24-70L Redux
Le sigh.
Despite an excellent batch of preliminary test shots at home when I got it 10 days ago, in the field this week it did not perform up to expectations. Don't get me wrong, I'm in LOVE with the AF speed, the focal length (50-70mm really is the sweet spot for most of my shooting), the bokeh, the "magic dust" it puts into contrast and colour... Despite the months I've waffled on, hesitated and hemmed-and-hawed, using it shows me exactly why it's the choice of so many, and it really is pretty much the perfect lens for me.
BUT.
I fear that while it is pretty darned good at 70mm even wide open, from 35mm and wider it wasn't much good until it was stopped down to about f4, and even some of those were iffy at 24mm if there wasn't enough obvious contrast for it to find. This was significantly worse the greater the distance to the subject.
Ever the optimist (and hoping it was simply user-error somewhere), I put it through some dollar bill tests and did some MFAF as a reasonable starting point. On a tripod, good light - augmented with flash so I could keep the shutter speed nice and high - approximately double the MFD (which I believe is Ziggy's formula for testing focus). Jpgs set to neutral, so no sharpening or other adjustments (jpg rather than raw simply because of speed of uploading and space!)
Ruh-roh ~best Scooby Doo voice~
Here are the results from testing today. I settled on -4 as the best MF adjustment; it might go to -3, but that's about it. These are ~100% crops (give or take). Shot on a 7d. It may have been about 4ft away in fact (I didn't actually measure it out), even though I noted 3ft on the images below.
Obviously at 70 5.6 it's really good, and nearly as sharp as a prime. but I NEED to shoot wider than that a lot of the time (both in FL and aperture). I did get some good 70mm 2.8 shots out of it earlier in the week, but I guess those were lucky, since the above shows that it's really not as sharp as it needs to be.
So, what now?
I reckon I have three choices:
1. Send it back to KEH and try another one. This will cost me at least $200-300 more since most of their other stock on this lens is in better external condition and priced accordingly. The other problem here is that, given the reputation for inconsistency with this lens, there's no guarantee another one will be any improvement (and may present with a different set of "copy variance" issues). If this is the recommended option, I will call them and talk to them and see what they suggest too, of course.
2. Keep this one, and send it to Canon for calibration. What does that cost, and is it likely to *really* fix the issues? I gather that the sharp-70-iffy-24 is not uncommon with this lens, although what I'm seeing above seems pretty extreme to me (other owners, by all means please chime in with your experiences at 24-35 2.8-4.0). For the record, it's a UU date code.
3. Send it back, get a refund, and wait it out until a v.II model comes along. HOwever, given recent events in Japan, I'm not counting on this being soon, even if it was in the pipeline before the earthquake (and please don't get me wrong - the human tragedy there is of far greater signfiicance than my paltry lens-purchasing needs, but I'm just trying to keep realistic facts in the picture as I deal with this specific issue in my little world. There WILL be an impact on consumer products, and it would be naive to think otherwise, so I may as well include that as a factor in how I proceed). It probabaly will require a mortgage to buy it, too. (Obviously, this is my least-preferred option!)
I'm so sad - I got used to the weight remarkably quickly, and everything else about this lens lives up to *exactly* what I need and want - I'm basically already in love with what it CAN produce (I've got some gorgeous shots out of it, despite this). But I DO need it to do so consistently, and at all focal lengths, reliably and without my having to compensate every time I click the shutter... which clearly it isn't doing at this point. :cry
Thoughts?
Despite an excellent batch of preliminary test shots at home when I got it 10 days ago, in the field this week it did not perform up to expectations. Don't get me wrong, I'm in LOVE with the AF speed, the focal length (50-70mm really is the sweet spot for most of my shooting), the bokeh, the "magic dust" it puts into contrast and colour... Despite the months I've waffled on, hesitated and hemmed-and-hawed, using it shows me exactly why it's the choice of so many, and it really is pretty much the perfect lens for me.
BUT.
I fear that while it is pretty darned good at 70mm even wide open, from 35mm and wider it wasn't much good until it was stopped down to about f4, and even some of those were iffy at 24mm if there wasn't enough obvious contrast for it to find. This was significantly worse the greater the distance to the subject.
Ever the optimist (and hoping it was simply user-error somewhere), I put it through some dollar bill tests and did some MFAF as a reasonable starting point. On a tripod, good light - augmented with flash so I could keep the shutter speed nice and high - approximately double the MFD (which I believe is Ziggy's formula for testing focus). Jpgs set to neutral, so no sharpening or other adjustments (jpg rather than raw simply because of speed of uploading and space!)
Ruh-roh ~best Scooby Doo voice~
Here are the results from testing today. I settled on -4 as the best MF adjustment; it might go to -3, but that's about it. These are ~100% crops (give or take). Shot on a 7d. It may have been about 4ft away in fact (I didn't actually measure it out), even though I noted 3ft on the images below.
Obviously at 70 5.6 it's really good, and nearly as sharp as a prime. but I NEED to shoot wider than that a lot of the time (both in FL and aperture). I did get some good 70mm 2.8 shots out of it earlier in the week, but I guess those were lucky, since the above shows that it's really not as sharp as it needs to be.
So, what now?
I reckon I have three choices:
1. Send it back to KEH and try another one. This will cost me at least $200-300 more since most of their other stock on this lens is in better external condition and priced accordingly. The other problem here is that, given the reputation for inconsistency with this lens, there's no guarantee another one will be any improvement (and may present with a different set of "copy variance" issues). If this is the recommended option, I will call them and talk to them and see what they suggest too, of course.
2. Keep this one, and send it to Canon for calibration. What does that cost, and is it likely to *really* fix the issues? I gather that the sharp-70-iffy-24 is not uncommon with this lens, although what I'm seeing above seems pretty extreme to me (other owners, by all means please chime in with your experiences at 24-35 2.8-4.0). For the record, it's a UU date code.
3. Send it back, get a refund, and wait it out until a v.II model comes along. HOwever, given recent events in Japan, I'm not counting on this being soon, even if it was in the pipeline before the earthquake (and please don't get me wrong - the human tragedy there is of far greater signfiicance than my paltry lens-purchasing needs, but I'm just trying to keep realistic facts in the picture as I deal with this specific issue in my little world. There WILL be an impact on consumer products, and it would be naive to think otherwise, so I may as well include that as a factor in how I proceed). It probabaly will require a mortgage to buy it, too. (Obviously, this is my least-preferred option!)
I'm so sad - I got used to the weight remarkably quickly, and everything else about this lens lives up to *exactly* what I need and want - I'm basically already in love with what it CAN produce (I've got some gorgeous shots out of it, despite this). But I DO need it to do so consistently, and at all focal lengths, reliably and without my having to compensate every time I click the shutter... which clearly it isn't doing at this point. :cry
Thoughts?
facebook | photo site |
0
Comments
What do you mean by MF -4? This was manual focus using 10x in live view.
<TABLE cellSpacing=0 cellPadding=2 width="100%"><TBODY><TR><TD vAlign=top width=110 align=right>Date Taken:</TD><TD vAlign=top>2011-03-27 12:22:22</TD></TR><TR><TD vAlign=top width=110 align=right>Camera:</TD><TD vAlign=top>Canon EOS 7D</TD></TR><TR><TD vAlign=top width=110 align=right>Exposure Time:</TD><TD vAlign=top>0.01s (1/100)</TD></TR><TR><TD vAlign=top width=110 align=right>Aperture:</TD><TD vAlign=top>f/7.1</TD></TR><TR><TD vAlign=top width=110 align=right>ISO:</TD><TD vAlign=top>100</TD></TR><TR><TD vAlign=top width=110 align=right>Focal Length:</TD><TD vAlign=top>24mm (38.4mm in 35mm)</TD></TR></TBODY></TABLE>
<TABLE cellSpacing=0 cellPadding=2 width="100%"><TBODY><TR><TD vAlign=top width=115 align=right>Date Modified:</TD><TD vAlign=top>2011-03-27 13:38:23</TD></TR><TR><TD vAlign=top width=115 align=right>Photo Dimensions:</TD><TD vAlign=top>1000 x 1500</TD></TR><TR><TD vAlign=top width=115 align=right>File Size:</TD><TD vAlign=top>332.97 KB</TD></TR><TR><TD vAlign=top width=115 align=right>Flash:</TD><TD vAlign=top>flash fired, compulsory flash mode</TD></TR><TR><TD vAlign=top width=115 align=right>Metering:</TD><TD vAlign=top>pattern</TD></TR><TR><TD vAlign=top width=115 align=right>Exposure Program:</TD><TD vAlign=top>manual</TD></TR><TR><TD vAlign=top width=115 align=right>Exposure Bias:</TD><TD vAlign=top>0 EV</TD></TR><TR><TD vAlign=top width=115 align=right>Exposure Mode:</TD><TD vAlign=top>manual</TD></TR><TR><TD vAlign=top width=115 align=right>White Balance:</TD><TD vAlign=top>manual</TD></TR><TR><TD vAlign=top width=115 align=right>Subject Distance:</TD><TD vAlign=top>89/100</TD></TR></TBODY></TABLE>
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
If the lens/camera combination tests well for the brick wall test then you can reasonably say that the lens is working basically well with the camera. Then you progress on to tests of the AF with foreground and background elements, trying to see how focus is affected by those distractions. The more discriminant the AF, the better you should be able to focus within "clutter" in the scene. You should find some variation between different lens/camera combinations, and that is normal. A fence line viewed from an oblique angle works well for this test but start with a high-contrast subject on top of a fence post in the center. If the camera/lens passes this test then remove the high-contrast subject and try focus on a single fence post in the line. The other posts and the wire of the fence are the distractive elements.
If that passes then move on to shooting through some branches at some more distant subject that also has stuff behind. Gradually you will see just how discriminating the AF system can be.
Keep working tests like these until you understand how "that" lens is working with you camera's AF system at typical distances. You can mix it up later to determine what your extremes will be, both distant and near. Remember too that minimum focus distance (MFD) at different focal lengths of the zoom range will be different. I believe that the MFD of the EF 24-70mm, f2.8L USM is rated at the long end of the range.
The Canon 7D can show magnifications that are rather extreme and potentially unfair to many otherwise useful lenses. In other words, only "very" large prints may ultimately show the problems you easily see on the screen.
Finally, the Canon 24-70mm2.8L is known to have sample variations (as you previously noted). You may just have gotten one of lesser quality. It is a very old line that is in need of a newer model and newer manufacturing and assembly techniques and tolerances than current methods allow. Another copy of the lens must be a consideration if this copy does not meet your needs.
Moderator of the Cameras and Accessories forums
MF = Microfus adjustment, ie I tweaked it down to -4.
Dan, do you have those unprocessed? I think it will be easier to compare like-with-like than try to guesstimate post processing. My shots were unmanipulated so that processing was out of the equation.
Ziggy, thanks for the further info. I remember somewhere seeing that 2x the MFD was the recommended distance for testing, hence why I chose that - clearly I remembered wrong (could it have been 10x?
I'll post back with more info when I've shot it.
Here are the links for those raws.
http://examples.danal.com.s3.amazonaws.com/IMG_0993.CR2?AWSAccessKeyId=1Z5VDRERTPVZGZKS0602&Expires=1306529984&Signature=1W4nK%2fc2vyrTWhdTmxMdWbKmZdk%3d
http://examples.danal.com.s3.amazonaws.com/IMG_0994.CR2?AWSAccessKeyId=1Z5VDRERTPVZGZKS0602&Expires=1306530004&Signature=sahbQJ2GwuNkIHSCa%2bisgZGYxTw%3d
After you get a chance to compare them to yours please let us know the results.
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
I think the kind of definitive answer you are looking for can only be got with a calibrating tool and watertight technique. Another Dgrinner is lending round his LensAlign (details needed), you could get in the queue. Or, do you know someone with the 24-70 who'd let you try it on your cam to test along with your copy (local camera club?)?
It's not for no reason that this lens has been the subject of a long discussion about IQ and calibration. But having to stop down for sharpness with some settings is a common characteristic of many lenses, as you know.
I don't remember using my copy at 3ft and closer, but all the images I got from it were a tiny bit on the soft side.
Neil
http://www.behance.net/brosepix
Still experimenting here, but have also sent KEH an email to find out what my options are. Idiot that I am I didn't notice that their 2-week return is from *invoice* instead of delivery, but since there's still a 60 day warranty (and I'm only out by a day and it's a Sunday when they're closed) maybe we can work something out.
Also still trying to find out ~cost of a calibration if I send it back to Canon. In many ways, that might be my best option if it's likely to truly fix the issues I'm having with it, as I think might still save me money over taking another copy. I really like this darned thing now I've got it - I just want it to work right for me! It's like the little girl with the curl, because when it's good it's very, very good, but.... Anyway, more as it develops.
(PS is NJ the only Canon service center? If there's one in the DC/Balt area I'd consider just taking it in person....)
As I say above, still researching both the performance, and my options. The frustration at the moment is the variability in "real life" situations - some shots are great, and others are el stinko, and I can't take all the credit for the rubbish being my fault on this occasion
ETA: I did NOT manual focus - all AF.
how did you decide on -4? yes..I am questioning your MF adjustment! Can we see equivelnt shots for -2, -3, -5 and -6? The reason I ask is if DOF focus..if you calibrated at apertures that are smaller, then when you open it up..you may be at edge of DOF for more open apertures. Did you MF at 2.8?
also as ziggy mentioned try objects further away. Minifmum focus distance may not be the best.
14-24 24-70 70-200mm (vr2)
85 and 50 1.4
45 PC and sb910 x2
http://www.danielkimphotography.com
For comparison, here's 24mm at -20, so you can see there is a significant difference between that and where I wound up...
You remembered correctly and it does look like that rule-of-thumb worked for the 70mm end of the zoom. Like I said before, even my 135f2L does poorly at close distances and I almost freaked when I saw the first tests at close proximity. At my normal shooting distances it is as close to perfect as I can expect.
Moderator of the Cameras and Accessories forums
I didn't do mirror lockup, but I was in liveview when I too the shot so that is the same has using mirror lockup, right?
And I did use a tripod.
BTW the DOF at 24mm f2.8 is relatively big, but I was careful to tape (and stretch) all sides.
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
Like Ziggy53 suggested, try it at the usual use distance.
But I admit, I think you got a piece that you may not be happy with no matta what! My finest lenses look fine close and very far too. my junk: I sold!
Um, there's one in California.
I thought I heard about a new one in Virginia, but I can't find Canon's list. I think it may be on the back of a warranty card or something like that.
Right.
Did you do the MA tuning after the first round of test shots? Did you see any hint before doing MAs of what you're seeing now?
From your shots immediately after you got the lens, my impression was that the lens had been recently calibrated. Had there been something wrong enough with the calibration when you first got it in your hands, I think it might have been obvious then. It didn't seem to be. I'm wondering if the new settings from the MAs have upset the calibration it had when you received it. I don't know enough about calibrating lenses to know if such an idea makes sense.
I think I would be talking to KEH asap. Perhaps you could explain that you would like to send the lens to Canon, if they would extend the return period until after you get it back. I am as convinced as you that it either needs checking or retuning. Let KEH help you make that decision, perhaps.
I'd also retest and use a more stable target, and one where you can judge back or forward focusing, because those problems, and shallow DOF, obviously can give a result like what you are seeing on a flat target, and a flat target doesn't explain much. Make sure too that the histogram is bunched right up the right end. Any degree of underexposure seems to accentuate a bit of softness.
Neil
http://www.behance.net/brosepix
And if it turned out that this lens really wasn't very good at taking pictures of five dollar bills at 3 feet using 24mm f1:2.8, I would probably avoid taking those, and look to see if it did some other things well. I don't own this lens, but I've seen some stunning images from it (well not this particular copy, but ones like it).
Another obvious choice for you, since you have a crop camera, is the efs 17-55mm f1:2.8. Maybe you should rent one of those and compare. It could suit your style better, who knows?
I have a 17-55 2.8 lens and love it, but I would not spend time on this experiment. I think I would need to be lucky to get good results. Taking flat objects relatively close is always difficult and this lens on my 40D would really struggle to give the right focus lock. I would be in manual focus and live-view in no-time, pixel peeping on ++++, and likely be changing lenses shortly after, probably to my 100mm macro. In my simple-minded world I would not be dreaming about getting my 17-55 calibrated or selling it on for not doing something it was never designed to do - but perhaps I am too easy going.
In theory the DOF should not be a problem, but getting the camera to focus is. I hear the 7D is a miracle worker in this respect but even so ....
I feel for Divamum's frustration. When you buy the L glass you expect miracles and rightly so. When you buy an expensive zoom for its 2.8 properties on a wide angle you are not likely to be consoled that it performs brilliantly at f5.6 at 70mm. And once you start doubting your equipment the comfort factor is lost for ever. Best to get shot of it.
For the record, I only started taking "clinical" test shots such as these because I noticed a problem when using it out in the field in real life conditions, which are all that counts to me. But having detected a potential problem, this was a good way to start checking what was going on in somewhat controlled conditions. I'm really not one of life's pixel-peepers-for-its-own-sake - I only do it to check sufficient sharpness when culling/processing and then move on. I'm far more interested in real-life tests than numbers!
Also, I do shoot in the 4-8ft range more than you might think, so that's not necessarily entirely arbitrary. I have a small working space, so sometimes am closer than ideal.
MF a complete no-go for me. I shoot too fast and my eyes aren't good enough - even bad AF typically produces better results than manual for me in situations where things are moving! Using live-view and slowing things down would be pretty much impossible in theatre situations, and would really change my shooting style/energy in portrait sessions in a way I don't particularly want to do - I like to handhold and keep moving.
As for the 17-55is ... One of my main reasons for adding this lens is to get a bit more reach at the long end (the Sigma 50-150 was actually my first choice, but I wasn't able to find one despite months of looking, and now they've discontinued it with the new version not expeted for a while), so the 17-55is would defeat the purpose since I already have a Tamron 17-50. The beauty of the Canon over the Tam is its blazing fast AF (even in dim light - so impressed by that) and the extra FL between 50-70, which is a place I have often felt lacking (I have 85 & 135 primes). Having used it, I'm happy as can be with the FL, as it absolutely supplies what I've been missing. (although I'm happy to know i have 17mm should I need it, so the Tam is staying in the bag for the time being).
So, waiting to hear back from KEH. There is much to like about this lens and since it does perform nicely at 45-70, it will still be useful to me regardless of how this plays out. But it would be even nicer if I can improve the results at the wide end, too! I'm not expecting it to be as sharp as a prime, but I'm not sure it's performing up to par as it stands. We'll see where this goes...
Okay Divamum... I have the same 7D body and this EFS17-55 IS lens. I was puzzled by goldenb's remark so decided to test this on my 7D because I think the 24-70L will have a very hard time to beat this lens on a 7D (actually, the 24-70L will not win this I think... help us here Ziggy :-)
The link to my test gallery incl. EXIF : http://sv-jedi.smugmug.com/Other/dgrin/17-55-test/16379070_Utcd7
The shots are f2.8, f4.0, f5.6 & f8.0 for both 17mm and 55mm. Single (center) point AF, Aperture priority mode. 1:1 crops
The f2.8 images for 55mm and 17mm:
#1 55mm f2.8
#2 17mm f2.8
Looks pretty perfect to me ! I would not hesitate to return that 24-70L and get the EFS17-55 instead !!
Edit: ah Tamron 17-50 ... well, sell it
ciao!
Nick.
Nick.
my equipment: Canon 5D2, 7D, full list here
my Smugmug site: here
KEH rock. Just got off the phone with them, I can send it back no problem (despite my goof on the dates), they'll take a look at it and see what's going on and either repair or replace it as appropriate and subject to availability. Frustrating to have to send it back since when it's delivering I'm in love with what it gives me (I'm such a bokeh junkie it's not even funny), but always pleases me to see a retailer living up to reputation - they're definitely one of the good guys
Love the 17-55is but the question is how it does on the 40D after reading goldenb's post.
I have this gap between 55m and 70mm too and to be honest I've been waiting for a new 24-70L model. The current one is from 2002 and I really need a good chance to get a good copy as I'm thousands of miles away from where they are sold... I went for the 17-55is after Ziggy warned me about the bad copies...
You should try and ask the good people at KEH to find you a good copy of the 24-70L. They must know there's good and bad ones and you might just charm them enough to do that for you :-)
ciao!
Nick.
Nick.
my equipment: Canon 5D2, 7D, full list here
my Smugmug site: here
Natural selection is responsible for every living thing that exists.
D3s, D500, D5300, and way more glass than the wife knows about.
Finally, I took the plunge with this very well-priced BGN-rated lens from KEH and, having used it, there's no doubt I *need* this lens- the focal length, the super-fast AF, colour, contrast, gorgeous bokeh, the excellent performance at the telephoto end, even wide open... Now if I can just get one that performs properly across it's entire FL, it will do very nicely indeed.
I've had great luck with KEH's bgn rating in the past (one of the best lenses I've ever owned was a BGN 100 f2 from them, and if it weren't that I had to sell it to pay for the 135L I'd never have let it go - that thing would cut steak it was so sharp). I'm delighted that they're delivering even when there's a problem, too - so nice to see a company respond to a customer issue on a case-by-case basis instead of resorting to the corporate manual diktum from on high. Seems like photography-land has a few of those (yes, Helen, I'm looking at you
Natural selection is responsible for every living thing that exists.
D3s, D500, D5300, and way more glass than the wife knows about.
I even have an "UG" grade (from KEH), Sigma 70-210mm, f2.8 APO that optically is one of my best lenses, compared even to the Canon equivalent "L" series lenses.
What made it "UG" grade? It has no tripod ring. It came without caps or hood. The rubber trim is a bit sticky. The lens will only shoot at f2.8 and the AF is very slow and sometimes fails altogether. (Manual focus is fine however and I do get focus confirmation that is very accurate.) For the little that I paid for the lens, it works fine for those high-risk occasions when I hesitate to use the "good" lenses in the range and I only need f2.8. One of my favorite images came from this lens too (during initial testing).
Titled: "Try to keep up."
I'm not suggesting that people "should" buy BGN and UG grade stuff but, with KEH the only real capital risk is shipping and insurance costs.
Moderator of the Cameras and Accessories forums
Stunning shot, Ziggy!
Manually applied in Photoshop as brush shapes. I tried to mimic what a star filter would do and I placed the stars over spectral highlights and varied the intensity of the pattern according to the intensity of the highlight, avoiding conflicts with subject matter as I could.
Moderator of the Cameras and Accessories forums
Firstly, the guys at KEH were absolutely terrific and went the extra mile to work this out. No problem with returning the first copy (Lens A). They also advised me to watch the website, and if there was another BGN one listed to call them so they could set it aside for me until the first one arrived back with them and they could swap it out.
Amazingly (since they don't turn up all that regularly - I know this because I've been stalking them for months!), another one (Lens
Lens B arrived yesterday. Although it's an older date code, it is in much better physical/external condition - a couple of "normal wear and tear" scratches on the housing, but nothing like the deep scuffs and dings of Lens A.
More importantly, some images. SOOC entirely unsharpened/unmodified in anyway (camera setting neutral, no PP) full-size jpgs shot on a 7d. Since we have some sun and some flowers, I used those as tests - photographing dollar bills on the door kind of makes me crazy
I deliberately chose a scene I thought might confuse the AF - here's the uncropped 24mm
This one seems to hit peak 2.8 performance around 45mm
70mm 2.8
It's still not as breathtaking at the wide end as one would hope, but having now looked at MANY sample shots from this lens from all the reputable reviewers, I think it's within expectations for what the model generally delivers; interestingly, too, it sharpens up considerably with as small a change in zoom as 27mm.
Zoom browser shows that in this shot the focus juussst caught that little branch in the foreground, hence the difference in focal point from the previous shots (I was handholding and there was a slight breeze, too). At least with Lens B there IS a point of focus - that was my big issue with Lens A at 24mm, where sometimes nothing in the image would be sharp.
Here's another (deep crop) at 70mm 2.8, entirely unaltered.
So, what think ye - better? I'm happier with this one so far, although I do find it hard to judge out of the context of an active shoot (which is my problem, not the equipment's!). I have a shoot next week where it may see some action, so will (as before) hold off final judgement until then, but so far this one does seem a LOT more predictable and consistent.
Once again, a shout out to the guys at KEH - really appreciated their prompt attention, reasonable approach, and good advice. They were great, and based on this experience I'd say they're not only good when things are going right, but you can buy from them with absolute confidence that they'll handle a problem well. A+++++++ all round.