I got to try one in my local store (Midwest Photo Exchange) yesterday. Very impressed. I would happily trade my f/4L IS for it, but the cost is prohibitive. The only way I could afford it is to sell my f/4L IS and my 100L macro, which I just can't do....but it's definitely on my radar, and if the price comes down or I miraculously get a lot of money, I'll be getting one.
After using the f/4L IS for so long, it also seems HUGE. I had the 80-200 f/2.8L, which isn't quite as big as the 70-200/2.8 IS, but it never seemed huge, but that was before having the f/4 IS for so long. Still, the image quality is outstanding, even at f/2.8. Really, really want it, but really can't see a way to do it now.
DigLLoyd is full of it. Now I can understand if he reserved judgement and said that he would test another sample, but to try to get attention by claiming it's a design fault.. Well, it would seem that Ken Rockwell has a competitor...
Anyway, test at 58 m (distance measured with a Leica DISTO D5), 200mm, 100% crops:
http://peltarion.eu/img/5D7D/digilloydiswrong.jpg
All shot raw and processed the same way in lightroom. Cropped and labeled in photoshop.
If you want to check the unprocessed RAWs yourself you can download them here (~170 MB zip file).
Tim Ashton wrote:
DigLLoyd's subscription IMO is worth every penny you pay for it.
re the 70-200 it is amazing at 2.8 and better at f/4
no sign of the prob at 18ft
not yet tested at intermediate; 50 ft
but it is not diffraction
He does think it is a design flaw rather than a bad sample which is most disturbing, because in the long run we all need the top end competition to improve the breed and if his observations turn out to be well founded, it will be a disaster following on the probs some encountered with the 1D3.
I can honestly say his is one of the best learning sites I have come across and is well worth the money
Pony up the money, it is not just about this one lens. There is a wealth of carefully crafted information
Disaster? I think that's a bit strong. If there actually is a problem (which I doubt) how would it be any different then other known problems like the 70-200 f4 L IS had at MFD at 200mm? (which my copy blatantly had)
Personally I'd rather get bad information for free at a site like RG's vs paying for it!
DigLLoyd is full of it. Now I can understand if he reserved judgement and said that he would test another sample, but to try to get attention by claiming it's a design fault.. Well, it would seem that Ken Rockwell has a competitor...
Anyway, test at 58 m (distance measured with a Leica DISTO D5), 200mm, 100% crops:
http://peltarion.eu/img/5D7D/digilloydiswrong.jpg
All shot raw and processed the same way in lightroom. Cropped and labeled in photoshop.
If you want to check the unprocessed RAWs yourself you can download them here (~170 MB zip file).
I agree that it looks like he is wrong from the data we have here.
That said, if so and he IS wrong, then it's one mistake. Rockwell is just nuts. Rockwell takes being wrong to a world class art form. Diglloyd will admit he is wrong when confronted with actual data.
Perhaps. I don't actually know too much about him so perhaps I'm jumping to conclusions. From what I read on his blog he did seem quite Rockwellish (although I would not presume to imply that anybody could be more wrong than Rockwell). Still, him claiming a design fault after testing one lens sample doesn't exactly inspire confidence.
I haven't read his exact piece on this, it's under an annual subscription fee and I haven't bought that.
He's been pretty accurate in the past. I'm not exactly sure what are the conditions of his testing. Often he gets into some areas that most photographers don't. For example, he's big into IR modification of cameras to do IR photography. He also is big on Zeiss on Canon/Nikon which is also quite a niche in it's own right. I've also talked with him extensively on his Zeiss stuff - very interesting and knowledgeable was my thought.
Maybe someone here who subscribes could tell us about the test conditions?
Now Rockwell's site - Phew! Another story entirely. The guy specifically says that he writes to be controversial. I think he cynically does it to drive hits to his sight and to his paying links. Either way, sorting out the flat out wrong stuff from what might be true is a waste of time.
He says he has tested a second copy with the same results. I don't know what to make of it. Either his camera body has a problem (but IIRC he tried it on two cameras), there are quality control problems with the MkII and he was unfortunate to get two bad copies (or I fortunate to get a good one) or there are some differences in our tests or he is making things up. I don't know. The samples on the blog are pretty dramatic.
While his test subject is better than mine, such huge differences would have been easily visible on the license plate test as well.
JohnJ80 wrote:
I agree that it looks like he is wrong from the data we have here.
I haven't read Digilloyd's specific comments, but even Denoir's pics show some resolution loss when stopped down. His lens seems to peak at f/5.6, then gets slightly softer at f/8.
That's odd, though it wouldn't bother me at all. I can't imagine ever using a 70-200mm lens past f/5.6. Maybe this was a design decision by Canon, a specific trade-off to get this lens as sharp as it is at f/2.8 - f/5.6
Why Lloyd's 2 copies evidently start to go at f/5.6 is a mystery, but even THAT wouldn't bother me much. This lens could literally stop at f/4, and I'd still want one.
I have now subscribed to Digilloyd's site and I'll partly retract my comments. He seems to do tests fairly properly. Not as good as the digital picture but still pretty much OK. So I'll at least retract the grave insult of associating him with Ken Rockwell.
His samples do show a massive degradation at the same distance/fstop combinations I showed above. There is definitely something strange about such a huge difference in quality.
He does say that he doesn't claim that it happens on all 70-200 MkIIs but just on the two samples that he tested. It could simply be a batch of bad copies that went out. It wouldn't be a first for Canon. As much as I like their cameras and lenses, their quality control is embarrassing.
denoir wrote:
Perhaps. I don't actually know too much about him so perhaps I'm jumping to conclusions. From what I read on his blog he did seem quite Rockwellish (although I would not presume to imply that anybody could be more wrong than Rockwell). Still, him claiming a design fault after testing one lens sample doesn't exactly inspire confidence.
He does post his findings for all to see
Maybe a thinking mans vesion of KR or is that an oxymoron ??
Tim
rd4tile wrote:
Disaster? I think that's a bit strong. If there actually is a problem (which I doubt) how would it be any different then other known problems like the 70-200 f4 L IS had at MFD at 200mm? (which my copy blatantly had)
Personally I'd rather get bad information for free at a site like RG's vs paying for it!
If through QC or design the top end of town lose confidence in either of the majors then the pressure to keep up the pace of development is gone
Maybe disaster is too strong a term but whatever, ultimately the whole photographic community will be the loser
BTW, there is a difference between bad news and bad info.
I have emailed Lloyd my results. It will be interesting to hear his take on it.
Tim: I agree. Either way it sucks. We did see from the tests at the digital picture that there is sample variation. One would expect this type of glass at that price to consistently perform according to factory specifications.
denoir wrote:
I have emailed Lloyd my results. It will be interesting to hear his take on it.
Tim: I agree. Either way it sucks. We did see from the tests at the digital picture that there is sample variation. One would expect this type of glass at that price to consistently perform according to factory specifications.
Please share the outcome of your dialog with Lloyd.
Folks - the sky isn't falling. Between AF system tolerances and sample variation, I'd bet there is not a major disaster brewing and this will all get sorted out shortly.
it is amazing to me how a lens can be owned by probably thousands of people with essentially unanimous praise...people talking bout replacing the best primes such as the 135L because it is so good...people raving about its sharpness and improvements over past lenses and one guy comes out suggesting he sees a problem and all of a sudden everybody is concerned....did all of the lenses of those who spoke glowingly suddenly become lousy?Do they no longer take the incredible pictures all spoke of?I am sure this gentleman will now have huge increases in traffic to his site ..that said,i have tested mine and see no sign of the problems spoken of...perhaps i am wrong,blind or dont know what i am talking about..who knows?somehow i will stick with the glowing reviews of pretty much everyone prior to this gentleman's findings!
JohnJ80 wrote:
Please share the outcome of your dialog with Lloyd.
He was surprised at the results and hopeful that it may in fact just be dud samples that he tested. We talked a bit about testing methodology and I'm now completely convinced that he does know what he is talking about and that his tests were correct executed. We agreed that the whole thing was strange. I explained the method I used for testing, shared my RAW files and he did not find anything wrong with how I measured and confirmed that my samples did not exhibit the same problem.
Anyway, the result was the following now posted on his blog:
Update March 28 (late): one subscriber in Sweden sent me a credible aperture series from his 70-200/2.8L IS II which does not show the issue I’m seeing. Obiously I have no access to his lens, so I can’t test it and confirm that directly. I also have no immediate availability to my usual sources for another sample, but I’ll try to obtain a 3rd sample somehow to confirm (or not) my findings with the first two brand new samples. See also Brand-new Blur — but the blur I’ve seen generally has always improved by stopping down.
jonbrach wrote:
it is amazing to me how a lens can be owned by probably thousands of people with essentially unanimous praise...people talking bout replacing the best primes such as the 135L because it is so good...people raving about its sharpness and improvements over past lenses and one guy comes out suggesting he sees a problem and all of a sudden everybody is concerned....did all of the lenses of those who spoke glowingly suddenly become lousy?Do they no longer take the incredible pictures all spoke of?I am sure this gentleman will now have huge increases in traffic to his site ..that said,i have tested mine and see no sign of the problems spoken of...perhaps i am wrong,blind or dont know what i am talking about..who knows?somehow i will stick with the glowing reviews of pretty much everyone prior to this gentleman's findings! ...Show more →
The thing is that this is a very expensive lens supposed to be at the top of Canon's range. The ISO chart tests over at the digital pictures showed that one out of three samples tested was flawed. Now this example of two samples exhibiting strange behavior. This is important information both for new and existing owners. For new ones that they are to expect sub par quality control and that buying top of the line Canon glass is no guarantee that it will actually work to factory specification. For existing owners it is relevant so that if their lens is affected that they can send it back to Canon for correction or at least be aware of in which conditions the lens should not be stopped down.
One interesting thing, which makes me more convinced that it's a question of bad samples is his comparisons to the 135L. According to his tests the 135L is clearly better at f/2.0 and f/2.8 than the 70-200/II at f/2.8. Only at f/4.0 does the 70-200 start to catch up. My own (much less formal and controlled) tests so far show that the 70-200 @ f/2.8 is sharper than the 135L @ f/2.0 and that the 135L @ f/2.8 only has a very slight advantage over the 70-200.
I very much doubt that you can get a copy that is better than the design/factory specifications which would mean that the copies he tried were problematic.
denoir wrote:
The thing is that this is a very expensive lens supposed to be at the top of Canon's range. The ISO chart tests over at the digital pictures showed that one out of three samples tested was flawed. Now this example of two samples exhibiting strange behavior. This is important information both for new and existing owners. For new ones that they are to expect sub par quality control and that buying top of the line Canon glass is no guarantee that it will actually work to factory specification. For existing owners it is relevant so that if their lens is affected that they can send it back to Canon for correction or at least be aware of in which conditions the lens should not be stopped down.
One interesting thing, which makes me more convinced that it's a question of bad samples is his comparisons to the 135L. According to his tests the 135L is clearly better at f/2.0 and f/2.8 than the 70-200/II at f/2.8. Only at f/4.0 does the 70-200 start to catch up. My own (much less formal and controlled) tests so far show that the 70-200 @ f/2.8 is sharper than the 135L @ f/2.0 and that the 135L @ f/2.8 only has a very slight advantage over the 70-200.
I very much doubt that you can get a copy that is better than the design/factory specifications which would mean that the copies he tried were problematic....Show more →
From what I've seen Sample #2 at the-digital-picture.com is soft at 200mm when wide open. Once stopped down sharpness improves dramatically. At F8, Sample #1 vs Sample #2 are very similar; in fact, center sharpness appears better on Sample #2.
This is not what was reported by Diglloyd. Diglloyd found that sharpness worsens when stopped down to F8.
Please share some samples exhibiting the reported problem if possible. I, and I expect others, would like to see them.
Also, it is very difficult for normal users like us to tell if sharpness is "within Canon specifications." As far as I know the closest thing to a sharpness spec is Canon's online MTF chart but there is no tolerance information there. I assume the chart is meant to be "typical" or "nominal."
Luckily, pixel peeping with a series of controlled shots is usually sufficient for amateurs to determine if a lens is of acceptable sharpness for him/her. Of course, this is not the same as knowing whether the lens is within specification.
