canon 24-70 L f2.8

[Felix K.]

Thank you, Jeff!

 

[Miguel "Macgregor" De Olaso]

You're assuming Miguel that everyone shoots lenses wide open for the genres they work on. That's simply not the case. Sure narrative guys regularly seek wide open apertures. But those who work in a lot of other styles and genres shoot a lot of mid-apertures.

Nononono, no talking about shooting wide open, i´m talking about a huge margin of error between Fstops and Tstops in this kind of lenses. That you might think, bah, it´s f2.8, that´s not so far from something fast like a 1.8 or f2. But when you go out and shoot you find your lens performing more like a f4.5 and that´s wide open, and since we don´t want to shoot wide open for better optics performance, we might consider closing the iris to a sweet spot, like f5.6 then it´s allready too dark if you´re trying to shoot a bautiful sunset. And then i think, uhmm, i spent nearly $2k in this lens when an ultra cheap 50mm f1.8 would increase the amount of light that hits the sensor by at least twice as much at the same given fstop...

So not saying it´s bad, just saying it´s too expensive and heavy for my point of view and it´s not even superwide or super tele, so you must carry other lenses. Then if you´re carring other lenses like the 14-24 why did you spend that much money in this midrange zoom? ;D

Just saying! and thanks for the comments!

 

[Steve Gibby]

Miguel,

I use the Canon 24-70 on my 5D bodies, not on RED One. I've been way happy with the lens on the 5D, as I have with the Canon 16-35 f2.8 MKII, and the Canon 70-200 f2.8L IS. As you might guess, I'm well aware of the difference between f stop and T stop. I'm not disputing that the 24-70 may have a higher T stop than 2.8 - what I am saying is that it has been a great lens for me and that nuance hasn't hindered my crew's shooting at all. If you're generalizing by saying that all 35mm stills zoom's T stop are significantly above their F stop, I disagree. Each zoom lens is different - and each individual lens within a lens make and model can be different. I don't use any of those Canon lenses on RED One, but I will test them out on Epic and Scarlet. I anticipate being happy with them on Epic and Scarlet, but that will remain to be seen.

I use 35mm stills lenses exclusively on my multiple RED One cameras - no cine lenses whatsoever. I own 28 Nikon and 2 long Canon lenses I use on my REDs. They are all hand picked, and I've been happy with each of them for the genres me and my crews shoot (listed a few posts ago). Keep in mind as I posted before, the genres I shoot are usually between f4 and f11, we eye focus, and expose via zebras and histogram - thus T stops aren't critical as when someone shoots cine-style. We obviously know how to get the DOF we want/need. We do no narrative or cine-style shooting with RED One - by choice. That said, I was also trained in cine-style shooting earlier in my career - I just like hybrid non-hardlined EFP style much better now.

All my Nikon lenses are stock - no gears, no conversions - I like it that way. The Nikon lenses me and my crews use the most on my RED cameras are:

14mm f2.8, Tokina 11-16 f2.8 with Nikon mount, 14-24 f2.8, 17-35 f2.8, 28-70 f2.8, 80-200 f2.8 AF-S, 80-200 f2.8 AF, 180 f2.8 AIS, 50-300 f/4.5 AIS, 400 f3.5 AIS

The Canon lenses we use are:

FD 150-600 T5.6L with Century PL mount conversion and de-clicked iris.
FD 800 T5.6L with Century PL mount conversion and de-clicked iris.

I own a lot more Nikon primes I didn't list, some AF-S and some AIS. The genres we shoot usually call for zooms, which we use as variable primes.

The results of our mobile hybrid shooting equipment and techniques are very cinematic - but as you can see we arrive at those results via very different equipment and techniques than cine-style shooters.

Once I get Epic and Scarlet, with all the mounts (PL, RED, Nikon, Canon) I'll test out all of the above lenses on those cameras and draw some conclusions on which lenses work best for what we shoot. I expect the new RED Electronic Lenses to be excellent. If they test out real well, I will most likely sell off almost all of my Nikon and Canon lenses (except the 150-600 and 800), and use the RE lenses exclusively.

To each their own...choose as you wish/need. The above setups and lenses have worked very well for me and my crews - for the genres and hybrid non-hardlined EFP style we shoot.

Fun times ahead though...:thumbup:

 

[Kim Frank]

Miguel... That was really enlightning.
I didn't know there was a difference between F and T stop and
that the rating of lenses does vary.
I own this lense with my 5D and in fact often was wondering,
that I had to give it an extra stop of light, to get the exposure I wanted.
What lenses would you recommend.
Fixed Canons or do you prefere different ones?
Also is there a better zoom you would reccomend?
Regards
KIM

 

[david_winters]

The problem with the 24-70 or the 70-200 is that the Tnumber might be closer to T4.5 than you think, while a prime lense like the 50 f1.4 could be a T1.6 (

I'm seeing the zoom as only 1/3 stop light loss vs the Canon primes. Not 1+ stop. The lens does get noticeably darker on the 70mm side, which I haven't compared to the 85mm L prime.

Here is a really quick test using daylight (clear sky today), 5DII, monopod, RAW and Canon Digital Photo Professorial app. The 5DII exposure meter indicates a -1/3 stop on the zoom vs prime, all lens set to f2.8

Canon 24-70 vs Canon EF 35mm & f1.4L Canon EF 50mm f1.4.

 

[Miguel "Macgregor" De Olaso]

Excellent David. That´s exactly what DPreview and others should do. Thanks so much for this.
One question: Did you have the lens compensation on (whatever it´s called in the 5D menu)? THe camera could be trying to compensate the optics.

Where are you reading a -1/3 of a stop? I can see at least -1 stop by looking at the histogram at 35mm and about half a stop at 50mm, but it´s hard to say. Probably focusing only on the grey card could be a better measuring tool together with some lightmeter measurements.

Still if we´re getting one third of a stop at 50mm sounds great to me, isn´t it? Not bad at all.
Could you try the 70-200 too if you have it?

Edit: also try using a rounded "native" ISO like 100, 200 or 400 for better results. thanks.

 

[david_winters]

Did you have the lens compensation on (whatever it´s called in the 1D menu)? THe camera could be trying to compensate the optics.

Yes, I considered this when I shot the frames. In the samples above the only "lens aberration correction" feature enabled is "peripheral correction" which corrects for lens falloff around the parameter of frame, but not the center. I'll post new images with this mode disabled (I shot RAW). Got distracted and forgot to disable.

Where are you reading a -1/3 of a stop? I can see about 1 stop by looking at the histogram, but it´s hard to say. Probably focusing only on the grey card could be a better measuring tool together with some lightmeter measurements.

Still if we´re getting one third of a stop at 50mm sounds great to me, isn´t it? Not bad at all.

I set exposure and shutter with the fist lens up, the 35mm prime, f2.8 1/160 sec 640 ISO. Then I changed to the 24-70mm f2.8, same camera position. The camera exposure meter indicates -1/3 stop vs the 35mm prime.

Grey card would be ideal, I'm at home and all my gear is at the office. I considered shooting a white sheet of copier paper too, but I was lazy.

I also shot the primes at 4.0 for comparison, which I will post later.

I don't have a 70-200 or 85mm to test, wish I did. Anyone in O.C. CA that wants to meet up? This is very relevant to my work as I shoot a lot of Canon glass on R1.

Edit: also try using a rounded "native" ISO like 100, 200 or 400 for better results. thanks.

My understanding, based on 5DII tests I saw on a forum, which I can't find now... ISO 160, 320, 640 produce the lowest noise images on the 5DII which I assume is bceause these are the native "analog gain" sensor settings, and the in-between ISO have an additional layer of digital gain. But I don't know if any of this is accurate.

I'll post updated images in a few hours.

Thanks

 

[Miguel "Macgregor" De Olaso]

Hi David, thanks again for the time you took to shoot this.

"The camera exposure meter indicates -1/3 stop vs the 35mm prime." I understand tht but if you check the Canon histogram that you provided with stops marks you´ll see it is way less than that (about a full stop at 35mm and almost half a stop at 50mm). Are those marks accurate? The only way of knowing that is by using a lightmeter and grey cards.

It´s true that at ISO160, 320 and so on you´re getting less noise, but that´s because the camera appears to be using ISO200 (for when setting 160) and using the software to lower the exposure, therefore lowering the noise floor (good), but also lowering the dynamic range (bad). It´s basically playing within the raw linear image limits (I call that "cheating") somehow similar to enabling highlight recovery. It can be ok for quick Jpgs, but i´d rather have the best dynamic range and adjust the noise floor myself without clipping highlights.

Again, thanks a lot.

 

[Sanjin Jukic]

Sometimes it happens that you have to choose the right lens between batches of the same but if you are lucky enough that you can even choose and make your own choice.

Canon EF 24-70mm f/2.8L USM performed pretty good at the latest Canon EOS Movie about Canon EOS-1D MkIV a brand new Canon HDSLR camera>>>

Shot with Canon EF 24-70mm f/2.8L USM on Canon EOS-1D MkIV, wide open @ f/2.8, ISO 1250

Shot with Canon EF 24-70mm f/2.8L USM on Canon EOS-1D MkIV, wide open @ f/2.8, ISO 1250

The footage link>>>

 

[david_winters]

Waiting for the haze to burn off so I can shoot a few more tests. ISO100?

 

[david_winters]

Ok, I bailed on the 5D tests, went out for lunch and came home with a R1 w Birger mount.

Test setup with 18% grey card, one 4' tube kino daylight. Canon 24-70mm 2.8L, Canon 35mm 1.4L, EF 50mm 1.4. R1 set to 4k HD, RC28, 5600, ISO320, 1/48th.

Overhead lighting is on for this ref photo, off for tests. The only light burning for the tests is the 4' Kino.

Incident meter reading at center of card, dome aimed at light source.
FC = 97

Spot reading at center of card.
FL = 14

R3D clips saved as TIFF half (1920x1080), REC709, REC709. Image slices pulled from the vertical center of each TIFF, full frame width - to show falloff. Each slice is the exact same pixel area, via a mask in photoshop. JPG displayed here is scaled 50%.

To me it still looks like the zoom is only approx 1/3 stop slower vs the primes. No where close to 1 full stop slower. Compare the prime f4 vs zoom f2.8 at various focal lengths. The zoom at f2.8 is much closer to the primes at f2.8 not f4.

Not sure why the zoom appears slightly darker at 50mm vs 35, 60, 70. I double checked all my notes, slates and layers. This is in fact the correct layer for 50mm and stop.

 

[Joe Taylor]

Helllo Stephen Strangways

You posted what might be the ultimate "tip of the week." Your idea has solved all sorts of problems that allows me to once again use my MB-16 and toss my clip-on LMB 5 to the side where it belongs. Cheers!

 

[Miguel "Macgregor" De Olaso]

Wow, i didn´t expect you to take the effort of arranging this setup!!!!!

Indeed, very good results. Your lens seems to perform very good. Thanks for the supertest.

It doesn´t give the same results that i have experienced with the 24-70 and 70-200 but you did your work well so there is not margin for error. As you say, the gap is between 1/3 of a stop and half a stop. Not a huge difference.

Now if you had more lenses... ;D


Edit: i played with the bars and i found that the 35mm at f2.8 is lighter than the 50mm at f2.8!!! Also as you get away from the center all bars get darker.

 

[Steve Murray]

Zoom problem solved

Zoom problem solved

Excellent lens IF you get a good one. You should be aware that these lenses seem to vary greatly from one to another. It took 5 for me to find one that was sharp at F/2.8 the first time and 3 for the 2nd one I bought. Some were really soft. Yes, they can be adjusted by Canon but I prefer to start with a sharp one out of the box.

Also see this post for a solution to problem with the lens zooming in and out: http://www.reduser.net/forum/showthread.php?p=507414#post507414

 

[Kim Frank]

Excellent lens IF you get a good one. You should be aware that these lenses seem to vary greatly from one to another. It took 5 for me to find one that was sharp at F/2.8

How did you make that? Did you test them at the store to compare?
And how does it work to let it adjust by canon? Is that expensive?

David thanks for the great Test!

 

[Steve Gibby]

Thanks for the tests David...your tests have confirmed that there is just a slight increase in T stop over F stop with the 24-70 versus primes - at least with your lens. My 24-70 performs similar to yours. If I'd have had the time I would have done the tests myself.

Before David did his tests, I posted this in my post #23 of this thread, I stand by it, and David's tests support it:

"I'm well aware of the difference between f stop and T stop. I'm not disputing that the 24-70 may have a higher T stop than 2.8 - what I am saying is that it has been a great lens for me and that nuance hasn't hindered my crew's shooting at all. If you're generalizing by saying that all 35mm stills zoom's T stop are significantly above their F stop, I disagree. Each zoom lens is different - and each individual lens within a lens make and model can be different."

Steve Murray then touches on the last point in my post - that individual lenses within a lens model can vary somewhat in performance. That's why all 30 of the 35mm still lenses I use on my RED cameras have been hand picked for their performance. I tested each lens on my REDs, and returned a few until I got good ones. Overall, only a few of my 35mm stills zooms showed unacceptable performance. But most of them did fine right out of the box. If they were used lenses, I simply sold them off, and bought another version of the lens. I had to do this with just a couple lenses out of the 30 I bought. It is obviously possible to get a bad 35mm still zoom - but in my experience that is a relatively small percentage of the lenses.

 

[david_winters]

Wow, i didn´t expect you to take the effort of arranging this setup!!!!!

Indeed, very good results. Your lens seems to perform very good. Thanks for the supertest.

It doesn´t give the same results that i have experienced with the 24-70 and 70-200 but you did your work well so there is not margin for error. As you say, the gap is between 1/3 of a stop and half a stop. Not a huge difference.

Now if you had more lenses... ;D


Edit: i played with the bars and i found that the 35mm at f2.8 is lighter than the 50mm at f2.8!!! Also as you get away from the center all bars get darker.

Thanks for the thanks! Yeah, I didn't plan to clock most of the afternoon on it, but once I got into it I wanted to do it right.

The falloff is there, but essentially consistent with all the other decent glass I've worked with in PL and B4 for the past 10 years. These three lenses falloff a lot more in FF, but the Canon DSLRs do a such a decent job of correcting, or not, that is doesn't matter.

 

[david_winters]

Excellent lens IF you get a good one. You should be aware that these lenses seem to vary greatly from one to another. It took 5 for me to find one that was sharp at F/2.8 the first time and 3 for the 2nd one I bought. Some were really soft. Yes, they can be adjusted by Canon but I prefer to start with a sharp one out of the box.
[/url]

The Canon Service Center in Irvine CA can calibrate Canon EF lenes. At least my friends at that office tell me they can, I haven't personally had glass serviced there, yet.

Several pro photographers have told me they send in all their new lenses for a calibration check, when purchased new. They also send their lenses back for recal about once per year. Canon has some type of annual service membership thing which covers all of this for a flat annual fee. Even at pay per drop off it's not that expensive relative to the cost of the gear.

I personally don't have the time or desire to hassle a retailer to test 3-5 copies of the same lens. I'd much rather spend a small bit of cash to have a tech zero it out for me. With a new lens, if it's out of spec, I'm confident Canon service would make it right no charge. But, I drive by the service center every day on my way home.

Excellent lens IF
Also see this post for a solution to problem with the lens zooming in and out: http://www.reduser.net/forum/showthread.php?p=507414#post507414

Just leave the lens hood attached, as intended. With the 24-70 the hood also protects the lens components which expand. It doesn't take much force to eff up the zoom mechanism and focal plane on this lens. I've worked with several rental copies which have slop and the focal plane shifts because of users leaving the hood off and the front half of the lens bumping into things.

No mattebox coverage issues with the hood attached plus 2 panavision size filters plus 138mm round. Here is a photo with prompter.

 

[david_winters]

Here is the 16bit 1920 wide PSD file, if anyone wants to re-sort and play around. The layers are slices from the vertical center of 16 bit TIFF files exported from RA as half size (1920x1080).

20MB
http://office.innovatemedia.com/winters/tests/24-70T-test.zip

 

[david_winters]

Edit: i played with the bars and i found that the 35mm at f2.8 is lighter than the 50mm at f2.8!!! Also as you get away from the center all bars get darker.

Interesting. The 50mm tested is the "non L" version. The front element is significantly smaller than the L 50mm ver which is 1/2 stop faster, which I'm sure your aware of. Yet another aspect to compare, but I'm done.

Goodnight.