right are farther from you than the objects directly ahead of you, even though they lie on a plane with each other - and thus they are further "behind" the arc of the depth of field that surrounds the arc of focus.

I suspect that when shifting extensively this becomes a greater factor, as the objects now projected onto the sensor are even further away, in a straight line, than those directly in front of you when you focused.

Not sure I'm explaining it very clearly, but I have a feeling that this is an issue.

Kevin

Thanks for the crops and comments.

Light falloff at f/11 certainly doesn't seem to be a serious issue. A bit more troubling is the slight mushiness at the upper right border at infinity. After all, this is not really at the edge of the extended image circle.

I'm considering to get the 24TSE for my 5D2 in my longstanding quest to find an excellent 24mm lens for landscape work. Currently I'm using the 24/2.8. My copy is actually pretty good on full frame, much better than some tests would make you believe. However, for critical applications CA needs to be corrected in PP, and there is certainly some mush in the corners. Even in the center, microcontrast is not quite on the level of the 35/2 or especially the 50/1.8. Sometimes when I need 24mm FOV I'm using the 35/2 in portrait for a conventional lens-rotating three stitch panorama.

I don't really need tilt/shift (however, manual focus is no problem at all because I use LiveView almost all the time for landscape work anyway and I hate the sometimes crappy manual focusing of AF lenses). It would be easier for me to justify the 24TSE if it could also be used for wide interiors using the three image shift-stitch to produce a 17mm equivalent. Your crops indicate that results will be very good with some room for improvement at the edges.

Chez Wimpy said:

Stitched with the 550D, the 24 looks very good out to the corners in almost every respect. Here there is no sharpness falloff at all.

Click to expand...

Yes, it makes sense to sample the non-extended 35mm image circle with a higher pixel density sensor when the lens is as sharp as the 24TSE. For me however 21MP is most of the time enough. And if that's not enough (for really big prints), I might want even more than a three image stitch and conventional panorama setups may be the better solution (no issues with softness at the borders of the extended image circle, no MP limit)

I think what you describe is residual field curvature. In most lens designs this is viewed as a flaw. I would expect Canon's TSE lenses to have at least as little field curvature as other lenses across the normal 35mm image circle but the issue can of course get more prononunced in the extended image circle.

The corner crop is indeed amazingly sharp. What was the f-stop?

It seems if you want across the frame resolution at 17mm, the 17mm TSE may indeed be better than a shift-stitch with the 24mm TSE.

It was f10 though I'm sure I could have got away with a wider aperture, certainly if I'd used a tiny bit of tilt - which I don't think I did....

The 17 certainly is a phenomenal lens and whilst I know the 24 is meant to be that little bit sharper, I've not seen anything yet to suggest there's much in it at all.

One of the reasons I went for the 17 is I wasn't at all happy with certain aspects of Canon's UWA zooms' performance, whereas I'm pretty happy with my 24-70 at 24, even if it's not up there with the TSE.
--
Lizzie
----------------------------------
http://www.lizzieshepherd.com

mattr said:

And if that's not enough (for really big prints), I might want even more than a three image stitch and conventional panorama setups may be the better solution (no issues with softness at the borders of the extended image circle, no MP limit)

Click to expand...

Ah, but the reason to stick with the TSE lenses is that (for my work) the tilt makes the "infinite DOF" possible. The parasol flowers there with the focus wedge extending out to the distant mountain, without bumping into diffraction (keeping the camera at f7.1), also means I can keep the ISO down to avoid motion blur from the wind... and I can use a PL. At this point I can't even imagine going back to my pano rig.

--
-CW

よしよし、今日も生きのいい魂が手に入ったな

rlshep said:

The 17 certainly is a phenomenal lens and whilst I know the 24 is meant to be that little bit sharper, I've not seen anything yet to suggest there's much in it at all.

Click to expand...

At f10 or so, I doubt there is any difference. Where the 24 really shines is being tack sharp corner to corner on FF from f3.5 onwards. Since this is tripod work, I doubt most would care, but it does come in handy keeping ISO down shooting flower detail on windy mountaintops and in doing "moonscapes" at night.

--
-CW

よしよし、今日も生きのいい魂が手に入ったな

Good point - and I certainly can't do 3.5 on my 17TSE! ;-)

Seriously though, I've not really tested it at f4 because I haven't yet had the need - only taken a few very high ISO snapshots on holiday inside a dark church. Corners are certainly decent and not significantly degraded but I can't make any serious judgement as to how much loss of detail there is because these were just rushed shots. From the reviews, I believe there is some loss of quality though - even if it does wipe the floor with any competitors - 24mm is always more likely to be kinder on the corners wide open I'd think, but for an UWA the 17 does very well. Would be nice to have both!
--
Lizzie
----------------------------------
http://www.lizzieshepherd.com

Chez Wimpy said:

Ah, but the reason to stick with the TSE lenses is that (for my work) the tilt makes the "infinite DOF" possible.

Click to expand...

Yes, given the right scene, a tilt lens is absolutely fantastic for various reasons, especially when we want to make full use of sensors with higher pixel density (less stopping down required and less diffraction).

The problem is, when I look at the scenes I tend to photograph at 24mm, there are an awful lot where tilt would not have helped much (e.g. tall trees in the frame or close subjects that are not at an edge).

It really is a bit frustrating: when I want something better than my old 24/2.8 for stopped down landscape photography, Canon forces me to buy a TSE that I don't really want or a crazy fast 24/1.4 that I most definitely don't need.

rlshep said:

One of the reasons I went for the 17 is I wasn't at all happy with certain aspects of Canon's UWA zooms' performance, whereas I'm pretty happy with my 24-70 at 24, even if it's not up there with the TSE.

Click to expand...

I'm starting to think that the 17TSE would for me perhaps also make more sense... especially when Canon updates the 24-70L with performance at 24mm that beats my 24/2.8.

I could use the 17TSE with my 500D, or with an upcoming 60D with hopefully a 21MP APS-C sensor This would be a very nice wideangle rig (around 20-28mm) when some shift-stitching is done.

And with the 5D2 I would have really extreme wideangle capabilities that will never be matched by any zoom.