Wth all these discussions on focus issues I was curious as to how much I could realistically expect to zoom in on a photo in Lightroom 3 and still have a clean view? Should I be able to go 1:1 and expect it to still look good? Maybe to put it another way, at what zoom ratio does it no longer matter if the shot looks good because it will never be printed large enough to see it that close? Hope this is understandable. I am having problems coming up with the correct terms this early in the morning.

Thanks

Visit my Nikonians gallery.

In general viewing an image on screen at 1:1, my opinion is doing so with a 16MP sensor is problematic.

Monitors are in general low-resolution devices. Some of the higher end monitors are around 133 PPI, but most are between 86 and 100 PPI.

My monitor is 86 PPI and at 100% view (or 1:1 viewing) with a D7000 full resolution image you are at about a 60 times enlargement. My viewing distance is about what the generally stated viewing distance for an 8x10 print would be. The CoC tolerable for that viewing distance is about .3mm. The diffraction spot size at f/5.6 is about 6 to 7 microns (depending on whose table you look up). With the 60 times enlargement you would be able to see the softening of detail due to diffraction if the image was taken with an f-number of between 4 and 5.6 or greater.

And I think that is where some of these back/front focus threads are coming from. We routinely now look at images at much greater enlargement than we ever did with film and expect crystal clear images. So is it that the D7000 is a flawed design, or are we now seeing an issue with AF systems that was there all the while but never noticed because of the lower enlargement factor. Phase Detection AF systems do not provide precision focus, but tolerable accuracy.

I have not found any references from Nikon, but there are interesting discussions over Canon’s AF system. For instance this article discusses the Canon AF Phase Detection system. It states that the tolerable out of focus range for their AF Phase Detection system is .035mm for full frame and .02mm for APS-C sized sensors (you might notice it’s the tolerable CoC when calculating DOF for an 8x10 print at normal viewing distance). So if the AF system chose a focus distance that was at the edge of the tolerable range, when you view a DX 16 MP image at 100%, that blur circle is now 1.2mm or 4 times larger than the tolerable .3mm for an 18 inch viewing distance so it will look soft. And while someone may be disappointed that their image is not tack sharp when viewed this way, it is within the tolerable range of operation of the system and sending in the camera to be checked may result in a report that there is nothing wrong with the camera.

Just my thoughts on this, any comments or corrections welcomed.

Pete

Visit my Nikonians gallery.

Generally, I don't use anything greater than 50% magnification to judge sharpness and image quality, and 25% is my norm. That's with my 12MP cameras - with 16MP images it would be even less.

Brian
Welsh Nikonian
Check out the Nikonians Team pages

Excellent discussion above. I also think it depends on your individual standards, applications and lenses.

Here's a shot, taken largely at random from the keeper pile:



D3, 400/f2.8, TC-20EIII, f/11, 1/500th, ISO 800, monopod - normal processing in LR3. Of course this is a 12mp image - I have no camera exceeding 13mp anyway. And here's a 100% crop from the center:



There are many things about this shot that are less than optimal from a sharpness standpoint: 800mm focal length at 1/500th, a subject moving at least 65mph almost directly at the camera, a complex surface for the AF to lock onto, aperture at f/11, the use of a 2x TC, and a mild dose of noise reduction. Also, it's worth observing that while I shot this on AF-C, the subject is not only fast but also changing speed. After the cars crest this hill, they accelerate through the corner that's just a few feet in front of this shot. And this is a Formula One car, probably accelerating at "only" 3g here. That means that the camera's dynamic tracking mode can't even predict accurately the speed a few milliseconds later when the shutter flies.

I'd call this a standard high-quality capture using good (but likely not excellent) technique on top-caliber equipment. This file would make at least a good 16x20 with a little further processing and good technique in printing. I expect to do at least this well for any money shot, and I think it holds up pretty well at 100% on my 133 dpi monitor. If it looks unsharp, look at the large N a little to the left of center in the crop. (And I don't want to hear that this "isn't a good AF target" - it's my subject, and it's what I have to shoot in the real world. And anyway, the result looks pretty good to me.) Besides, Google Maps tells me that the shooting distance was 250 feet. At 800mm, f/11, the DOF is twenty feet - more than the entire car from nosecone to rear wing, so to a first order, everything in this crop is in focus. Unsharpness seen here is far more likely due to subject and/or camera motion than AF errors.

A different shot, this one from the 200/f4 Micro-Nikkor demonstrates what I think is a realistic maximum:


D2x, 200/f4 Micro-Nikkor, f/7.1, 1/350th, ISO 100, tripod/remote/Mup, normal LR processing. And here's the 100% crop from the focus area:



It probably doesn't get a whole lot sharper than this, or at least I don't know how to do much better. In this case, the subject is totally motionless (not even a hint of a breeze from the fans - it was shot indoors in a conservatory), a top-caliber lens used squarely in its design center, at an optimal aperture, careful manual focus, a fully sufficient shutter speed, camera and lens locked down on a very solid tripod, with mirror up and a remote, and base ISO so no noise reduction applied. It's shot on DX so there are as many pixels as possible in the cropped subject area. I assume that I could print this one at 24x36 or better without any difficulties, since it looks pretty darn good at 100%, not that I think that this is a good way to evaluate. But clearly it holds up well EVEN at 100%, certainly the size equivalent of a window, if not a door.

By "normal" processing I mean standard global adjustments, such as import from raw, crop, sharpening strength and masking, and in the first one, some noise reduction. There's no local optimization for micro contrast or anything like that. These took less than two minutes each for processing.

Now these were shot with different cameras, and I suppose that one could argue that the D3 could do #1 better than a D7000, since it is in fact an AF problem. But the second one was manual focus and not done with LiveView, which the D2x lacks and the D7000 has. There is no reason that the D7000 can't duplicate #2 if not exceed it, given the same lens, tripod, remote - and equal or better skills.

And that's the rub. The result that gets to Lightroom zoomed 1:1 is the worst of the quality of the camera, lens and AF system - and also the support, support technique, AF technique, capture parameters, capture mode (eg JPEG vs raw), processing quality, and probably some other things that I didn't mention. If any part of that connected chain of factors is less than expected, the net result suffers. In my opinion, at least 95% of the complaints about "the AF" or for that matter lenses fail to account for errors in the rest of the pipeline. And most of them have readily identifiable shortcomings in areas other than the hardware, shall we say. The above is shown to demonstrate what's possible, at least by me. I believe that other photographers are more skilled than I am, sometimes by a wide margin, and I'm confident that these aren't really the limit of what can be done.

----

Finally, note that going PAST 1:1 on screen is simply not productive. You're looking at the pixels that are being synthesized by the software, and unless you're going to print THOSE pixels, you're beyond the information you've captured. I wouldn't judge pixels zoomed past 1:1 except as the finished artifact - the print itself on paper. (Of course, zooming in past 1:1 IS useful if it allows you to wield a smaller and more precise tool for editing. It's just not useful for judging what pixels you have captured.)

_____
Brian... a bicoastal Nikonian and Team Member

My gallery is online. Comments and critique welcomed any time!

Attachment #1, (jpg file)
Attachment #2, (jpg file)
Attachment #3, (jpg file)
Attachment #4, (jpg file)

Well Brian, I might not be up on all the latest complaints and debates but there are about 1000 of them that would be shut down completely simply by viewing those two crops. Something about a picture worth a thousand words.
I would say pretty darned impressive, good work...to you AND Nikon.

I know that when I do a good job of the basics, the D7000 delivers the most detailed and sharpest images I have ever taken...mostly based on my film days. One thing I find a little frustrating is AF, it was so easy to get just the focus plane I wanted with MF using a big bright 35mm micro-prism and split D. With big aperture lenses like my 85 1.4 having a good MF optimized focus screen would be sweet. I was planning to get a Katz-Eye screen for the D7000 but things have come up that give the possibility to order a D4 if the announcement is in January when I am back in California for 3-4 weeks. I would be anxious to hold one but the 6-8 months needed to get one would be tolerable since the D7000 is so satisfying.
Stan
St Petersburg Russia

Visit my Nikonians gallery.

> good work...to you AND Nikon

Thank you

> there are about 1000 of them that would be shut down completely simply by viewing those two crops.

I actually considered not posting them, because I think they can be misinterpreted. As in "well mine aren't this good, so my camera must be broken" or similar. And there's also the possibility of the complaint "well those are better cameras than a D7000 so they should be better." #2 should be reasonably proof against that, though.

Maybe I'll rent a D7000 for one of my next trips.

_____
Brian... a bicoastal Nikonian and Team Member

My gallery is online. Comments and critique welcomed any time!

There is a good probability that some, many, would jump to the conclusion that the camera failed them long before they jumped to a conclusion that they failed the camera. A more observant reader would note that you made a great many small aggregating steps to make them turn out this good...by giving the equipment a chance to shine with proper exposure, speed, support, mlu, delay, timing, remote etc.

I have noticed that there are really two types of users who are using the upper end hobbyist cameras and the difference between them is not subject matter or experience etc, it is how they go about solving problems or identifying unknowns. Those whose first inclination is to assume the camera is at fault are going to have one long series of frustrating relationships with all sorts and brands of cameras. "It must be the camera because I have been shooting a long time and did not change anything" is less than encouraging approach to solving problems or understanding the concepts that are fundamental to the process of getting images as intended.
Stan
St Petersburg Russia

Visit my Nikonians gallery.

I can surely attest that at 1:1 on my laptop monitor I can often spot less than perfect focus on the D7K pictures. I know this for a fact, when in the same picture or the same series of shots I can also identify areas in perfect focus (just not where I wanted them), so don't come with the capture technique or monitor performance, or other usual arguments.

I'm not analizing the AF error fault (machine or operator) merely stating it can easily be seen in a regular monitor at 1:1, unless the image is contaminated with other causes for blur (noise, subject movement, camera shake,

Now, whether if this close-but-not-quite focus is adequate for most uses, that's a whole other argument.

http://egozarolho.blogspot.com
1. Good content, good aesthetics and good tecnique. On that order.
2. Light is more important than glass and pixels.
3. In the digital photography process, software is as important as gear.

>I know this for a fact, when in the same picture or the same series
>of shots I can also identify areas in perfect focus (just not
>where I wanted them), so don't come with the capture technique...

I'm not sure I follow what you mean there. Are you saying that, if something is in focus in an image but not what you intended, then that cannot be caused by faulty technique? If so, I guess it depends on what you include in "technique". Certainly, if something is sharp, then the problem is probably not down to imperfect hand-holding technique, but I would argue that setting up the camera systems to suit the subject and the situation counts as "technique" too

And if it's necessary to enlarge the image to 100% to see any focus error, does it really matter anyway - unless you are planning to view a huge print at less than arm's length?

Part of the problem with these discussions is that many people have an unrealistic expectation of AF systems. Today's systems are very, very good, but NONE of them are capable of achieving perfect focus every time.

Brian
Welsh Nikonian
Check out the Nikonians Team pages

The situation i was considering was what led me to believe I had a problem. On a headshot portrait focused on the eye, the eye would be blury, but the hair and the ears would be razor sharp (viewed at 1:1). This eliminated camera shake and subject movement, while at the same time proving what the lens was capable of.

http://egozarolho.blogspot.com
1. Good content, good aesthetics and good tecnique. On that order.
2. Light is more important than glass and pixels.
3. In the digital photography process, software is as important as gear.

Thanks for the explanation. I agree that such a problem wouldn't have been caused by camera or subject movement. But... it wouldn't prove a camera or lens fault without first eliminating the possibility of inappropriate AF settings and/or target selection. We probably shouldn't go there again in this thread

Brian
Welsh Nikonian
Check out the Nikonians Team pages

Couldn't agree more. Thank your your expertise, as well as your common sense, patience and dedication.

http://egozarolho.blogspot.com
1. Good content, good aesthetics and good tecnique. On that order.
2. Light is more important than glass and pixels.
3. In the digital photography process, software is as important as gear.

At 1:1 you are seeing additional data that was created by the software rendering it that interpolates the addition date needed for display due to the mismatch of image data and screen resolution. It might be fooling you. A representative sampling will be more accurately reflecting actual recorded data at 25% or 50%.

Stan
St Petersburg Russia

Visit my Nikonians gallery.

Brian,
Regardless of the debate about AF systems or issues with the D7000, I just have to got to say...you make some impressive artwork with your eyes, hands, cameras. Both images you posted are amazing, and now I'm sitting here sifting through your gallery wanting to see more.

Thanks for sharing your work with all of us.

Kevin

>at what zoom ratio does it no longer matter if the shot looks good because it will never be printed large
>enough to see it that close?
This is part a "how long is a piece of string" type question.
Assuming you are printing in 2:3 format (the same as the camera) the D7000 has 4928 pixels on the long dimension.
If you consider mathematics in isolation and accept printing at 150 dpi is your minimum standard for printing you can print a 32 inch wide print.
If 200 dpi is your minimum stand you can print 25 inch wide. If you want 300 dpi for magazine reproduction (to avoid moiré) the magazine can print 16 inches wide.
Straight away this messes up any constant zoom ratio
Under critical examination a picture at 6400 ISO is unlikely to be as sharp as one at 100.
jpeg basic is not likely to be as good at big enlargements as RAW well post processed.
Interpolation software "queses" missing information when the mathematical limits are exceeded - often with considerable success at double the non interpolation dimensions.
A 13 inch £250 laptop screen is not going to be as good for assessing sharpness as an £800 separate monitor.
Brian's suggestion of never more than 50% and often no more than 25% is similar to the standard I use.

Photography is a bit like archery. A technically better camera, lens or arrow may not hit the target as often as it could if the photographer or archer does not practice enough.

Len Shepherd