First piece of L glass - how to test

Hi

Just got my first piece of L glass (24-105 L IS USM) for my 20D. First impressions are good but my problem is how to assure I truly have a good copy after reading all sorts of horror stories on various forum topics with L quality control issues. In other words - how do I know how good L glass should be... any suggestions on how to test thoroughly to make sure my money has been well spent or comments on the sample below would be much appreciated.

Thanks
Al.

Pic here is 100% crop sample, shot RAW and processed with ACR on full auto with no sharpening
ISO100
35mm
f/5.6
1/4s on tripod

Tripods and ruler shots are good for sharpness tests, but I personally think 1/4 second leaves too much to chance. If you're stepping back from the camera when the shutter is open you'll create enough vibration to possibly fuzz the shot on you. Shutter speed shouldn't be a factor in lens sharpness, so I'd focus on focal length and aperture as your variables, and try to keep the shutter speed way up.

I'd say ratchet up the lighting until you can shoot at 1/100 and ISO 200 for sharpness tests for that lens.

That said, I never did any ruler, mathematical or lab-type tests when I got my 70-200 f/4 or my 28-70 f/2.8. I just shot handheld with plenty of light at each focal length and f-stop until I got a "good enough" result. Like you, I shot in RAW and left everything unsharpened. That let me know that the lens was capable of taking good pictures, and any bad ones were my fault. If I recall correctly, it took about an hour and about 300 shots per lens to get that, but certainty is a wonderful thing for us engineers.

I'm beginning to learn that sharpness is relative. "Resolving power" isn't, but it's much harder to test/compare that accurately.

jb

--

--Laws alone can not secure freedom of expression; in order that every man present his views without penalty there must be spirit of tolerance in the entire population.
--Albert Einstein

don't use a focusing chart!
causes more trouble then it is worth
for a VERY NON SCIENTIFIC TEST

go round your house taken shots of things at different focal lengths and at different distance ... shoot wide open (F/4) .., and have a look at the results

--
AJ
http://www.pbase.com/manjade

Jason said:

Tripods and ruler shots are good for sharpness tests, but I
personally think 1/4 second leaves too much to chance. If you're
stepping back from the camera when the shutter is open you'll
create enough vibration to possibly fuzz the shot on you. Shutter
speed shouldn't be a factor in lens sharpness, so I'd focus on
focal length and aperture as your variables, and try to keep the
shutter speed way up.

Click to expand...

Thanks Jason - that makes sense... was indoors and only had on-camera flash and wanted to keep it at ISO100 but agree that I could have gone to ISO200 to get the shutter speed up a bit. Will get it out of the house this weekend for some natural light shots to get the shutter speed way up.

Jason198448 said:

I'd say ratchet up the lighting until you can shoot at 1/100 and
ISO 200 for sharpness tests for that lens.

Click to expand...

Agreed - will do

Jason198448 said:

That said, I never did any ruler, mathematical or lab-type tests
when I got my 70-200 f/4 or my 28-70 f/2.8. I just shot handheld
with plenty of light at each focal length and f-stop until I got a
"good enough" result. Like you, I shot in RAW and left everything
unsharpened. That let me know that the lens was capable of taking
good pictures, and any bad ones were my fault. If I recall
correctly, it took about an hour and about 300 shots per lens to
get that, but certainty is a wonderful thing for us engineers.

I'm beginning to learn that sharpness is relative. "Resolving
power" isn't, but it's much harder to test/compare that accurately.

Click to expand...

Yep - a thorough days shooting to eliminate the inevitable user error is whats required!

Could you explain more what you mean by resolving power vs sharpness

Manjade said:

don't use a focusing chart!
causes more trouble then it is worth
for a VERY NON SCIENTIFIC TEST
go round your house taken shots of things at different focal
lengths and at different distance ... shoot wide open (F/4) .., and
have a look at the results

Click to expand...

Yeah - based on jason's post too, I think it is a numbers game and just a question of spending some serious test time this weekend with my L!

If you think you've got a bad apple, the only way is to take a lot of pictures under different circumstances. My first copy of the 24-105 had a lot of CA, my second and last has non, but is less sharp at the wide end.
Seems you cant get it all with such a cheap lens.

Kind Regards
tage

Thesun said:

If you think you've got a bad apple, the only way is to take a lot
of pictures under different circumstances. My first copy of the
24-105 had a lot of CA, my second and last has non, but is less
sharp at the wide end.

Click to expand...

Well - thats my problem... I am not sure I do have a bad apple since I have no reference to compare to. Certainly kicks the kit-lens in to touch, but not seeing any improvement over my 50mm 1.8.

Thesun said:

Seems you cant get it all with such a cheap lens.

Click to expand...

Yep - absolute bargain

If you expect it to even come close your 50mm, your in for a fall. At f4 the 50mm will perform better in every way. A good copy (a few lucky ones have them) should give you L performance from 24mm and up at all aperatures up to f8 - f11.

Many of the pictures in my gallery is taken with 30d and the 24-105:
http://www.solberghifi.no/galleri-index.htm

Tage

go take some real photos. If you like them, be happy.

--
http://www.pbase.com/mkelpie

I'm at work, so I'll have to make this somewhat quick, and the optics section of physics was my weakest subject, so if I'm vague it's because I'm trying to just go with what I know for certain. LeeJay (ljfinger) or Doug Kerr are much better at splitting these hairs than I am.

Sharpness is a somewhat relative or loosely used term to describe how much detail can be attained in an image. It started out as a descriptor for resolution, but it's overused and poorly defined in a lot of places. For a more exact measure of how "sharp" a lens is, you would have to check it's resolving power (usually measured in lines per inch or LPI) against whatever specs it was built to provide. You'd need to take into a account the documented resolving power of the lens, the resolving power of the sensor, and then any aberration you'd get from the anti-aliasing filter, sensor noise, flare, or whether or not Mars is in retrograde and other such celestial events. You'd also have to test at apertures that are known to be free of any refractive errors, which IIRC would cover f/4 to f/16 on that lens, but I'm not certain of that.

In short (too late), the tests for resolving power are much more sensitive than simple focus charts and ruler shots. They involve a lot of stuff that isn't necessary to take good, thought-provoking photographs, which is why I never really focused on that stuff too much. (pardon the pun) Like I was saying, stick to what you feel comfortable with, which would be taking a lot of photos using a lot of light and good technique, then review them and make sure that you've got at least one "good" (defined by you) photo for each focal length and aperture.

Here was one from my 28-70 that I felt was "good enough":



and full size:
http://www.nightyear.net/gallery/Favorites/herpoofyness?full=1

More detail to follow, hopefully by those that can fill in the rather gaping holes I left in this post,
jb

--

--Laws alone can not secure freedom of expression; in order that every man present his views without penalty there must be spirit of tolerance in the entire population.
--Albert Einstein

Mark Caplin said:

go take some real photos. If you like them, be happy.

Click to expand...

Oh yes... will be out and about this weekend. I guess I just hate the idea of paying a lot of money for something that can vary in quality. If I buy an iPod I get an iPod... they are to all intents and purposes identical. If I buy a BMW car, I get a BMW car.

But with high quality optics, that doesn't seem to be possible and you are then left to the mercy of the stock picker to see if you end up with something that blows you away or not!

You will need 12 more pieces, if you buy them in groups; if you want to glue them together yourself, you need 17 more pieces.

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

then your lens clearly is a lemon.

Otherwise, it looked ok to me. Congratulations.
--
When I ask which Canon lenses are best,
people tell me to 'go to L.'

Not to criticize, but I find it somewhat laughable that so many people get hung up on "good copy / bad copy" etc.. and instead of going out into the field and shooting something meaningful with their new lens and seeing the results.. the first shots are done in poor light, usually in a kitchen or other cramped area - and its usually of a ruler and a piece of fruit, dollar bills, coins or other boring subjects.

The photo of the ruler and apple in this thread looks worse than a snapshot from a PS camera.. doing a tech shot like that is somewhat tricky and usually requires side / top lighting to get the most out of the lens and a very steady tripod, mirror lockup and remote trigger.

You can check your copy for sharpness by going out and shooting some buildings with x/y/z angles, wide angle architectural photos and some close ups of people and animals (using a tripod or high shutter speeds, not low ones)..

Then and only then will you know if your lens is producing the images you would like.. not to mention we would get to see some interesting photos of pleasing subjects!

Mark

--
http://www.sixgunphoto.com

Certainly there are some basic areas to test like correct AF calibration and image sharpness across the frame. Unfortunately testing is a rigourous and exacting processes seemingly unsuitable to many individuals, seasoned photographers and novices alike. There are so many ways to introduce error and/or mis-interpret results that rather than lay out testing specifics and suffer an inevitable dialog of confusion, the community seems to let it go 'every man for himself'. If you are happy with it - its good enough for you.

So here's what I do. I do some basic testing for front and back focus in the store. If I suspect its out of calibration I ask to try another copy. Once I'm reasonably satisfied within the limitiations of my ability to test, I'll take it home. At home I'll liesurly do a more controlled AF test and test for sharpness across the frame. If something clearly stands out as trouble I'll exchange it. Then I'll take it out and put a whole bunch of frames on it. If after that it gives me some reason to doubt it based on my experience with other proven lenses I'll exchange it. If the lens recieved in exchange gets this far and it too leaves me with doubts I mark it up to design characteristics and either accept it or say that lens model isnt for me.

The problem you are going to have is your base-line for comparison is the 50/1.8 which is has aspects that are not fairly comparable to zooms with IS.

For every lens you buy from here on out, unless the new lens blows your mind right out of the box youll be asking yourself if you got enough 'wow factor' for your money. In the end you have to go with your gut feeling about it.
--
http://imageevent.com/chetcarson

Not to name names. I work with a former Canon Rep. We are in the mortgage banking industry. He was a Canon rep for quite a while. He shared some pretty interesting stories.

He had to accomidate a well known pro on a lens purchase at the facility. They went through 8 copies before the pro made his pick. The scenario ended with coments of off the line quality. I really value the info my co worker was able to share.

Im with Canon to stay. If im spending the bucks for L glass im going to make sure I'll qualify the lens.

This forum is a really good place for making people paranoid about whether their equipment is up to snuff, ESPECIALLY lenses. Take pictures with it and if you like them then it's a good lens. 'Nuff said.

--
Todd Walker...Canon TenD, Canon GThree, Canon EssDee500, Sony H2, Sony R1

http://www.toddwalker.net
http://www.pbase.com/twalker294
http://www.twphotography.net

Yep - thats the overriding theme. Will get out and about with it this weekend and try and get some serious clicks under it and then form an opinion from there

Thanks to all for their suggestions

So you say you want to test your lens to see if you've got a good copy. First, I think in order to adequately test the lens, you need to shoot real world stuff.

Second, you need to have discipline. Sadly, most people have no discipline. They run down to Mom's basement, turn on a single 40 watt bulb, turn on the camera and snap away, then they run upstairs, not bothering to figure out how to post samples or EXIF data, and in a huff, they start moaning about the poor quality of their L lens.

If you put stock in that type of a lens review, good luck.

Frankly, the bulk of the junk posts out there that talk about "quality control" are just that. Junk. Someone knows someone who heard of someone who bought an L lens who couldn't get shots in focus.

Frankly, there are a lot of newbies out there that like to talk BIG! Talk is cheap. And it's tremendously easy for one of these big talkers to knock someone else's gear.
Rarely do we see samples.
Even rarer still do we see samples taken by an EXPERIENCED photographer!

Most of these big talking newbies have never seen an L lens, never used an L lens, yet they repeatedly have to put down L lenses because they don't think it's worth the extra money... or they saw one shot on some post somewhere that was out-of-focus or not "tack sharp".

I'll tell you this: It's far easier for me to believe that the photographer is incompetent than it is for me to believe in "quality control issues" with Canon L lenses.

If there were really quality control issues with L lenses, you would have read magazine articles about it. There would have been stories written about it.

I'll go so far as to say that I don't trust sample shots or lens tests unless I know the person, or have a good idea of their skill set. There are WAY too many morons who don't understand the basic laws of physics. They don't understand light, lenses, or cameras, yet they insist on posting drivel about "quality control".

If you want to "test" your lens, use plenty of shutter speed to negate camera shake. Focus on high contrast, well-lit subjects. Shoot.
Look at your pictures. Shoot some more.

Don't go download some fancy test target without knowing what you want to accomplish, and how to go about it.

Good luck.

--

'In 1983, the game of golf had a firm grip on the waist of my boxers and was administering the death wedgie. I had a dose of the atomic yips and after missing 10 of 11 cuts by a single shot, I was ready to quit and apply for a job as a wringer-outer for a one-armed window cleaner.'

• David Feherty

Ah, yes. My sister bought a BMW too and it was a total lemon.

This is how things are done now. It's called consumer testing - assured quality is gone.

AlThomson said:

If I buy a BMW car, I get a BMW car.

But with high quality optics, that doesn't seem to be possible and
you are then left to the mercy of the stock picker to see if you
end up with something that blows you away or not!

Click to expand...