Sony 24-70 F4 Lens

[Guy Mancuso]

Seems to be too many optical challenges for this lens. Sharp here, not so sharp there....distortion. It's a lens that I would really love to have used, given the convenience factor, but the 35 (at least my copy and the 55) will suffice for the range of what I need...alas, hope my mind will be change, but I'm not seeing a resounding thumbs up here.

Tim, I may have missed this, but it seemed early on that you really dug the lens, but now it appears that you returned it....not as impressed?

Part of this Ashwin is expectations folks have which I totally understand. People see this lens as a catch all in focal lengths. For me too but as usual a tough lens to build at all the focal lengths.

 

[fotografz]

FE 24-70/4 arrived yesterday.

Am comparing it to the ZA24-70/2.8 and ZA24/2.

It is 1/2 the size of the ZA24-70/2.8, and much lighter. The ZA24/2 on the LAEA-4 adapter is not much bigger, but is heavier.

Compares very favorably in terms of color/contrast. So all three lenses produce synergistic files rather than different signatures.

Native distortion is well corrected with the ZA24/2 but there is no profile for it yet either, so I do it manually. The ZA 24-70/2.8 is next best but requires more software correction than the ZA24/2 … and the FE24-70/4 is by far the worse of the three. I haven't seen distortion like this since some old Canon wides.

I do not know how well a custom profile will work to correct distortion on the FE24-70/4, nor how much of the image will be lost doing so.

My copy seems to have equal slight softness at the L and R edges at 24mm, but not a deal breaker.

50mm looked very good, and 70mm also looked good … quite crisp in center, slight but equal fall off of sharpness on edges.

In practical use, it looks to be an all around casual use lens for vacations and walkabouts.

Depending on how well the profile works, I'm not sure I'd depend on this lens for critical work doing architectural type images or anything where distortion is an obvious detraction. If it requires massive software correction, then we'll be losing some of those pixels we paid for.

I saw it as a general use lens and that seems to be exactly what it is. My M21 is a much better corrected lens, as is the ZA24/2 … so I opt for those for anything of a bit more critical nature.

Jury is still out on whether I'll keep it. At least is seems reasonably balanced side to side and 70mm looks good with nice bokeh when you can get distance between the subject and background.

- Marc

 

[mark1958]

My lens arrived today. Finally started raining in CA…. so could only do a little bit of testing. I have to agree-- I have never seen such distortion.. Corrected in the in camera jpgs

FE 24-70/4 arrived yesterday.

Am comparing it to the ZA24-70/2.8 and ZA24/2.

It is 1/2 the size of the ZA24-70/2.8, and much lighter. The ZA24/2 on the LAEA-4 adapter is not much bigger, but is heavier.

Compares very favorably in terms of color/contrast. So all three lenses produce synergistic files rather than different signatures.

Native distortion is well corrected with the ZA24/2 but there is no profile for it yet either, so I do it manually. The ZA 24-70/2.8 is next best but requires more software correction than the ZA24/2 … and the FE24-70/4 is by far the worse of the three. I haven't seen distortion like this since some old Canon wides.

I do not know how well a custom profile will work to correct distortion on the FE24-70/4, nor how much of the image will be lost doing so.

My copy seems to have equal slight softness at the L and R edges at 24mm, but not a deal breaker.

50mm looked very good, and 70mm also looked good … quite crisp in center, slight but equal fall off of sharpness on edges.

In practical use, it looks to be an all around casual use lens for vacations and walkabouts.

Depending on how well the profile works, I'm not sure I'd depend on this lens for critical work doing architectural type images or anything where distortion is an obvious detraction. If it requires massive software correction, then we'll be losing some of those pixels we paid for.

I saw it as a general use lens and that seems to be exactly what it is. My M21 is a much better corrected lens, as is the ZA24/2 … so I opt for those for anything of a bit more critical nature.

Jury is still out on whether I'll keep it. At least is seems reasonable balanced side to side and 70mm looks good with nice bokeh when you can get distance between the subject and background.

- Marc

 

[tashley]

Seems to be too many optical challenges for this lens. Sharp here, not so sharp there....distortion. It's a lens that I would really love to have used, given the convenience factor, but the 35 (at least my copy and the 55) will suffice for the range of what I need...alas, hope my mind will be change, but I'm not seeing a resounding thumbs up here.

Tim, I may have missed this, but it seemed early on that you really dug the lens, but now it appears that you returned it....not as impressed?

In a way, the opposite is true: as a fully paid up realist, I know that designing a lens like this is going to involve compromises so I can live with the distortion and the fact that some parts of the range will inevitably be less impressive than others. I've said before and I'll say again that I think that overall, this is the most useful mid range zoom for full frame that I have yet tried, and I have shown how, with a bit of attention to shooting and processing, I can make it hard to distinguish a file from this lens from a 55mm f1.8 file.

So I am really impressed. I just want to make sure I get a copy that has been put together really well. Out of two copies so far, they both had different weaknesses, and that is the story that's emerging from other folk here.

I will settle for a reasonable copy, rather than expecting to find a perfect one. My criteria is that by f5.6 it should not be obviously weaker on one side at any focal length and that there should be at least a central range where by f8 it appears reasonably clean at both edges.

 

[fotografz]

Understood - I don't use ACR, but rather LR. I'm lazy.

ACR is the same whether used from Photoshop Bridge or Photoshop Lightroom.

I guess I'm going to have to download the new ACR release candidate today and test the lens at various focal lengths to see how well it corrects the distortion, and at what cost in image area.

From what I can tell so far, the ZA 24/2 looses the least image area, and the FE24-70/4 loses the most.

However, (and I have to confirm this):

When I had the camera locked down on a camera stand, and aimed at a tile floor, then just changed lenses … the 24 end of the FE24-70 seemed to capture more area than the ZA 24/2. Don't know if that is because the adapter moves the lens forward or if the ZA24/2 isn't actually 24mm … all I know is that the film plane never moved.

- Marc

 

[tashley]

Some thoughts:
1) I posted here a link to a profile you can install as Custom in LR and which works well.
2) The distortion does not show in, and therefore does not need correcting for, quite a wide range of subject matters, strong though it is. Placing horizon lines on vertical half way mark helps, where possible.
3) I posted uncorrected distortion examples at all focal lengths here (#135 & 137)
4) If you compare corrected and uncorrected RAW files, you will see that resolution and micro contrast takes quite a hit in some parts of the frame at the worst affected focal lengths.

 

[fotografz]

I'm on the same page Tim.

1) Don't correct the distortion if you don't have to. For me, a vast majority of the editorial sort of candid images won't need it. Church pic below needed it, tree-trimmers and tree shots didn't.

2) Be aware when you are shooting images that you know will need distortion correction, and shoot wider than normal to give yourself enough image to "correct".

Here are some test pics I did yesterday. Forgive the boring subjects, but it was -7º F outside … plus I promptly stepped off into a deep drainage ditch hidden by snow, and fell in … took my friend Jim Bulin and one of the tree-cutter guys to get me out Quickly took the snaps and made for the car immediately. Brrrrr.

Distortion test: Interior shots are totally untouched … just exported them as is.

ZA 24/2 is clearly the winner here. Less distortion and note the B&W framed image top right compared to the other shots.

ZA 24-70/2.8 isn't too bad, but for me is far to big on the A7R to be anything other than a back up if needed.

FE 24-70/4 is the worse. However, note that it provides more image area that the other two images … keeping in mind that the camera was locked down on a stand and never moved … all I did was change lenses. Makes me wonder if the LAEA-4 is slightly cropping the 24mm? Or is it that the adapter is moving the ZA lenses forward compared to the FE 24-70? Whatever, it is what it is.

Outdoors;

Church pic shows "as shot" on top, with finished corrections below. I used the Sigma 24-70 profile and then manually tweaked from there … "some hit to edge resolution", but not too bad.

Tree and Gazebo image were at 70mm … the little plaque over the Gazebo entrance is completely readable at 100%

Tree trimmers shot @ 45mm is the sort of image that doesn't need profile correction. I think it shows the remarkable DR of the A7R in such high contrast 11 AM winter lighting.

BTW, the church wide shot, then 45mm of workers, then 70mm of the tree and Gazebo illustrates the usefulness of a mid-range zoom … who the hell wants to swap lenses three times in sub zero weather? Or on a warm sandy beach with island breezes kicking up fine sand spray … where I wish I were now thank you very much. :toocool:

- Marc

 

[Guy Mancuso]

In regards to image area. Long story short my Leica 19 r with modified hood worked fine on my D800 e camera with no vignetting. Same lens and hood will not work on the Sony A7r without vignetting. Bottom line it seems the Sonys image area is bigger or using more of the sensor size than Nikon. I have put the hood away in its storage bin as I can't modify it any more without grinding metal away. I have a custom made solution but I was forced to do that.

 

[Guy Mancuso]

Just checking Photozone tests for distortion

ZA 24mm F2 is 1.7 percent barrel distortion

ZA. 24-70 at 24mm is 2.6 barrel distortion

My new Zeiss 25mm F2 for Canon is 1.8 percent barrel

Nikon 24mm 1.4 G is 1.5 percent barrel

Interestingly the Nikon 14-24 at 14mm is 4 percent which this FE 24-70 at 24mm seems to be .

I would try correcting this with that 14-24 profile at 14mm if you can

Seriously this in my mind is excessive barrel for a 24-70 zoom.
Honestly this is exactly why I bought the Zeiss 25 or even the ZA 24 was to avoid this zoom at 24mm.

I call it unusable at 24mm maybe better said is not preferred at 24mm. After that it seems okay. YMMV as some folks won't care at all and some picky *** people like me do. Lol

 

[tashley]

Bingo! I got a good one

My third attempt and I have one that has no notable asymmetries and seems to be totally useable, believe it or not, at all focal lengths and all apertures, if you judge by the appearance on a Retina screen. On a traditional 100dpi monitor it looks less pretty at 100%, quite acceptable at 50%, but that little extra pixel density of a retina screen at 220dpi makes them all look at the worst OK and at the best really very good. 220DPi will equate to a 33" print. That is really not bad!

I am uploading an aperture series as we speak. Mine is really quite good at 24mm, its less impressive range is 28-35 but still OK, 50 is excellent (for a zoom) and 70 is just fine if you stop down.

This lens will not even leave my cold, dead hands. I will be buried with it. The holy grail has arrived: a mid range zoom which can make acceptable thru very good prints at exhibition sizes at every focal length, and mounts on a camera that, with a change of lens, becomes effectively medium format. Yay!

BTW on the FOV issues, maybe Sony gave its nominal FOVs to the post-correction (i.e. distortion corrected JPEG) file, so you really get a true 24mm after correction?

Here's a link to a gallery with my usual harbour side aperture series. Shot RAW, OSS off on a tripod, I shot two series and as usual the AF series was overall better than the MF series so that is what I have posted. Files were developed in LR with no lens corrections, with Camera Standard profile, with sharpening 60/0.7/70/20 (it can take up to 80 as the main setting quite well too) and with no clarity adjustment though giving it +12 and 80 sharpening makes the files look pretty nice at Retina resolution...

 

[Deleted member 7792]

Re: Bingo! I got a good one

This lens will not even leave my cold, dead hands. I will be buried with it. The holy grail has arrived: a mid range zoom which can make acceptable thru very good prints at exhibition sizes at every focal length, and mounts on a camera that, with a change of lens, becomes effectively medium format. Yay!

Congrats! My thoughts exactly. :thumbs:

BTW on the FOV issues, maybe Sony gave its nominal FOVs to the post-correction (i.e. distortion corrected JPEG) file, so you really get a true 24mm after correction?

Have you tried the ACR Release Candidate 8.4 Beta with either Photoshop CC or CS6? The lens profile for the FE 24-70mm is in that release and seems to work well, IMHO.

Joe

 

[dwood]

Re: Bingo! I got a good one

My third attempt and I have one that has no notable asymmetries and seems to be totally useable, believe it or not, at all focal lengths and all apertures, if you judge by the appearance on a Retina screen. On a traditional 100dpi monitor it looks less pretty at 100%, quite acceptable at 50%, but that little extra pixel density of a retina screen at 220dpi makes them all look at the worst OK and at the best really very good. 220DPi will equate to a 33" print. That is really not bad!

I am uploading an aperture series as we speak. Mine is really quite good at 24mm, its less impressive range is 28-35 but still OK, 50 is excellent (for a zoom) and 70 is just fine if you stop down.

This lens will not even leave my cold, dead hands. I will be buried with it. The holy grail has arrived: a mid range zoom which can make acceptable thru very good prints at exhibition sizes at every focal length, and mounts on a camera that, with a change of lens, becomes effectively medium format. Yay!

BTW on the FOV issues, maybe Sony gave its nominal FOVs to the post-correction (i.e. distortion corrected JPEG) file, so you really get a true 24mm after correction?

Here's a link to a gallery with my usual harbour side aperture series. Shot RAW, OSS off on a tripod, I shot two series and as usual the AF series was overall better than the MF series so that is what I have posted. Files were developed in LR with no lens corrections, with Camera Standard profile, with sharpening 60/0.7/70/20 (it can take up to 80 as the main setting quite well too) and with no clarity adjustment though giving it +12 and 80 sharpening makes the files look pretty nice at Retina resolution...

Great news that a good copy finally landed at your doorstep! It's such a bummer though, and this is surely not a Sony specific issue, that one has to go through the hassle of multiple iterations of lenses until the proper one shows up. If a lens, or any product, doesn't meet spec at the manufacturing QC stage, it shouldn't be shipped. It makes me think that either the window on what 'meets spec' is way too large, and/or the QC folks aren't doing a good enough job. Kinda sucks for consumers of this not-inexpensive kit.

 

[tashley]

You could look at it like this: current standards (and therefore costs) of manufacturing and QC mean that the price is set where it is: were they to try harder, it would cost more. So if you're one of the ones that ploughs through three copies to get a good one, you're effectively getting a bargain compared to the average lens purchaser (same price, better product on average) and therefore in a sense you're getting paid for your effort!

I had a horrid experience today: lens#3 was from a smaller local dealership and when I took it in for a refund (so immaculate that easily re-sold as new) he wanted evidence. I gave him an SD card and said, 'do you have Lightroom or Photoshop?' and he said 'oh you shoot RAW do you, we can only look at JPEGS". I replied that there was indeed some jpegs on the card too, and he put the card into one of those 5x7 print terminals and thought that by making a print that size, the problem should be evident - and if not, it probably wasn't a problem...

I won, but I had to use the L word (legal, not lens) and I felt sorry for the guy. If Sony QC don't catch it, how is he going to? He simply doesn't have the means and possibly not the experience to tell a dud from a good copy...

 

[Godfrey]

Re: Bingo! I got a good one

... It's such a bummer though, and this is surely not a Sony specific issue, that one has to go through the hassle of multiple iterations of lenses until the proper one shows up. If a lens, or any product, doesn't meet spec at the manufacturing QC stage, it shouldn't be shipped. ...

In any manufacture there is variation.

I remember back to the 1980s when the project I was working on needed Nikkor 105mm lenses for ground truth data collection. They'd get in a dozen lenses and we put them on the optical bench to find the closest match of three, send the rest back.

All 12 were typically good quality meeting the deliverables spec and QC. But all varied somewhat ... we just needed three that were very close in resolution, contrast, etc for our needs.

 

[sandrahmart]

Last I checked B&H they show end of Feb. we're on the pre order list so all we can do is wait.

 

[dwood]

Re: Bingo! I got a good one

In any manufacture there is variation.

I remember back to the 1980s when the project I was working on needed Nikkor 105mm lenses for ground truth data collection. They'd get in a dozen lenses and we put them on the optical bench to find the closest match of three, send the rest back.

All 12 were typically good quality meeting the deliverables spec and QC. But all varied somewhat ... we just needed three that were very close in resolution, contrast, etc for our needs.

There is certainly variation in the manufacturing process. There are a lot of variables. In my day gig, I run sales/mktg. for a company that designs and manufactures pro-audio gear. Our customers can be pretty tweaky when it comes to performance expectations, which they should be. If we had to ship 3, 6, 12, etc. samples of a product until one was deemed 'right', we'd be looking for a new line of work.

It sometimes feels like the products we use to make pictures are more prone to sample variations than other stuff. Could be that we, as customers, expect too much, or that consistent adherence to a product spec. can't be met in the manufacturing process, or folks in final QC are maybe letting things slide, etc. I, for one, don't really enjoy the multiple versions of 'buy it, test it, send it back and try again'. Just kinda stinky, that's all.

 

[Michiel Schierbeek]

Thank you Tim for all your tests!
So if I want a good copy I have to plough through several testing periods to find the right lens or I could be lucky and find a good one at once.

 

[tashley]

Just listen to Pharell Williams and Daft Punk while you shop, and you may not have so stay up all night to get lucky!

 

[jonoslack]

Re: Bingo! I got a good one

Great news that a good copy finally landed at your doorstep! It's such a bummer though, and this is surely not a Sony specific issue, that one has to go through the hassle of multiple iterations of lenses until the proper one shows up. If a lens, or any product, doesn't meet spec at the manufacturing QC stage, it shouldn't be shipped. It makes me think that either the window on what 'meets spec' is way too large, and/or the QC folks aren't doing a good enough job. Kinda sucks for consumers of this not-inexpensive kit.

I remember when the Nikon 17-55 came out I bought it from KP Professional in Cambridge - a big dealership which has long since gone to the wall.
I returned the lens 5 times for decentering - they were perfectly okay about it and replaced it (the Nikon rep got involved as well)

At the end we were discussing the issue, and I asked him how many lenses he got back because the IQ wasn't good enough . . . . he said it had never happened before.

I suspect it isn't so different now - if people can't tell the difference between a good copy and a bad copy of a lens, then it isn't worth spending the money on making sure that the lenses that come out of the factory are good.

As Tim says - if you're fussy you're getting a good deal - more than you are actually paying for.

This isn't just an issue with Sony (actually I think it's Less of an issue with Sony)
All the best

 

[vjbelle]

I'm on the same page Tim.

1) Don't correct the distortion if you don't have to.
- Marc

+1...... Any correction has its detrimental affect on pixels. I have found that there could/can be a difference between correcting with a profile and doing it manually in PS. Sometimes PS has a 'gentler' affect on pixels. But, if you don't have to then don't....

Victor