Canon 1DsMk3 meets Leica

[neverfinder]

hello,

Now I finally got my Canon 1DsMk3 and I really like it. Here are a few sample shots from it and a very short test with my old Mark 2.

Everything shot on a Tripod, cable release, ML. Converted in Lightroom, Sharpening is the same for all files. All files are upsampled to 1DsMk3 level.

First image is an overview over the setup.
[attachment=4201:attachment]

The first three crops are from the BLUE area:

- First Leica M8 + Leica 21mm f8
- Second 1DsMk2 + Leica 28mm f8
- Third 1DsMk3 + Leica 28 f8

[attachment=4202:attachment]
[attachment=4203:attachment]
[attachment=4204:attachment]


The second pair of images is a from the Canon vs Leica lens. On the 1DsMk2 the differnece was there but not as huge. I mean it was always a debate between handling and quality.

With the Mark3 the difference got quite big so that I'm really that I own the Leica 28

First two images from the blue area second two from the red area:

- Fist 1DsMk3 + Canon 28mm @ f11 (24-105)
- Second 1DsMk3 + Leica 28 @ f11

[attachment=4205:attachment]
[attachment=4207:attachment]

red area:

- Fist 1DsMk3 + Canon 28mm @ f11 (24-105)
- Second 1DsMk3 + Leica 28 @ f11

[attachment=4206:attachment]
[attachment=4208:attachment]

The last image is just an example. I really love my Leica 19mm and I think the crops show why.

Overview:
[attachment=4209:attachment]
Green:
[attachment=4210:attachment]
Red:
[attachment=4212:attachment]
Blue:
[attachment=4211:attachment]

Some more comments:

I tested also the 70-200 f4 IS against my 70-200 f2.8 IS and my 135L.

I did not post any images because the difference is not that big and in print there is no real difference. I would say 135>70-200 f4 IS>70-200 f2.8 IS but the difference is REALLY small.

Another thing about stopping down, so many people were arguing about how far we can stop down before killing the image quality.

Canon:
More or less I would say everything from f4-11 is absolutly fine. Going up to f16 is still ok but you notice that the image is slowly getting softer. I would not use f22 on the Mark3 but it was quite the same on the Mark2 the image just is getting to soft,

Leica:
From f4-16 is really good and printed it is nearly impossible to see the difference. Also f22 is still quite good and IF I need to I would use it on the Leica glass.

I don't have enough to post examples of that topic but if I find some I will post tje, later.

I hope I coud give some useful information.

[Boris_Epix]

This is amazing. I would have never believed that the difference is that big. At the same time I never considered to get the 24-105  :-)

Thanks for your effort.

[kaelaria]

That IS a huge difference.  Are you sure about the focus of the Canon lens?

[juicy]

Hi!

That IS a huge difference.

Unfortunately this seems to be the case except for the very best of Canon's lenses. It may be unfair to compare AF-zoom to MF-prime because it's probably very difficult to manufacture lens constructions which at the same time contain extremely tight tolerances yet there are easily movable and light weight lens elements for AF. Anyway, that does not change the fact that sample variation in most of today's AF-lenses are huge and if Canon does not do something to sort this problem really soon, many customers will increasingly start looking for other options especially in the wide angles. It's extremely welcome to have Nikon competing in the FF-market with full force and with some excellent new lenses.

You might also check:   the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx
 
Cheers,
J

[Ray]

I can't believe it. All good lenses are pretty similar at f11. The blurring of the 1Ds3 shot with the Canon lens looks like a case of mirror slap to me.

Whenever you get suspiciously incredible results like that, you should retake the shots and examine your methodology.

[juicy]

I can't believe it. All good lenses are pretty similar at f11. The blurring of the 1Ds3 shot with the Canon lens looks like a case of mirror slap to me.

Whenever you get suspiciously incredible results like that, you should retake the shots and examine your methodology.
[a href=\"index.php?act=findpost&pid=160798\"][{POST_SNAPBACK}][/a]

Hi!
What is it that you can't believe? That most of Canon's WA-lenses are crap in the corners is no news to anyone. Check out 16-9.net just for one good example site with some eye-opening    comparisons. The most recent is 5D with Nikon 14-24 vs 24L. The results are very much in line with other people's findings.

Cheers,
J

[Mort54]

You're comparing a zoom (Canon 24-105) against a prime (Leica 28). Also, the 24-105 isn't considered Canon's sharpest by any means. So I guess your results aren't that surprising.

[Boris_Epix]

Mort: You can compare a Porsche to a Ferrari just as you can compare a Ford to a Ferrari.

I personally didn't expect such a HUGE difference because so many people talk about the 24-105 as if it was a godsend. I continue to be happy AND unhappy with the Canon 24-70mm 2.8L which can be very sharp one day in one setting and pretty unsatisfactory in another light. I used to love the 70-200 2.8L IS although it wasn't that sharp. But it was convinient. It was so useful before it suddenly started to backfocus on my 1Ds Mk2. I focussed at the front eye and the ears were sharp.

I understand that Canon glass (expensive L glass I'm talking about) can't be as good as some of the new digital MF glass. But the huge tolerance with Canon lenses is really annoying. They need to work on their quality (assurance) but probably a certain deviation is considered normal and keeps the pricepoint of the lenses "lower".

Cheers
Boris

You're comparing a zoom (Canon 24-105) against a prime (Leica 28). Also, the 24-105 isn't considered Canon's sharpest by any means. So I guess your results aren't that surprising.
[a href=\"index.php?act=findpost&pid=160853\"][{POST_SNAPBACK}][/a]

[Mort54]

Mort: You can compare a Porsche to a Ferrari just as you can compare a Ford to a Ferrari.

Most comments I've heard from owners of this lens are that it's soft. That said, I certainly wish Nikon had something like that, since it's a very versatile focal length range, and has IS (the Nikon 24-125 VR has an even worse repulation for softness, unfortunately, or else it would be a perfect walk around lens).

[CatOne]

I can't believe it. All good lenses are pretty similar at f11. The blurring of the 1Ds3 shot with the Canon lens looks like a case of mirror slap to me.

Whenever you get suspiciously incredible results like that, you should retake the shots and examine your methodology.
[a href=\"index.php?act=findpost&pid=160798\"][{POST_SNAPBACK}][/a]

There's also some pretty large CA in the shot with the colorful books.

The difference is marked.  I don't know that the 24-105 is one of Canon's sharpest lens by any means (also maybe not so fair to test a zoom against a prime), but those differences ARE huge.

[Mark D Segal]

My 24~105 is very sharp - I have confidence it would deliver considerably better results than those shown here. There is quality variance from one piece to the next over a range between their minimum acceptable standard and the best few percent they produce. With zooms I am given to understand the issue is generally the accuracy of the alignment. Zooms are much more complex lenses than primes, with much glass for the light to travel through, therefore comparing a zoom with a prime is rather pointless. Apart from this, very interesting set of comparisons.

[Ray]

Hi!
What is it that you can't believe? That most of Canon's WA-lenses are crap in the corners is no news to anyone. Check out 16-9.net just for one good example site with some eye-opening    comparisons. The most recent is 5D with Nikon 14-24 vs 24L. The results are very much in line with other people's findings.

Cheers,
J
[a href=\"index.php?act=findpost&pid=160828\"][{POST_SNAPBACK}][/a]

This is what I can't believe.

[attachment=4277:attachment]


I can't find these books on the overview. It's not clear that the crop is from an extreme corner and it is surprising that this degree of difference should exist even at f11.

But again, how small a crop is that and how far into the corner?

[Rob C]

This is what I can't believe.

[attachment=4277:attachment]
I can't find these books on the overview. It's not clear that the crop is from an extreme corner and it is surprising that this degree of difference should exist even at f11.

But again, how small a crop is that and how far into the corner?
[a href=\"index.php?act=findpost&pid=160987\"][{POST_SNAPBACK}][/a]

So, CSI reaches the Far East too. Mamma always told me not to believe everything I read or see...

Ciao - Rob C

[juicy]

This is what I can't believe.

[attachment=4277:attachment]
I can't find these books on the overview. It's not clear that the crop is from an extreme corner and it is surprising that this degree of difference should exist even at f11.

But again, how small a crop is that and how far into the corner?
[a href=\"index.php?act=findpost&pid=160987\"][{POST_SNAPBACK}][/a]

Hi!

This is the crop marked in red in the upper edge, not a corner.
I know nothing about the methodology of aligning the camera used in this test (whether it was done carefully with a spirit-level and if MLU was used etc) but my point was simply that I find it rather depressing to be forced to test several lenses every time when bying a lens. I mean it should be the  manufacturer's task to test the lens before attaching the price tag! It won't help Canon in the long run if customers start to associate the L with Lemon instead of Luxury.

Cheers,
J

[Ray]

Hi!

This is the crop marked in red in the upper edge, not a corner.
I know nothing about the methodology of aligning the camera used in this test (whether it was done carefully with a spirit-level and if MLU was used etc) but my point was simply that I find it rather depressing to be forced to test several lenses every time when bying a lens. I mean it should be the  manufacturer's task to test the lens before attaching the price tag! It won't help Canon in the long run if customers start to associate the L with Lemon instead of Luxury.

Cheers,
J
[a href=\"index.php?act=findpost&pid=160996\"][{POST_SNAPBACK}][/a]

No it's not. The crop I'm referring to and which you have repeated in the above post does not appear anywhere in the overview. The area highlighted with a red square is different.

I agree it's a PITA to have to test lenses before you buy. Furthermore, once you start doing that and you discover how much variability there can be, you become fearful of buying another lens ever without testing it first.

The last lens I bought without testing was a Canon 50/1.4, in Bangkok. When I got back to Australia and one day compared it with my Canon 50/1.8, a much cheaper and lighter lens, I found it wasn't as sharp. It's supposed to be marginally sharper.

I've been advocating for years a policy where ever lens is sold with an individual, real MTF chart (as opposed to a theoretical one) as a part of standardised quality control.

But no-one listens to me   .

[mcbroomf]

It's definitely in the overview.  Top edge in center.  
The 1st overview image highlights it correctly.  In the 2nd overview the red square is moved off to the left.

[Ray]

It's definitely in the overview.  Top edge in center. 
The 1st overview image highlights it correctly.  In the 2nd overview the red square is moved off to the left.
[a href=\"index.php?act=findpost&pid=161048\"][{POST_SNAPBACK}][/a]

Okay! My apologies! I was concentrating more in the corners and missed it. Considering this comparison is far from the corner, it's a disgraceful result for Canon.

Extreme corners are the furthest from the centre, about 22mm maximum. The short edge of the frame, in the middle, is 18mm maximum. If this obvious lower resolution of the Canon lens at 28mm and f11 had been in the extreme corner, I would have been more forgiving.

[neverfinder]

So thanks for so many coments. I was gone for some days doing some shooting with my new baby and didn't have any access.

So here are some answeres.

-MLU was used on both cameras all the time. Both cameras were set on a tripod and were switched with L-brackets.

- The second pair of images is still from the first overview so sorry for that confusion.

- I know the leica is a prime and the Canon is a zoom, but this is the stuff I have and use. A few things you have to know. The 24-105 is/was a quite good lens on my 1DsMk2. I also have to admit after seeing the results I consider getting a 35 prime. Canon's tele zooms are really great, they just can't make anything wide ... don't ask me why ^^

- My 24-105 is certainly not the best out there, but the bnest I could find out of 5 tested lenses, so I think is quite ok. ON the other hand you really have to know that the Leica 28 IS really the best 28 lens out there. I have never seen anything close. I think it is one of the lenses which really gets very close to leicas M-lens line.

- The whole test is done for my own sake. I just wanted to share it with you. It is not made to show anything specific just the things I saw and the things I really was surprised.

[Christopher]

Just to make one more thing clear, which totally sliped my mind. ( Thanks to the person pointing it out )

I was asked whether I had a posting history here or just new to the forum. I'm the same person as "neverfinder". I just did all the testing on my old workstation back home and there I was still logged in with my old user name. So I uploaded these tests with the wrong login name.

Christopher

[Mark D Segal]

Hi!

This is the crop marked in red in the upper edge, not a corner.
I know nothing about the methodology of aligning the camera used in this test (whether it was done carefully with a spirit-level and if MLU was used etc) but my point was simply that I find it rather depressing to be forced to test several lenses every time when bying a lens. I mean it should be the  manufacturer's task to test the lens before attaching the price tag! It won't help Canon in the long run if customers start to associate the L with Lemon instead of Luxury.

Cheers,
J
[a href=\"index.php?act=findpost&pid=160996\"][{POST_SNAPBACK}][/a]

No. It works like this. They have a minimum standard for the lens. Nothing with an L on it should leave the factory that doesn't meet the minimum standard for that lens. Above that minimum standard quality improves varying by the piece. But as long as the lens meets the minimum standard for an L lens, Canon is satisfied (doesn't mean the customer is) that the lens is fine. Now, if they were to raise the standard to only market quality at the top 10 percentile of what they now market, the price would be multiples of the price we now pay. So you can have absolutely the finest they can produce at very high prices, or you can have anything above their minimum standard for current prices. They've made that commercial decision.