cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Trying to evaluate my lens and I need help - thanks

mangurian
Rising Star

I have taken a number of shots with my EOS-M100 using the 15-45 kit lens. I do not have a good eye and I would like help in determining if the lens is up to par for its type.

 

I took these shots in auto mode and converted the RAW photos to uncompressed tif files using IrfanView with no processing.  I have uploaded a sample to my Google Drive (link below).

 

If you would take a look and tell what you think, I would greatly appreciate it.

 

p.s. They look ok to me, but that means nothing.

 

Thanks in advance,

 

link to photos

1 ACCEPTED SOLUTION

TCampbell
Elite
Elite

A few things:

 

1 - is there a reason you are concerned about lens quality?  You mentioned they look ok to you.

 

2 - when trying to evaluate "lens quality" (specifically), ordinary photos are not really suitable.  The reason why can be a bit technical (and require a lot of reading).  If you're trying to test a lens for optical flaws, and you don't have a lab target, then it's best to use a large flat surface with lots of fine textured detail... such as a big flat brick wall.  The camera would be pointed at the brick wall so that it's perpendicular (not tilted left/right nor up/down).  You'd take some test shots both at "wide open" and also at "f/8".  

 

3 - it's good to know what someone is after before answering such questions.  For example... sometimes the reason someone isn't happy with a lens is because they are trying to achieve a certain "look" and the lens they are using isn't actually suitable to the task (it's not a question of the lens being flawed... it's an issue of the lens being the wrong type).  A common example is the desire to have a tack-sharp subject, with a nice creamy out-of-focus blur to the background.  This effect benefits from lenses that have a long focal length, but a very low focal ratio, and a close subject w/distant background - one would struggle to pull this off with a short focal length lens or a lens that doesn't have a low focal ratio.

 

4 - sometimes lenses that do not score the best, have some attribute which causes people to really like the images from that lens.  I'll use an example:  Lomography makes a lens called the "Petzval".  It's a reproduction of an old-world lens (and it's designed to "look" old).  Anyway... it has an optical effect that results in the background out-of-focus blur having a soft swirling characteristic to it that many find very pleasing.  This is technically an optical flaw.  But many people like it.  If you were to do some lab testing on the lens, it would score poorly.

 

Lens tests often score these optical qualities using something called "Modulation Transfer Function" (or MTF).  Here's the MTF chart for your lens:

 

Screen Shot 2018-08-08 at 12.13.24 PM.png

 

You can learn what these mean by reading this article: http://learn.usa.canon.com/resources/articles/2013/reading_MTF_charts.shtml

 

This chart is created using what Canon considers to be a good representative copy of the lens.  

 

Before getting into what this means, let me just say that this is actually a very good lens (based on that chart data).  This is better than the Canon EF-S 18-55mm f/3.5-5.6 IS STM lens which is the kit lens for most new Rebel bodies and it's substantially better than the EF-S 18-55 f/3.5-5.6 IS II lens that preceeded it.  It's not *quite* as good as the highly coveted Canon EF 24-70mm f/2.8L USM II (sometimes regarded as probaby the best lens in it's class ... from any vendor) but that's an "L" series lens costing $1600.  

 

I've committed a bit of a sin in that I'm comparing completely different lenses using MTF chart data (a bad idea) but I did this to point out that, as "kit" lenses go... this is lens scores better than most kit lenses.

 

Basically  the higher the lines are on the charts, the better the optical performance of that lens (e.g. users will generally regard that lens as being "sharper").  BTW, these tests do not test for everything.

 

The horizontal axis (labeled 0, 5, 10, and ending roughly at 14 even though it isn't marked 14) are the scores of the lens at that distance, measured in millimeters, from the center axis.   In other words the height of the graphed lines at the "0" point (far left edge) indicate how "sharp" the lens is at the precise optical center.  The far right (abouy 14mm away) is now sharp it is in the extreme corners of the frame.

 

The vertical axis (ranging from 0 to 1.0 in ".1" increments") is the quality score.  Generally anything higher than ".6" is considered "good" and higher than ".8" is very good to excellent.

 

The black lines represent how the lens scores when shot at "wide open" (whatever wide open is for the given lens being tested at the given focal length) and the blue lines indicate the performance of the lens when tested at f/8 (always at f/8 regardless of lens.)

 

The fine lines represent how the lens does at resolving fine detail.  The bold lines represetn how the lens does with contrast (coarse detail).  These measurements are made using blocks of parallel lines.  The "resolution" test uses fine lines very close together.  The "contrast" test uses bolder lines and they have larger gaps between them.

 

 

BE CAREFUL when reading MTF charts.   They can be misleading when trying to compare two different kinds of lenses.  E.g. if I were trying to compare a Canon 70-200mm f/2.8 version I to a Canon 70-200mm f/2.8 version II then I could use these MTF charts to get meaningful data.  But you would NOT want to use the charts to compare the performance of a Canon 70-200mm f/2.8 lens to, say... a Canon 18-55mm f/3.5-5.6 lens... that would be a meaningless comparison because nothing about those two lenses is very similar.  (Wait, didn't I just do something like this a few paragraphs early?  Yes.  Yes I did.  Bad me.)

 

Another thing to be aware of is that sometimess lenses that do NOT necessarily get the best MTF score... can produce more pleasing images.

 

For example, Lomography makes this reproduction Petzval lens with waterhouse stops (physical tabs that you slide into the lens to change the aperture size).  It's really NOT a high-quality lens.  But it creates this out-of-focus blur (bokeh) that has this very interesting swirling quality to it that some people find really attractive.  Technically that's a flaw.  The lens is less optically correct than a lens that does not do this.  But like I said... sometimes a flawed design can result in an attractive result that people actually like.

 

In other words ... don't get too obsessed with the scores.

 

Lastly, to test a lens, you really want to be able to control the conditions of the test.  

 

Suppose a lens was installed with de-centered optics.  That would be a manufacturing error that would effect that single copy of the lens.  This would result in the area of best focus not being in the center of the image.  It would also result in the amount of blur in one corner being greater than the amount of blur in the opposite corner (even though there should be symmetry ... there wont be if the optics are de-centered).

 

When you take an ordinary picture... there's no way to know if you've got de-centered optics.

 

It's generally a better idea to take a photo of a flat surface that has lots of detail and contrast... with the camera lens being perfectly perpendicular ("vector normal to the plane").  Now you can measure if one corner is sharper than the opposite corner.  This is something you can't do with a typical photo of a day out with the camera.

 

Tim Campbell
5D III, 5D IV, 60Da

View solution in original post

9 REPLIES 9

ebiggs1
Legend
Legend

As to techinical aspects the photos look fine to me. IQ, exposure, etc.  Composition could be better.

 

But what I wonder is, "I took these shots in auto mode and converted the RAW photos to uncompressed tif files using IrfanView with no processing."

You know enough to shoot Raw but not enough to not use auto mode?  The two don't really go hand in hand.

As to converting to tiff, again why?  Did you do a lot of post editing? If not converting to a tiff only increases the file size and gains you nothing.  A reason to use tiff is if you did a lot of processing with layers and masks and such.  A tiff will preserve all of that where a jpg won't. If you are just wanting a file that anybody can view just use jpg with a setting of 10.

EB
EOS 1DX and 1D Mk IV and less lenses then before!

You know enough to shoot Raw but not enough to not use auto mode?  The two don't really go hand in hand.

 

Since I was just thinking about the lens, I wanted to interfere in the process as little as possible I don't think manual vs auto will change the evaluation.  The RR shots were also my first outing with the camera and I didn't want to possibly screw up with manual settings.  It's the first "real" camera I have had - so no experience.

 

As to converting to tiff, again why?  Did you do a lot of post editing? If not converting to a tiff only increases the file size and gains you nothing.  A reason to use tiff is if you did a lot of processing with layers and masks and such. 

 

No post editing.  I wanted to post full size images and not everyone can open EOS-M100 .CR2 images in their graphics program.  I had the tiffs available so that's what I posted.  No space concern (maybe some download concern for responders).

 

A tiff will preserve all of that where a jeg won't. If you are just wanting a file that anybody can view just use jpg with a setting of 10.

 

The program I was using allows for very high quality jpegs, but not with zero compression.

BTW, Thanks for the feedback. I feel better knowing I don't have to return the lens as some others have had to.

If they look OK to you then why are you bothering to ask ?

 

It would have been nice to have the EXIF information because without knowing what aperture was used it is very difficult to assess the lens quality. Almost everyone on this forum should be able to open your .cr2 files so in future they would be far better to post than tif files.

 

To me the shots seem to have some corner softness and possibly some decentering showing as softness on the extreme left hand side but that may be perfectly normal for the lens if it was used at a wide aperture. I do not have any experience of the lens you have used so it is difficult for me to say whether this is normal or not.

"If they look OK to you then why are you bothering to ask ?"

 

Because that is how newbies learn.  Now I will look for "corner softness"  and "decentering"  (whatever that means).

 

Thanks,

I have now posted some CR2 files from the same shoot for those who wanted more info.  Same link.

TCampbell
Elite
Elite

A few things:

 

1 - is there a reason you are concerned about lens quality?  You mentioned they look ok to you.

 

2 - when trying to evaluate "lens quality" (specifically), ordinary photos are not really suitable.  The reason why can be a bit technical (and require a lot of reading).  If you're trying to test a lens for optical flaws, and you don't have a lab target, then it's best to use a large flat surface with lots of fine textured detail... such as a big flat brick wall.  The camera would be pointed at the brick wall so that it's perpendicular (not tilted left/right nor up/down).  You'd take some test shots both at "wide open" and also at "f/8".  

 

3 - it's good to know what someone is after before answering such questions.  For example... sometimes the reason someone isn't happy with a lens is because they are trying to achieve a certain "look" and the lens they are using isn't actually suitable to the task (it's not a question of the lens being flawed... it's an issue of the lens being the wrong type).  A common example is the desire to have a tack-sharp subject, with a nice creamy out-of-focus blur to the background.  This effect benefits from lenses that have a long focal length, but a very low focal ratio, and a close subject w/distant background - one would struggle to pull this off with a short focal length lens or a lens that doesn't have a low focal ratio.

 

4 - sometimes lenses that do not score the best, have some attribute which causes people to really like the images from that lens.  I'll use an example:  Lomography makes a lens called the "Petzval".  It's a reproduction of an old-world lens (and it's designed to "look" old).  Anyway... it has an optical effect that results in the background out-of-focus blur having a soft swirling characteristic to it that many find very pleasing.  This is technically an optical flaw.  But many people like it.  If you were to do some lab testing on the lens, it would score poorly.

 

Lens tests often score these optical qualities using something called "Modulation Transfer Function" (or MTF).  Here's the MTF chart for your lens:

 

Screen Shot 2018-08-08 at 12.13.24 PM.png

 

You can learn what these mean by reading this article: http://learn.usa.canon.com/resources/articles/2013/reading_MTF_charts.shtml

 

This chart is created using what Canon considers to be a good representative copy of the lens.  

 

Before getting into what this means, let me just say that this is actually a very good lens (based on that chart data).  This is better than the Canon EF-S 18-55mm f/3.5-5.6 IS STM lens which is the kit lens for most new Rebel bodies and it's substantially better than the EF-S 18-55 f/3.5-5.6 IS II lens that preceeded it.  It's not *quite* as good as the highly coveted Canon EF 24-70mm f/2.8L USM II (sometimes regarded as probaby the best lens in it's class ... from any vendor) but that's an "L" series lens costing $1600.  

 

I've committed a bit of a sin in that I'm comparing completely different lenses using MTF chart data (a bad idea) but I did this to point out that, as "kit" lenses go... this is lens scores better than most kit lenses.

 

Basically  the higher the lines are on the charts, the better the optical performance of that lens (e.g. users will generally regard that lens as being "sharper").  BTW, these tests do not test for everything.

 

The horizontal axis (labeled 0, 5, 10, and ending roughly at 14 even though it isn't marked 14) are the scores of the lens at that distance, measured in millimeters, from the center axis.   In other words the height of the graphed lines at the "0" point (far left edge) indicate how "sharp" the lens is at the precise optical center.  The far right (abouy 14mm away) is now sharp it is in the extreme corners of the frame.

 

The vertical axis (ranging from 0 to 1.0 in ".1" increments") is the quality score.  Generally anything higher than ".6" is considered "good" and higher than ".8" is very good to excellent.

 

The black lines represent how the lens scores when shot at "wide open" (whatever wide open is for the given lens being tested at the given focal length) and the blue lines indicate the performance of the lens when tested at f/8 (always at f/8 regardless of lens.)

 

The fine lines represent how the lens does at resolving fine detail.  The bold lines represetn how the lens does with contrast (coarse detail).  These measurements are made using blocks of parallel lines.  The "resolution" test uses fine lines very close together.  The "contrast" test uses bolder lines and they have larger gaps between them.

 

 

BE CAREFUL when reading MTF charts.   They can be misleading when trying to compare two different kinds of lenses.  E.g. if I were trying to compare a Canon 70-200mm f/2.8 version I to a Canon 70-200mm f/2.8 version II then I could use these MTF charts to get meaningful data.  But you would NOT want to use the charts to compare the performance of a Canon 70-200mm f/2.8 lens to, say... a Canon 18-55mm f/3.5-5.6 lens... that would be a meaningless comparison because nothing about those two lenses is very similar.  (Wait, didn't I just do something like this a few paragraphs early?  Yes.  Yes I did.  Bad me.)

 

Another thing to be aware of is that sometimess lenses that do NOT necessarily get the best MTF score... can produce more pleasing images.

 

For example, Lomography makes this reproduction Petzval lens with waterhouse stops (physical tabs that you slide into the lens to change the aperture size).  It's really NOT a high-quality lens.  But it creates this out-of-focus blur (bokeh) that has this very interesting swirling quality to it that some people find really attractive.  Technically that's a flaw.  The lens is less optically correct than a lens that does not do this.  But like I said... sometimes a flawed design can result in an attractive result that people actually like.

 

In other words ... don't get too obsessed with the scores.

 

Lastly, to test a lens, you really want to be able to control the conditions of the test.  

 

Suppose a lens was installed with de-centered optics.  That would be a manufacturing error that would effect that single copy of the lens.  This would result in the area of best focus not being in the center of the image.  It would also result in the amount of blur in one corner being greater than the amount of blur in the opposite corner (even though there should be symmetry ... there wont be if the optics are de-centered).

 

When you take an ordinary picture... there's no way to know if you've got de-centered optics.

 

It's generally a better idea to take a photo of a flat surface that has lots of detail and contrast... with the camera lens being perfectly perpendicular ("vector normal to the plane").  Now you can measure if one corner is sharper than the opposite corner.  This is something you can't do with a typical photo of a day out with the camera.

 

Tim Campbell
5D III, 5D IV, 60Da

TCampbell -  Thank you.

 

I believe I followed what you laid out so well.  I was a physicist and I am used to similar measures (although not identical).

 

I guess what I was looking for was a qualitative evaluation just to help me decide if my lens was defective enough to return for replacement while it is still new.

 

Thanks again

Announcements