is my 24-105 a soft copy?

Messages
577
Name
Dan
Edit My Images
No
I really cant see any difference in sharpness between my 24-105L & my 18-55is kit lens.

Having never owned an L lens before i'm not sure if i have a soft 24-105 or a very sharp 18-55?! neither are as sharp as the 50mm 1.8 prime i have

Im slightly dissapointed, i was expecting to see noticable improvement in detail being an 'L' lens n all that! ive tried to take sample pics to compare....

24-105

GONK:
photostream


HAND GEL:
photostream


18-55is

GONK: (sun going in)
http://www.flickr.com/photos/dan-jam/6059079688/in/photostream

HAND GEL:
photostream




is there anywhere i can get my lens tested profesionally so i can decide if its a bad'n or not? (Swindon area)
 
oops, not all links appear to be working. there are only 5 pics on my flickr & they are tagged at the bottom of the pic which lens was used.

its probably too hard to see from these pics but when viewed at 100% i cant see any differance in detail. i am a complete amature btw
 
Looking at the two gonk photos the lighting is completely different so you cant compare.

We'd also need a little more detail, like did you shoot raw (and if so how did you process to jpeg), speeds etc.
 
That's about the worst test target possible.

Get outdoors and pick a nice detailed scene and analyse the full frame of your camera for a start.

24-105mm should be pin sharp unless missfocused.
 
hmm okay, i'll try a different scene.

should i use centre focus so i know where the sharpest point should be?
 
We need info on apertures, ISO, focal length, tripod, mirror lockup, etc.
These can degrade image quality when viewed at 100% to the point that there will be no difference between the two lenses even when the same settings are used.
To get the best IQ requires proper technique AND a good lens.
The kit lenses weak spot will be on the edges at the widest aperture; the 24-105 at the same settings should be WAY better.

Having said that....
I've been through this loop and my 24-105 was a soft copy; they do exist.
I did a set of back to back tests with other 24-105s and learned that getting images that can be meaningfully compared is surprisingly difficult, but not impossible.
Having gathered the evidence and emailed it to Canon they were extremely helpful and sorted it for me.

And having admitted that there are soft copies.....
Even once I got my lens sorted and had upgraded to a 5DII, my image quality was only tolerable, it still looked rubbish at 100%
I spent a long time working out that the biggest problem was my technique, not my kit.
I spent over a year seriously working out what was wrong, and fixing it.
If I had to go back to my 20D and kit lens, I'm sure I'd now be able to take WAY better images.
 
I see, so some can spend hours picking holes in their lenses justifying its price, even though they'll never print at full size or cant see the entire image on their screen at 100% mag.
 
Out of interest, why would you look at your pics at 100%?

Isn't that the best way to determine if your lens is a good copy or not?

I may not have the experience or ability as some of the best of you on here but that doesnt mean i should accept a product that isnt as it should be!
This lens was certainly not cheap for my budget & although im aware it's the person behind the camera that takes a good pic, i would certainly like my equipment to be upto scratch.

I would like to have a great deal of knowledge to call on to go about testing lenses, but i dont. However what i can see on my p.c is that what i would call like for like images looking very similar in terms of detail (sharpness) between my new L lens & the kit lens.
i'm just wanting to know if its 'operator trouble' as opposed to faulty equipment before i stress whether to send it back or not , or were my expectations of this L lens just too high?
 
Yeah - lets not start that old argument :bang:

To take an image that looks sharp at 100% corner to corner requires perfect technique and is rarely achieved in the real world.
But if you know HOW to take a top quality image, then at least you are aware of the compromises when you take an image and how they will affect the final image when viewed or printed.

But the OP is wondering if they have a soft lens copy.
Pixel peeping is needed, but it MUST be combined with proper technique.
 
looking at the exif 18-55 F10 and 24-10f F8 there probably isn't going to be a lot of difference at centre sharpness.........
 
I see, so some can spend hours picking holes in their lenses justifying its price, even though they'll never print at full size or cant see the entire image on their screen at 100% mag.

err... no.
 
If it was resting on a solid surface or tripod when this shot was taken try again with IS off. 100% doesnt always show softness any better than you can see softness when looking at a pic as a whole image such as your flickr pics.

I think people are surprised when they expect huge leaps in quality over a kit lens because theyre paying so much, often kit lenses in my experience are pretty good stopped down in sharpness. The extra cost is sometimes for minimal increases in sharpness but youre also paying for a faster lens thats sharper wide open, colour, contrast, etc. depending on the lens.
 
Last edited:
Glad we cleared that up, I was beginning to worry about your blood pressure. :LOL:
 
Ah... the old 24-70 versus 24-105 debate :bang:
Opinion is split is about 50/50 on which is the best - they are both good lenses (but neither are perfect).
It's pretty pointless arguing as the OP already has the 24-105.
 
Last edited:
Ah... the old 24-70 versus 24-105 debate :bang:
Opinion is split is about 50/50 on which is the best.
It's pretty pointless arguing as the OP already has the 24-105.

Who's arguing
I pointed out that I had a 24-105mm Lens and was never happy with it
and I suspect any other member of this forum with a Lens they have or had would also say whether they had a good or bad copy
the original poster a has asked about the sharpness of the Lens
personally after my experience with it I would never buy it again, a good mate of mine currently has one and loves it
I dont
my 2p
 
Out of interest, would a local camera store have the ability to put my new lens through its paces?
For a small fee ofcourse
 
looking at the exif 18-55 F10 and 24-10f F8 there probably isn't going to be a lot of difference at centre sharpness.........

Sorry, nothing to do with this thread (except when I was on the light side I had a 24-105 and loved it - sharp as a really sharp thing) but your bug shots are fantastic Mr Dogfish_magnet:clap:
 
Sorry, nothing to do with this thread (except when I was on the light side I had a 24-105 and loved it - sharp as a really sharp thing) but your bug shots are fantastic Mr Dogfish_magnet:clap:

The bug in the water droplet is just Amazing
 
Duncan, how about giving the OP some advice on HOW to take a top quality image as he is new here? Or at least point him to a thread if there is one for a bit of guidance. You excel at this kind of info ;)

Caroline

Yeah - lets not start that old argument :bang:

To take an image that looks sharp at 100% corner to corner requires perfect technique and is rarely achieved in the real world.
But if you know HOW to take a top quality image, then at least you are aware of the compromises when you take an image and how they will affect the final image when viewed or printed.

But the OP is wondering if they have a soft lens copy.
Pixel peeping is needed, but it MUST be combined with proper technique.
 
Caroline - you have a point.
Having just read another thread where someone advised checking front/back focussing by simply focussing on an angled newspaper and having a look - it's not that simple! Sometimes reading this stuff makes me grumpy, but most of the time I let it go... But since you asked...

I learned this stuff when I upgraded from a 20D to a 5DII and my image quality was still very disappointing. The problem was almost 100% my technique. I spent two years looking at what I was doing wrong and learning what I should do instead.

My definitions... Acceptably Sharp - good enough for a decent print at A3+ (may look awful viewed at 100%). Critically Sharp - looks sharp zoomed to 100%; also known as pixel peeping; only of value when printing to A1 or bigger, or as a technical exercise in improving your camera skills.

Auto Focus - is only acceptably sharp. For tests manually defocus and then half press to auto focus; defocus both closer and further than the subject. There will be quite a lot of variation; take at least 3 and keep the sharpest. For testing lenses don't use auto focus; use live-view zoomed to 10x and manually focus; there's no front/back focus issues, it's WYSIWYG.

Image Stabilisation - is only acceptably sharp. IS hunts and settles, during an exposure this shows as motion blur. How much of a problem this is depends on the equipment. For testing lenses, the safest thing to do is turn off IS.

Mirror Lock Up - essential! There are some exposure combinations that produce motion blur on my 24-105 with the 5DII, even on a tripod with IS turned on and the results are bad enough to see viewing full frame. An alternative is to use live-view, which tests indicate gives better results than MLU; it's certainly a lot more convenient.

Cable Release - essential. No matter how careful you are, If you use the shutter button you will end up with motion blur when the image is viewed at 100%. A top quality tripod will only reduce the problem, it cannot eliminate it. An alternative to the cable release is to use the self-timer.

Tripod - essential. Needed for MLU, cable release and in order to do a comparison, the the images should be as similar as possible.

Hyperfocal distance - only acceptably sharp. A perfectly set up hyperfocal distance image will always have a soft horizon when viewed at 100%. The key term to Google for is Circle of Confusion. To be critically sharp you need to rely on far less depth of field than normal.

Aperture - For diffraction, the rule of thumb is that f16 on an APSC sensor (f20 on full frame) is only sharp enough to produce an 8x10" print; this is caused by diffraction and is nothing to do with the lens. Going the other way, opening up the aperture shows any inherent softness in the lens. Consequently, for a critically sharp image then the only useful apertures are f5.6 or f8 (special caveat for using just the centre of the lens, or the amazing 24mm TSE II)

ISO - one of the side effects of upping the ISO is a reduction in the amount of fine detail. For a critically sharp image, use base ISO.

UV filters - I've seen some shocking results from poor quality filters being used to protect the lens. Usually a loss of fine detail and a loss of contrast, especially in high contrast scenes. Need to do your own tests, there will always be some degradation but need to make your own decision about whether it is significant. When testing lenses take off the UV filter.

I'm pretty sure that's not an extensive list - I've written it straight into this thread without referring to anything. My main point is to illustrate that getting a critically sharp image is not trivial, it requires perfect technique. There may be one or two points people may choose to regard as contentious; that's life.

P.S. I don't aim for every image I take to be critically sharp - that would be crazy. Taking images is a compromise needed for real world conditions. But at least I know what effect the compromise is having on my finished image, and most importantly, whether it will affect an A3+ print.
 
Great post Duncan!

Now for the benefit of those of us who can't contemplate going through the procedures you suggest, and are resigned to accepting what we've got, what are your must do steps when taking a pic out in the wilds of Snowdonia or Scotland?

Caroline
 
Great post Duncan!

Now for the benefit of those of us who can't contemplate going through the procedures you suggest, and are resigned to accepting what we've got, what are your must do steps when taking a pic out in the wilds of Snowdonia or Scotland?

Caroline

Hmm - this started as a fairly quick response, but seems to have turned into war and peace...
And I'm beginning to think this isn't the right thread for this stuff to be in...
However - it will be interesting to see peoples reactions as although what I've written works for me I know that many people will have serious reservations.

First - and most important
Image quality is mostly good technique, not good equipment.

I'll happily share how I use/abuse the settings on my 5DII, but other cameras will be slightly different, I'm afraid there is no substitute for doing your own tests. Also, these recommendations are aimed at landscape photography and may not always be applicable.

The short version... On the 5DII, more than 90% of my shots are:
- hand-held
- Aperture priority, f11
- Auto ISO
- Auto focus
- White balance: Sunny
- EV -1/3
The rest of this post explains why....

Aperture - normally between f5.6 and f11; anything else is a creative decision. Smaller aperture softens the image through diffraction. Wider aperture softens the image through imperfect lens construction. Because aperture has so little flexibility I usually set the camera to aperture priority. I do use other apertures, but there is always a purpose such as a shallow DoF or artificially lengthening the exposure. The cool trick my TSE lens pulls off is that I can get extreme DoF (e.g. 30cm to the horizon) using only f5.6 resulting in wonderful rich texture than can be printed at any size.

Focus - I've never been very good at focussing manually. If I'm hand-holding I'll use my experience to pick a suitable point to use for auto-focus, focus and then recompose; I'll often check I got it right by reviewing the image I've just taken. If I'm on a tripod, then I'll use live view with DoF preview to check focus is exactly where I want.

Know your lenses; with the aperture wide open, image corners will be soft and the centre will be much sharper, how bad the edge softness is and whether it can be overcome by stopping down depends on the lens. Zooms have different results at different focal lengths. All lenses have weak spots, for example the 24-105 is slightly soft across the whole frame at 105mm f4 and stopping down to f5.6 only regains centre sharpness. Another example is that the 24 TSE II has a spherical plane of focus which catches out a lot of people who think it may have soft corners. I recommend checking the performance of your lenses by reading reviews such as those at SLR Gear.

ISO - I use Auto ISO. It works extremely well. It takes ISO up to a maximum of ISO 3200. Through testing I know an ISO 3200 image can be printed with no noise reduction and create a clean A3+ print with accurate colours.

Shutter Speed - whatever the camera wants (aperture priority and auto ISO). From experience I know my hand-holding technique is good enough that I can ignore shutter speed until I hear a pronounced clunk-clunk when I take a shot. That means the camera has already got to ISO 3200 and is struggling with low light; time to either open the aperture right up (compromise image quality) or up the ISO (compromising image quality) or start using a tripod. Oddly, this rule seems to apply to all my lenses including the 70-200 f4 IS L; I've found that 4 shots taken hand-held and standing with no support at 200mm 1/6s will give about 3 acceptably sharp images.

Tripod - using a tripod is a creative decision, most of my images are hand held... IMPORTANT - as soon as the camera goes on the tripod a LOT of settings have to change; ISO 100, MLU (or live view), cable release (or self timer), hot shoe bubble (or I know my horizons will be even wonkier). Live view is great for checking the image has a strong composition and for checking clutter in the background or around the edges. Use live view to check focus (including stopping down the lens). The tripod slows me down and making me less impulsive; sometimes this is good for my images, sometimes bad.

RAW - always... I stopped using JPEG when RAW Shooter Essentials demonstrated that image workflow was faster in RAW than JPEG as well as having all the well known RAW advantages. I reckon I process about 200 landscape images an hour including removing duplicates, removing duffers, checking sharpness, adjusting the tones, cropping, grading, setting copyright and adding keywords - everything except printing and publishing to my Blog.

White balance - for landscapes I usually set WB to Sunny. Colours look 'right'; a storm will look moody and woodland shade looks invitingly cool. Auto WB is to be avoided as it will look at a woodland scene and go 'Ooo Flourescent' and dial out the green cast and look at a sunset and go 'Ooo Tungsten' and dial out the orange cast. And of course, taking in RAW means the WB can always be changed in PP.

Exposure Compensation - for landscapes I find the 5DII needs -1/3 EV (for comparison, my 60D needs no compensation). Through experience I find this gives the best compromise for recovering highlight detail and still retain detail in the shadows. All cameras are different.

Histogram - the 5DII screen gives a very good feel for the exposure (it senses the ambient light), so I use histograms far less that I do with other cameras; but these notes are still relevant. Always use the three RGB histograms (not the white histogram), this is because a strongly coloured image can blow one of the channels without showing a clipping indicator, or showing any sign of trouble on the white histogram. The classic problem is a red sunset where the bright bits of the sky are captured yellow instead of red; this is an artefact of the red channel blowing; the fix is to underexpose until the yellow turns back to red.

Graduated Filters - a landscape scene often has more dynamic range than the camera is capable of capturing. Graduated filters are a way of reducing the dynamic range and are a good alternative to exposure bracketing merged in post processing. If I've gone to the trouble of setting up my tripod, then I'll use a graduated filter if it is needed. I'll also consider using a graduated filter hand-held as my objective is not to produce the perfect image straight out of camera, but to control the dynamic range; if the edge of the grad isn't perfect then I'll fix it in PP.

In summary:
1) Practice makes perfect
2) Look critically at your images and learn from your mistakes
 
Just wanted to thank you Duncan for taking so much time to put these posts together. I'm quite new to photography, have loads to learn and tips like these only make the learning process quicker and more enjoyable.

You have some wonderful shots on your website, I really enjoyed looking through them.

Thanks again!
 
Thanks (y)

I would just like to re-itterate...
1) My first chunky post does not apply to real world shooting (unless shooting for billboards). It's for testing lenses and to satisfy pixel peepers looking for ultimate sharpness.
2) My second chunky post is how I use my 5DII and may not apply to other people using a 5DII (we are all different) and probably doesn't apply to other DSLRs. Think of it as a guide for things to try out and see what works for you.
 
Only just seen your excellent post Duncan... Well explained :)

2) My second chunky post is how I use my 5DII and may not apply to other people using a 5DII
It applies here (a few minor differences here and there - I mainly use f8 for example)
 
Andy - cheers :)

I'd rather use f5.6 all the time; my lenses give their best results at that aperture.
But it depends what mood I'm in.
a) If I'm a wandering through the landscape taking what I see, then it's hand held f11; this give me a little lee-way for the 24 TSE guessing where the focal plane is and the 70-200 gives me a DoF that is deeper than 1".
b) But if I'm spending half a day on the same 200m stretch of coast then it's tripod and a slow considered approach at the shallowest aperture I can get away with, checking DoF preview in Live View the whole time.
 
Hi Duncan. I think that has to be some of the best tips I've read on this forum so far. How much of that I can use on my D90 I'm not sure but it has certainly given me food for thought. You have a knack for explaining things in easy to understand language which most will appreciate. Thanks for taking the time to post.(y)
 
Thank you for your two chunky posts Duncan. The first was interesting and the second genuinely illuminating - a very helpful insight into your technique that has given me lots of food for thought.
 
Cheers Paul (y)
Yeah - the first post was a bit of a pixel peeping geek-fest :help:

I reckon I take only a handful of images a year that look sharp at 100% corner to corner right across the whole frame.
I've had a few printed to A1 and it's amazing.
I call it an immersive viewing experience - where the image looks good at a distance, but as you walk close the detail draws you in till you end up studying a tiny piece of the whole image; there's so much detail that even viewed only a hand span from the print it feels like if you could get closer there would be more detail to be found.

Why don't I try to make more images like this?
Easy - the real world requires compromise.

It takes a special subject to allow the perfect shot to be taken (e.g. shallow DoF), and a very geeky mindset to make the most of it.
At this point in my photographic journey I'm quite happy aiming for A3+ prints; I can enjoy my photography without getting caught up in the techy stuff.
I'm aware of the compromises I'm making and I'm happy with the sacrifice in ultimate image quality.
 
Back
Top