Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: John R Smith on May 14, 2010, 10:21:50 am
-
After my initial ghastly focus results using my old Hasselblad 500s with a CFV-39 DB, and posting the topic here on LL, I had to take stock of my shooting technique. What, exactly, was going on?
Well, we had all the technical explanations for the cause of Focus Anguish in that thread, and they all seemed to make good sense. The question was, what to DO about it. After all, when you have just paid out the most money for anything you have ever purchased in your entire life (my house excepted) for a little black box with some buttons on it, used up your entire savings and had to sell a vintage guitar in the process – and when you are producing worse pictures with it than you were with a three-quid roll of HP5, that tends to be pretty depressing.
So I did some serious thinking, and testing. The curious thing was that some of my shots were perfectly in focus, so obviously I was doing something right, some of the time, without realising it. The other piece of evidence came from focus testing on a tripod. Using my 120mm S-Planar lens at close range (say between 6 and 3 feet) I could in fact get perfect focus every time. For example, with the leaf of a fern outside my door, where the leaf was coming from the roots towards me at an angle of 45 degrees, I could focus on the tip, or half-way down the leaf, or at the base and I would be bang-on for each shot. This, I reasoned, was because in this case I had something big and clear to focus on which filled most of the screen, and I was in the area of lens helicoid movement with the most sensitivity and finesse. So then I knew that there was nothing intrinsically wrong with the focus screen position, the mirror plane, or the alignment of the DB to the body. A good start.
At the opposite end of the spectrum, distant shots with the 150mm or the 80mm focused at infinity were also perfectly crisp. It was the shots where the subject was between say 20 to 60 feet away which were the big problem, with any lens. This, of course, is just the area where the lens focusing movement is very small (and pretty stiff on the old Zeiss ‘C’ lenses) and has the least finesse. Fair enough, so what goes wrong in practice?
There we are, out for the afternoon, and the perfect subject lies before us. A splendid tombstone, aged and weathered, with fine lettering and the light slanting just so. Framing it perfectly and about 20 feet behind are the mellow 16th century church porch and an ancient yew tree (yes, my middle name is cliché). So we frame up in the viewfinder and, using the WLF magnifier (or a prism), wind the focus ring until the subject appears to be sharp. Now this is where we encounter a problem which I shall term the Area Of Uncertainty (or AOU). We can stop as soon as the subject appears to be sharp, but we find that we can move the focus ring even further, back again, and we can see no change. This is the AOU, and it exists because we do not have enough magnification on the focus screen to judge the tiny changes in focus within this critical area. A 6x or 8x loupe would probably help, but with the standard finder we cannot judge focus correctly inside the AOU, which is the range of helicoid movement within which we can see no change.
We humans may not be able to judge focus within this zone, but the sensor in the MF DB most certainly can. This malignant little rectangle of silicon will mercilessly throw the results back in your face at 100% or even 50% when you download your cherished results that evening. And the typical scenario goes like this –
We raised the camera to our eye, with the lens focused at infinity. We rotated the focus ring out, focusing to the near field until the tombstone appeared to be sharp. We hit the AOU and stopped, and tripped the shutter. On screen, we can see the awful truth about what really happened. Fooled by the AOU, we had stopped too soon and we were actually focused behind the subject. Consequently we have an image where the subject itself is soft and everything immediately behind it is pin-sharp, highlighting the fault even more. And because there is less depth of field in front of the plane of focus than behind, the grass and flowers in front of the tombstone are totally OOF, making things even worse.
So what can we do? The answer is to make use of the AOU to our own advantage, accept that there will be a focus error, but make it work for us rather than against us. We must make use of the fact that there is more DOF behind the subject than in front of it. Here is the simple rule – focus the other way around – not from infinity to the subject, but from close-up to the subject. So again we raise the camera to our eye, but this time the lens is set not to infinity (as I always used to do) but to the near field, perhaps 3 or 5 feet. This time we rotate the focus ring in, not out, and stop when we hit the AOU and the subject appears to be sharp. On the PC screen, although the plane of focus is still in error and this time slightly in front of the tombstone, the subject appears sharp because there is sufficient depth of field behind the focus point to keep it in. Even better, the grass and flowers in front are also pin-sharp. Behind the subject, DOF fades away gracefully, and though the porch and yew tree are slightly soft our eye accepts this as they are in the background.
So there is the simple trick which has really improved my manual focusing on the old ‘Blads. Perhaps I should coin another acronym – Focus From The Front, FFTF, a bit like ETTR, perhaps. It may be that loads of you chaps already do this, or it may be that you think it is a load of pants. Fair enough. But it might be worth a try.
John
-
Hi John,
How about whenever you are in the AOU, not just focus from the near field and stop, but then tweak it just that little bit further. Are you able to get the subject actually in focus then? Sounds like something I should try on my Bronica/Ixpress tomorrow morning! Have you tried doing the same, but then reconfirm in Live View? (<--- Yes, I know you'll have to do this indoors, or at least somewhere where you can safely setup a laptop or something)
Jason
-
With all due respect, if you have to think about it that much, something in the design of the camera or back or their interaction is truly not working. End of story.
Or else, use V bodies only on still life jobs, while on tripod, completely unhurried.
No thanks.
-
Jason
I can't do live view as I have no kit for that, and don't work in a studio. But yes, in theory you would be able to pull perfect focus like that. I am out in the fields and churchyards all the time.
John
-
Not to be simple minded about this, but I always use a split-prism and then find a vertical line to work with. Micro-prism or even fresnel without a lupe is just not enough differentition between in-focus and out of focus, especially with wide angle lenses.
FWIW, there are other approaches: for example, the Hy6/AFI has pretty good focus confirmation. The older 6008AF required a chimney finder lupe, which works well enough, but is big. SOme have adapted after-market lupes to put on the cameras... perhaps this could done for the V Hassys.
It seems that the Hy6 doesn't need this - as the focus confirmation is not only right on, but a small enough target to actually get the right thing in its sights.
-
Or else, use V bodies only on still life jobs, while on tripod, completely unhurried.
*********
That is the bottom line if critical focus at 100% is desired. Tripod, mirror up, remote release., etc. I have tried most magnification options, including the loop attached to the prism and found that the focusing hood that Hasselblad made for awhile for use with digital back is the best option. The 4 x 4 DPS hood has 5.5x magnification and combined with the Acute-Matte D, #42215, screen probably provides the maximum focusing accuracy with the "V" series.
Steve
-
I should perhaps add that I have magnified things somewhat in my description to make the process clear. We are talking about tiny differences, in reality. On film, for all the reasons which have been explained in the original thread, one rarely notices these issues, but of course they are still there. It is because (on my monitor) I am looking at the equivalent of a six-foot wide print from a 39MP file at 100% that focus problems scream at you. A number of the photos I have rejected would have been perfectly acceptable printed at 10x8 ins, and in my darkroom days I would have done so.
gwitf's comment is fair, but only in the context of modern camera systems. We have to remember that the 'Blad 500 and the Zeiss lenses I am using are all 50 year old designs, and they were never intended to be married to a sensor which would allow this level of forensic examination. So I do cut the old girl a bit of slack.
Actually, I think I may have made things a bit over complicated. I do have this terrible tendancy to get carried away and become verbose. So we could sum up the thrust of the argument as -
* We are likely to get focus errors on the 500 cameras with the old 'C' lenses in particular (rather than the CF or CFE, which are different beasts).
* Front focus error is better than back focus error for the sort of work I do (John's churchyards).
* So stack the odds in your favour and Focus From The Front.
YMMV, as they say.
John
-
It works, right? Than it is a good system
-
Thanks for sharing your experiments here John. Not sure if it will be of particular use to me as I use much more modern gear (just recently the Afi II-10), but it's another option I'll bear in mind.
Philip
-
Well written John, clear and compelling to try.
-
Sounds like you are photographing static subjects on 'free' digital memory
BLF
Bracket Like F c
thats what I do on both my H1 and my mamiya when I had one
S
-
gwitf's comment is fair, but only in the context of modern camera systems. We have to remember that the 'Blad 500 and the Zeiss lenses I am using are all 50 year old designs, and they were never intended to be married to a sensor which would allow this level of forensic examination. So I do cut the old girl a bit of slack.
I didn't want to come off as flip in the way I responded, but at some point, you need your brain to be concentrating on what's in front of you, and not be dicking around with the camera, and worrying about it. That's all I was trying to say. I have two recent memories illustrating that:
1. I shot a job last week where they wanted about three inches in focus. I shot it with the 85mm f1.2 II on the 5d2, but I couldn't trust any of those outer focus points, and couldn't get the prop overlapped on those outer focus points anyway, so I just gave up and manually focused. That 85 is voodoo; it's just got a mind of its own. It focuses where it wants to; it doesn't give a damn what you're seeing in the viewfinder. I was wide open, which is suicide, so I just started "focus bracketing" which is such a waste of time and energy. All that fucking technology in that studio, and in the end, it came down to "you can't trust what you're seeing in the viewfinder". Are we moving forwards or backwards, with all this so-called precision digital technology? So just shoot, shoot, shoot, and focus bracket, and maybe you'll get some sharp frames.
2. I was getting props on another job last week. I was rolling this kart down the sidewalk on 28th Street, and it was rickety, and full of props, and the front wheel would come off if you hit a crack in the sidewalk. So you've got to devote most of your brain to watching out for the wheel to come off, rather than thinking about the job, or something worthwhile. Wheel comes off in the middle of 8th Avenue, and I'm chasing it, dodging cabs, and trying not to turn over the kart.
Point: When you've got gear that you can't trust, or is so high maintenance that it requires sizable brain power just to bring back a frame that's in focus, it's time to switch gear. Again, we're not talking about whether the picture you're about to shoot is a Pulitzer or CA-worthy -- we're talking about whether it's in focus! The gear works for you -- not the opposite. I just feel that the V body served its time, but now its time to put it out to pasture.
-
John,
I had similar experiences with digital medium format. It's very demanding when it comes to focus. As others have mentioned, I did find that it took away from my being 'with the subject'. Others develop a technique and it becomes second nature. I suppose time will tell.
I just wanted to let you know that you're not alone. Best of luck with it--I hope you master the technique.
All that fucking technology in that studio, and in the end, it came down to "you can't trust what you're seeing in the viewfinder".
Did you try Live View?
-
Thank you all very much for your thoughtful responses. I appreciate it because it took me quite a bit of effort and time to put the post together - it's one thing doing something in the field but quite another to try to write a technical explanation of it for others to read.
Point: When you've got gear that you can't trust, or is so high maintenance that it requires sizable brain power just to bring back a frame that's in focus, it's time to switch gear. Again, we're not talking about whether the picture you're about to shoot is a Pulitzer or CA-worthy -- we're talking about whether it's in focus! The gear works for you -- not the opposite. I just feel that the V body served its time, but now its time to put it out to pasture.
Absolutely right. And for the same money that I laid out on my kit, I could have bought a Canon 1DS mk III or whatever and a couple of decent lenses, and got just as good or better pictures with a lot less effort. Which is the reason that Hasselblad quietly retired the V-system and developed the H-system instead. There is a difference, though - you earn your living with a camera but for me it is just recreation. So it's a bit like running an old MG or Triumph TR2 or something - I do this for fun and it does not have to pay the rent. We English are a nation of eccentrics, you see. And I do have a beard and glasses and wear a tweed jacket.
John
-
- it's one thing doing something in the field but quite another to try to write a technical explanation of it for others to read.
And I do have a beard and glasses and wear a tweed jacket.
John
Well, John, it depend what you are trying to do in that field - it can get quite cold near the sea and that wind can make you get your own back - take care!
Tweed jackets are never going to replace denim jackets; as you almost pointed out, it's all part of part of the Series 500 mindset, which I share as far as the Series 500 camera goes, but not as far as those Hebridean sartorial choices are concerned!
Rob C
-
it came down to "you can't trust what you're seeing in the viewfinder
As was remarked on above, if you use the Live View on the 5D mark II for manual focusing then what you see is what you get, spot-on focusing.
Philip
-
As was remarked on above, if you use the Live View on the 5D mark II for manual focusing then what you see is what you get, spot-on focusing.
Philip
Yes, as long as it's not living, breathing people that are moving around and changing position after every frame.
We have invested this much money in all this digital technology, and we have now sunk to a new low of being forced to say "the only way to shoot a photograph that's in focus is to use Live View"....? If we progress at this level, it won't even be thought of as photography any longer; it'll be thought of more like CAD Rendering or something, where the soul is actually sucked out of the process, and we all need tripods, computer screens, and crew in order to shoot a simple, sharp photograph.
I'm just raising my hand here, and saying that something is wrong with this entire equation. This is Progress?
-
Yes, as long as it's not living, breathing people that are moving around and changing position after every frame.
We have invested this much money in all this digital technology, and we have now sunk to a new low of being forced to say "the only way to shoot a photograph that's in focus is to use Live View"....? If we progress at this level, it won't even be thought of as photography any longer; it'll be thought of more like CAD Rendering or something, where the soul is actually sucked out of the process, and we all need tripods, computer screens, and crew in order to shoot a simple, sharp photograph.
I'm just raising my hand here, and saying that something is wrong with this entire equation. This is Progress?
Well, this echos - no, underlines my position made clear here quite often: were I faced with living my life again, I'm sure that a career as a photographer, even photography as a hobby, would never have entered my mind. It is so different - alien - to the photography that existed (and seduced me) with film and film cameras that the charm has vanished completely. I never did think of technology as anything but the simple understanding of why film makes an image, how to handle that and how to use different focal lengths. End of technological learning curve. Start of the artistic one.
Today's photography just appeals to a different mindset, and if one has the bad luck to be caught in the flux at the wrong stage in life - time to think again.
Rob C
-
Yes, as long as it's not living, breathing people that are moving around and changing position after every frame.
Sorry, I made the apparently wrong assumption from your earlier remark that you were shooting still-life/product. Yes, with moving subjects Live View is of little use.
I agree with your other points about technology getting in the way sometimes these days.
-
Sorry, I made the apparently wrong assumption from your earlier remark that you were shooting still-life/product. Yes, with moving subjects Live View is of little use.
I agree with your other points about technology getting in the way sometimes these days.
I dont see tehnology getting in the way
the only thing getting in the way is pixel peeping and rejecting images that are a little off at 6ft wide - many classic images are a little wooly if not outright 'out' - Capa on the DDday beach hardly wins any technical awards
as I have used my D3 recntly I shoot so much more for 'the moment' I did present an image to a mag of a kid doing something cute - the focus was a little off and they had a moan
-the focus was above their print resolution
-i had sharp frames that did not have 'the moment'
So I suppose technology is getting in the way - the clients ability to over analyise the image
S
-
Sam
Reading the posts it seems to me that the problem is way beyond pixel peeping sizes - if you end up with focus six feet before or behind your subject you need VERY low res or tiny telephone-sized screens to get away with it!
And if you get the reflections on the lips crisp when the subject you wanted was actually the makeup around the eyes...
It'll never happen, but perhaps if all those guys who do that kind of work decide, en masse, to return to film for such shots, maybe somebody somewhere will listen and things would change. I suspect that part of the problem might be that cameras are trying to cover too many bases - be all things to all men. Yes, I know that they 'market' some aimed at sports and others for more 'fine grained' situations by virtue of different sensors, but is that the case with MF too? As I've said before, they should perhaps have worked more on getting a sensor that covers the full 6x6 version of life and retained the perfected body shape that already exists. You pay so damn much for those MF cameras already - I would imagine that proper FF wouldn't be much of a spoiler for those in that rarified market. It is always nicer to be able to crop if you want to than have it dictated to you.
Rob C
-
Yes, as long as it's not living, breathing people that are moving around and changing position after every frame.
.........snip......... This is Progress?
None of these suggestions you'll like (or even do) but if you put a 7" marshall monitor on your 5d set on live view you'll hit focus 9.9 times out of 10 if your talent moves into position, hits pose holds for a brief second, because at 7" the view is so detailed it's hard to miss focus.
It takes some getting used to because at first there is a slight disconnect feeling standing a few inches back from the camera and framing, focusing and shooting but after a while it becomes second nature. Kind of think of it as a smaller view camera with a big bright ground glass turned right side up.
Let's face it, you use to do this with a fuji 680 and you hit some, missed some but nobody lost their mind and the fuji was a hand holding beast compared to a 5d2 and a 7" monitor.
For client review, set up a field monitor from the marshal monitor using bnc cables. Its analog, but it will let them see what your shooting and they won't complain.
Now if your subject is moving running on the far side of the frame, rent/buy a 7d, set it on continuous focus and mark the prop by depressing the shutter halfway until those little blocks of focus areas light up. The focus points of a 7d or any 1.5 crop camera cover most of the frame and it will usually track true, as long as you keep the shutter pressed partially down. And you'll have to use a 50 instead of an 85* but it'll look pretty much the same.
You'll know it's a cropped camera, nobody at the tech station or the retoucher will.
But in medium format land, I think it's doable. A few weeks ago we shot 900 or so images with the Contax and the manual focus 120mm. Mostly full length models to 3.4 crop. I haven't shot solely manual focus in a while and maybe it was my lucky day, maybe my eyes have changed, but it worked to the point the few times I used other lenses like the 80 or the 140 I just left them on manual focus and shot.
I would imagine if the Contax with it's smaller viewfinder will do this the H-system won't be a problem, with that true focus thing probably better.
As far as progress, yes, I'd have given anything in the manual focus days to have a 7" ground glass that showed me my exposure.
Now if you want to try something that is progress go borrow one of those panasonic G cameras and set it on face recognition. The one that uses the lcd as a viewfinder. As the model walks through the frame you can see the yellow square just track with her face. Soon you'll be able to touch screen the area you want to track and it will follow. Not today but soon.
BC
*I rarely if ever use the 85 1.2., nearly always the 85 1.8. The glass is so large on the 1.2 it takes it too long for the focus to react and I think the 1.8 is sharper anyway.
All IMO.
edit
Working in live view, either video or stills I've noticed with the Canon lenses, there seems to be this weird kind of disconnect between turning the ring and hitting exact focus. Not that the lenses aren't sharp, I actually like the look of most Canon lenses and the prices are good, but they just don't have that locked in focusing feel of older manual focus nikkor lenses, or my contax lenses.
When these come out you might want to give them a try.
http://www.zeiss.de/c125756900453232/Conte...125756f003e6703 (http://www.zeiss.de/c125756900453232/Contents-Frame/042839dea0e28e5fc125756f003e6703)
They are true manual focus lenses with gears for follow focus and are probably the only true future proof product on the market today, cause the mounts will interchange between Canon, Nikon and PL cine, so when you finally buy that RED Epic you'll have the lens set. (insert smile here).
-
*I rarely if ever use the 85 1.2., nearly always the 85 1.8. The glass is so large on the 1.2 it takes it too long for the focus to react and I think the 1.8 is sharper anyway.
Working in live view, either video or stills I've noticed with the Canon lenses, there seems to be this weird kind of disconnect between turning the ring and hitting exact focus. Not that the lenses aren't sharp, I actually like the look of most Canon lenses and the prices are good, but they just don't have that locked in focusing feel of older manual focus nikkor lenses, or my contax lenses.
That is my feeling too -- there is a disconnect between the physical operation of the 85 lens, with what's actually being recorded onto the sensor. Very scary. I too, regret buying the 1.2; wish I'd bought the 1.8 instead. In fact, I think most anyone could make a case that no Canon lens should be much faster than about f4, if true focus dependability is needed.
To be clear, there are two issues at hand here:
1. That what you see in the viewfinder, even when shooting wide open, is not nearly indicative of what's rendered into the file.
2. That you could clearly be looking at something thru the viewfinder that appears tack sharp, and yet, it's rendered focused several inches to the front or rear, with no real pattern, and no sense of repeatability, and thus, no way to really attack the problem, (other than to switch to Nikon).
There's just not many things worse that standing in a studio full of all this expensive technology, and all these supposed precision cameras and computers, and have the Tech yell across the room, "Nope, you missed it. You're about four inches focused forward", followed by, "Nope, you missed that one too, you're about six inches to the rear", all in clear range of the client.
You write the big check for all this stuff, and then someone whispers, "Well, if you really want your photographs to be in focus, you've gotta buy the XXXXX focusing screen, or the XXXXXX external monitor, or you've got to shoot in LiveView, (making your snazzy 35mm camera operate as fast as a 1952 wooden Deardorff)". Basic focus. Very frustrating.
Obviously, Hasselblad is dealing with this factor with True Focus. But I have not tested the H4D40 on a real job situation. But at least their lenses are more realistic, more in the f2.8-4 range.
-
I'm far from being an experienced photographer, but from my humble experience so far I have never had an higher % of focused files (not viewfinder trusting as you point) than using MF + focus confirmation indicator. I've done many many times in different ways and, at least for me it works way faster and it is more reliable. Now I only look at the green dot, recadrage, et nothing more.
-
Sam
Reading the posts it seems to me that the problem is way beyond pixel peeping sizes - if you end up with focus six feet before or behind your subject you need VERY low res or tiny telephone-sized screens to get away with it!
And if you get the reflections on the lips crisp when the subject you wanted was actually the makeup around the eyes...
It'll never happen, but perhaps if all those guys who do that kind of work decide, en masse, to return to film for such shots, maybe somebody somewhere will listen and things would change.
Rob,
The thing is that at a given print size the digital image will always be at least as good as the film. The lack of focus accuracy impacts the ability to tap into the much higher resolution potential of digital. This shows badly when looking at images at 100% on screen, which we were not able to do with film. Put it otherwise, we have been living for years with poorly focused film images but never really noticed because 99% of applications are OK with this, just like they were in fact OK with 12 megapixels digital.
Anyway, the only real answers to increase the accuracy of digital focus are:
1. Live view
2. Reliable and accurate AF
24 mp class DSLR have the same pixel pitch as the P65+ and 99% of my images appear tack sharp when viewed at 100% on screen thanks to the use of live view. That is at least the case with longer lenses, wides are sometimes difficult to focus accurately even with live view.
Cheers,
Bernard
-
Bernard, I understand what you are saying about 100% viewing, but two things: I have scanned many many of my 35mm Kodachromes and find them perfectly all right at high percentages when I have to go there and spot; the sort of focussing issue that gwhitf is talking about would have ruined any film capture too. I don't believe the problem is as simple as 100% viewing making the difference; I'm pretty sure there is more to it than that. For example, this uncertainty with Canon lenses compared with Nikkors that both gwhitf and bcooter acknowledge... that's not normal photography experience.
Rob C