Testing EF Lenses On IP Camera

By Ethan Ace, Published on Nov 23, 2016

Glass.

IP cameras are often characterized as using lower quality 'glass' and that those lenses reduce image quality. By contrast, Canon EF lenses are widely viewed as superior and for higher quality photography work.

What if you put an EF lens on an IP camera? Would it make the image quality better?

We tested one of these models, the 70-200mm Canon EF70-200L against a high end Computar varifocal model, the 25-135mm H5Z2518C-MP, both on Axis' high end Q1635.

*******

** ***** ****** *** provide **** ****** ********** over ******* ** ***** lenses ** **** ******, providing ****** ********* ** details *********. *******, **** other **** ***** ****** lenses ****** ** ******** **** ** **** Lenses *** **********, ******* ********* ** long ****** *** ******* compared ** ******* *** levels ***** ******* ***** lengths.

************, ****** ** *********** and ********* ******** *** these (**** ******) ****** may ***** **** **** skilled ***** **** *********** them.

*** ****** **********, *** high *** ******** ***** very **** ***** ******* is ******** ** **** ranges, ** ****** *** be ***** *********** *** not * ***** ******* plus ***** *** ********* cost *** ****.

*******

** **** ******* ****** widely, **** ~$*** *** low **** ***** ****** to ********* ** ******* for **** ***** ****** varifocal ******. *** *******, the **-***** **** **** in **** **** ***** online ******* ***** $*,*** ***, ***** * **** fixed ********** *** ~$***. ******* **** ** typical */** ***** ******, which ********* **** *** ~$150, **** *** **** expensive ****** ~$***-***.

*** ******* **** ** connect ** ****** ** the ** ******* ** this ********* *** ~$**.

Lens *************

** *** *****, ** had ** ****** **** EF ***** ****** ******** properly ** ******** ** mount ****** ***** *** adapter ****. *** **** able ****************** *****.

********** *** ** ***** lens ** *** **** felt ******** *********. **** connecting ** ** **** with ** ***** ******, typically *** **** ****** rotates * ******* **** until ** ***** **** place. **** ****** ** have **** **** **** adapter ***** ******* **** the *******, **** **** tightly **** ** *******, and *** **** *** not **** **** *****. Users ****** **** **** the **** ** ****** when *********.

Test *********

*** ***** ***** ***** the ******* ****** ******, starting ** ~***' *** extending ** ~***' ** an **** *****.

Depth ** ***** ******

**** **** **** ***** long ***** ****** ****** such ** *** ** and ** ****** ****** here,***** ** ***** ********* ****** * *******. For *******, **** ***** optimized *** *** ****** cone ** *** ***** below, **** ******* ****** blurry (~***' *** ******, highlighted ** ***).

******* ** ***** ******, the ****** *** ********* before ****** ***** ******* at **** ******** ** eliminate ********* ***** ********. The ****** ** ** optimal ***** ** **** of *** *********** *****.

Image ***********

*** ** *** **** long ***** ****** ** the ** ****** **** in *** *******, ** started *** ***** *** very **** *****, ~***' (the ***** **** ** the ***** *****).

** **** ********, ******* delivered ** *** ** lens *** ******* *********, with *** ******* ********** blurry ** *** ** mount *****. ***** ********** is ************* ******* ***** both ******, **** *** very ****.

** ~***' (~*****), *********** become **** **********. ******* details *** *******, *** chart ********** ** ********* in *** ** ***** lens.

*******, ** ~***' (~*****), subject ******* *** ***** similar, *** ********** ***** slightly ********* ** *** EF ****.

Calculation **********

******* ** ****** *** generally *** **** ** surveillance ** **** */**-***** cameras, ***** *** ******* options *** *********** ***** of ****, ** **** calculators *** **** ***** lenses **** ** ** do *** ******* ******* for ***** ******* **** as */*" ** */*", nor ** **** ******* for *** ** ** CS *******, ***** **** additional ***** ******.

** *** *****, ** determined ****** ***** ** view ** ********* ********* FOV ***** ** * specific ********. *** *******, setting *** ***** ** lens ** *** ******* focal ****** (*****) ******** in * *' *** (the ***** ** * standard ** ******* *****) at ~**', ***** ******* in * ~*.* ****** horizontal ***** ** ****.

***** **** ****** **** to ****** ** ******* trial *** ***** ** the ***** ** ******** figure *** ***** ** view **** ***** ***** lenses.

Larger **** ****** ******* ******

** ****** *** **** larger ******** **** */** mount ******, *** ********* much ****** ** ******. Because ** ****, ***** must **** **** ***** outdoor ******** **** *** the ****** ******/**** ***********, as **** *** **** for ******* *** ******/** mount **** *******, **** shorter.

************, *** ***** ***** of **** ******** *** not ** ***** ****** to *** *** ****** EF ****' ********. ************* offering ****** ****** ******, such ** ******** **************** ********* ****** ******** such ** *********** *-************** ******* ****** ***** ******* because ** ****.

Comments (24)

Matt Ion

11/23/16 05:47pm

Probably worth noting that the Canon lens you used was their L-series glass, which is the top-of-the-line pro-grade stuff. You can get the EF70-210 F4 (non-L) for about a third the price.

Tamron and Sigma also make competing EF-mount 70-200 or 210, F/2.8 lenses, at about 1/2 to 2/3 the price of the Canon.

Undisclosed #2

In reply to Matt Ion | 11/23/16 06:57pm

Undisclosed #1

IPVMU Certified | 11/23/16 06:47pm

Interesting test!

Do you think that ppf or the distance is the real driver of the declining superiority of the EF lens vs the Computar?

Said another way, were you to use a sensor with the same physical size but twice the resolution and without changing focal lengths, would you expect the performance of the two systems to converge in the same way?

Or might the EF be much better at all three distances in that case, because enough pixels are available to render a lens with better MTF?

Sebastian Kratzke

11/23/16 07:12pm

Interesting!

Does anybody know an EF to the CS mount adapter that can drive the lense's iris electronically?

Undisclosed #3

11/23/16 07:26pm

Great test! I would love to see a similar test using the Axis Q1659 20mp EF mount camera when it releases using high end Canon L lenses along with value options from Canon, Sigma, Tamron, ect.

Undisclosed Manufacturer #4

11/24/16 09:58am

It's great for IPVM to do such a test to clear "myth" of how good a lens is, like super expensive lens used for security camera. There is a common misunderstanding that more expensive and more "professional looking" lens are better, which is not true. but this article has not discussed about that technical detail, so let me share some of my understanding.

Different lens are designed for different image size, which means big EF lens is not necessarily better than small C/CS mount or M12 lens. For example, on 1/2.3 inch 4K sensor, it needs the lens to have > 300LP/mm (Line Pairs per mm on imager) in order to resolve the full resolution, but regular EF lens is designed to have > 100 LP/mm, so we can say most of EF lens may fail for a tiny but high resolution sensor.

However, on some 1/2 inch 1080p sensors, it only needs to have 130LP/mm to resolve all pixels, then high quality EF lens can do better than regular C/CS mount lens. the Canon EF Tele lens usually have very high MTF, and the LP/mm is also higher, which can be helpful to get higher resolution, higher contrast, and less image artifacts.

Conclusion is, only on bigger sensor (like 1/2), and low resolution (1080p and below), EF (Luxury White) lens can have noticeable better performance than regular C/CS mount lens (under 100$).

Undisclosed #1

IPVMU Certified | In reply to Undisclosed Manufacturer #4 | 11/24/16 11:55am

For example, on 1/2.3 inch 4K sensor, it needs the lens to have > 300LP/mm...

Do you think that your standard cctv lens that comes with a 4k camera actually resolves 300lp?

we can say most of EF lens may fail...

You are not taking into account the fact that the MTF is far better for the center glass of a lens. And that's the part of the lens used with small cctv sensors.

Undisclosed Manufacturer #4

In reply to Undisclosed #1 | 11/25/16 01:58am

actually, we tested, and there are a few good C/CS-mount lens that resolve about 300LP/mm, so we are using such lens.

And even at the center of EF lens, the resolution is usually lower 150 LP/mm, it's not designed to be that high. otherwise on a full frame camera, the resolution is "too good" .

Carl Zeiss Prime lens for full frame camera is usually designed to resolve 50Mpixel, which means, it has > 120 LP/mm on a full frame camera. consider the very high MTF at corner, the center of lens has a resolution of about 150 LP/mm JUST.

We did test with Carl Zeiss Prime lens on 1/2.3 inch 4K sensor, and found very good CS-mount lens has better resolution.

small lens with world-class design usually has ultra high LP/mm, like the lens used on iPhone7, and also used by Gopro Hero5.

Undisclosed #1

IPVMU Certified | In reply to Undisclosed Manufacturer #4 | 11/25/16 02:45am

and there are a few good C/CS-mount lens that resolve about 300LP/mm, so we are using such lens.

Which ones, Fuji's?

Its hard to get real detailed info about cctv lenses for some reason.

Undisclosed Manufacturer #4

11/25/16 02:58am

No. Evetar and Sunex have such super resolution CS-mount prime lens.

Undisclosed #1

IPVMU Certified | In reply to Undisclosed Manufacturer #4 | 11/25/16 05:01am

Yes, but the lens in this test is a Computar, what do you suppose the lp/mm is there?

Also in this particular test the sensor is 1/2" and 1080p, 3.3 micron pixel size which would mean that a lens with 150 lp/mm would outresolve the sensor. (Check my math, please)

Do Evetar / Sunex provide MTF charts of any kind?

Undisclosed #1

IPVMU Certified | 11/25/16 06:22pm

@Ethan, is there any interest in trying this test on a 4k or higher resolution camera?

In case the one or both of the lenses are resolving to greater detail than the sensors pixel pitch can discriminate.

Undisclosed Integrator #5

11/28/16 01:50pm

I would say this test was... eye opening.

Nick Ames

02/27/17 11:13pm

Since the 70-200 Canon L lens is an autofocus lens, but this feature not used for surveillance test, a premium price goes along with the lens. Testing older fixed focal length SLR lenses from the 1960's for example may be a lower cost alternative. a Canon manual focus 85mm f/1.4, higher power, or other brands may be considered. I would suspect noticeable differences comparing the Canon f2.8 used in test to a manual focus f1.4 or f1.2.

Ethan Ace

IPVM | In reply to Nick Ames | 02/27/17 11:21pm

The 70-200mm Canon lens came from our Avigilon Pro 4K Camera Test and those cameras do use the autofocus features of the lens. They don't drive focus on all lenses, but on the ones we tested, they definitely do.

Undisclosed #6

03/02/17 03:23am

OK, so I'm glad someone else bumped this thread so I didn't have to be the necro-bumper, but either way, I have a question.

Per the thread that I posted awhile back, I have obtained a Bosch NBN-80052-BA 5MP box camera, the same Fotodiox CS>EF mount adapter used in this test, and a Canon 55-250mm zoom lens (specifically THIS one) that I just happened to already have for testing. I've been testing it over the past few weeks, and I've run into some issues that I'm hoping someone can point me in the right direction on. I'm admittedly green on this stuff so I'm sure there's something simple or straightforward that I'm missing, but I'm about to lose my mind over whether this is my issue or something being imposed by the camera/adapter/lens.

To make a long story short, I can only get the camera to focus within a fairly narrow zoom range. Regardless of what settings I change on the camera or anything else, actually changing the focus in the camera settings itself, there are many areas of the zoom (probably 70-80% of the camera) that I can never get to focus. I am able to zoom the camera to roughly 100mm and get it to focus, and then I can zoom it up to about 180mm or so and get it to focus there. But, anywhere outside of those ranges, it simply doesn't work. It's like the camera doesn't possess the ability to "focus" outside of those ranges, and the focus ring on the lens doesn't nothing whatsoever.

Is there something I can change in the camera to help this work? Anybody have any idea what would be causing this, or how I can fix it? I have some suspicions, but I'd be curious what everyone else here thinks, because this level of detail on the "why" a lens will/won't work is above my level of expertise.

For the record, the most frustrating part by far is that part of the range that I CAN get it to work will deliver an absolutely beautiful image (>40 ppf) at over 1100 feet away, with a depth of field of over 2100 feet. If I can figure out how to get the thing to focus, I'd have an absolute winner on my hands...

Ethan Ace

IPVM | In reply to Undisclosed #6 | 03/02/17 04:16am

Where did the lens come from? I'm not a photography expert by any means but apparently Canon does some funky things with their lenses and it may need to be "reset" if it came from a camera kit.

So since you called yourself green, I won't feel bad for pointing out the most common rookie mistake: is the lens set to MF (manual focus)? There's a slider switch at the base marked AF and MF. On some lenses, it disengages the rings and engages the autofocus motor in AF mode, so manual focus won't work.

Another basic question, are you resetting the Bosch camera's backfocus, then zooming, then autofocusing? Sometimes it hits a limit of how far it can focus, and on long focal lengths like this it's certainly possible.

Undisclosed #6

In reply to Ethan Ace | 03/02/17 04:52am

First, thank you. I appreciate the speedy (and detailed) reply.

Yes, I am setting to MF. I made that mistake once when switching back and forth, but it's been in MF for awhile.

I have tried every combination of backfocus, zoom, autofocus. zoom, backfocus, autofocus. etc that you can think of. No chance, no difference at all really.

I will try that reset function you mention and see what happens. However, I'm not sure that it would make a difference, because I have to power down the camera+lens to put it back on the Bosch camera, so it appears like it would go back into sleep mode either way? Could be an issue with this specific lens, perhaps?

Ethan Ace

IPVM | In reply to Undisclosed #6 | 03/02/17 04:17am

Also I'd be really interested to see the images you're getting at that range, if you're willing to share.

Undisclosed #6

In reply to Ethan Ace | 03/02/17 04:52am

I will take some this weekend and share.

Matt Ion

In reply to Undisclosed #6 | 03/02/17 06:17am

Off the top of my head, it sounds like the lens isn't sitting exactly on the same plane as the sensor - maybe the adapter isn't fully seated, or is cross-threaded and sitting crooked, or something similar is creating a "tilt-and-shift" effect.

That's also a fairly heavy lens, at nearly 14oz, and is probably a little front-heavy, so if its whole weight is being supported by the mount, it may be bending the mount, or even the front of the camera at the CS threads. That, too, could introduce the T-S effect.

As others have noted, seeing images from it would probably help diagnose the problem.

Undisclosed #6

In reply to Matt Ion | 03/02/17 03:52pm

Matt, thank you for your input. Much appreciated.

I just went and checked, and the lens does "sag" off the front of the camera by maybe 1/16" of an inch of movement. However, I checked the standard Bosch lens (12-50) that I got with the camera, and it does the same thing, so unless I've literally broken or weakened the mount on the front of the camera, I don't know that this is a major is issue. I can try to test for this later today, but with as much as i have fiddled with this thing, I don't feel like this is a likely culprit. I could, of course, be wrong.

For the images I'm getting when out of focus, they are literally so bad that it's not worth sharing. There is nothing recognizable in the image, and no changes to the camera will improve it whatsoever. I did just have a thought that perhaps the CS-mount adapter that I'm using is actually "jamming up" the movement of the sensor on the camera beyond a certain point, but I checked and I don't believe that is it either.

I'm starting to believe that maybe Ethan's point above about the lens going to "sleep" could be part of it. There is no other explanation for why the focus ring wouldn't work at all when it's off the camera as far as I can tell...but I'm also not sure if the reset will fix it. Will have to test later and find out....

Matt Ion

In reply to Undisclosed #6 | 03/02/17 06:23pm

Remember that the "S" in CS stands for "short", and a CS mount holds the lens a lot closer to the sensor than a C-mount, so it's possible at some focal lengths, the back element of the lens is actually bumping up against the sensor, or is simply too close to it to focus properly.

Canon's lineup actually has a similar dichotomy: the newer EF-S series lenses sit a lot closer to the sensor than the EF series, and are designed specifically for the crop sensors in their consumer and prosumer cameras. EF-S lenses have an additional tab on them that prevents them being mounted to EF cameras (EF lenses work fine on EF-S cameras), but that tab can be cut off, allowing the EF-S lenses to be used on EF mounts. There are benefits to this, particularly allowing the use of ultra-wide-angle lenses like the EF-S 10-22mm to be used on full-frame cameras, but it's a tricky proposition, because the back element CAN interfere with an SLR's pellicle mirror, and vignetting can be particularly bad with wide-angle lenses.

So putting these factors together, we may be on to something: if the adapter is designed for EF lenses, an EF-S may simply be sitting too close to the sensor to focus properly. You could see if there's an EF-to-C adapter available, or add a C-to-CS adapter to the stack to hold the lens a little further from the sensor.

Undisclosed #6

03/03/17 01:01am

Final answer?

Ethan Ace is a golden god.

Well, that's at least PART of the final answer.

The lens does indeed power down and make the trim ring nonfunctional. The lens only allows focus via the trim ring when the camera is powered.

However, this still doesn't full solve the issue of the camera failing to focus, sometimes grossly so. No amount of working that trim ring is going to make it focus I don't think, so it would just be for finer adjustments once the camera did the heavy lifting.

So, my next step is to follow your advice and work on getting a CS>C>EF adapter to see if extending the lens away from the camera body will improve my ability to focus any. I've got a few things i'm going to test as well just to see what happens, but if anyone else has any thoughts or ideas, please share them.

Read this IPVM report for free.

This article is part of IPVM's 6,587 reports, 888 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.