My apologies to begin with, because I distinctly do not like to single out anyone or their photography.

The "meaningful detail" phrase perhaps needs discussion of what is and what isn't meaningful detail.

When the noise is too great, the detail is not useful. That is true even if the detail is visible. Quantization distortion is visually very unpleasant to most people, though even then it depends of the nature of the image.

Your 3rd image is an excellent job of editing, and a very good example of exactly why editing a JPEG is a poor substitute. There is an astounding amount of quantization distortion visible throughout all tonal ranges of the image that are not virtually pure white or pure black. Likewise there are very visible "jaggies" along diagonal edges due to the block sampling in a JPEG.

Notice also the rather harsh tones for all colors and across all brightness levels. There are no "creamy" gradients anywhere.

I would not want to guarantee that a RAW file would produce a great picture either, but it absolutely would look better than that one.

Your images demonstrate exactly the same issues that dsturr demonstrated in his article posted a bit before yours.

I'll try to shoot something of interest in the next couple days with this thread in mind, and take some shots that are 3 to 5 stops under exposed, and then come up with examples showing the difference between editing JPEGs and RAWs. (That way we can insult my photography instead of anyone else's!)

No questions are dumb. But the answer might be irrelevant, because what the scene looked like would be different for everyone who saw it, and really isn't important anyway!

What the photographer wants a viewer to see in the photograph is all that counts!

Street Photographer Garry Winogrand (1928-1984) had a cranky way to put it:

--<http://www.youtube.com/watch?v=YQhZcKzbM9s>

Without seeing another image to compare? Probably not. However, the fact is that I do a lot of commercial printing and under some circumstances it is in fact very obvious. In others not at all obvious. In no case is that meaningful! It misses the point!


Pixels don't have much to do with it. What is "better" depends on the expectations of the photographer. The priorities of a photojournalist or a sports photographer, as two professional examples, often literally necessitate JPEG over RAW.

The average DSLR owner who would like to print images shouldn't even think about shooting JPEG only.

The point of the never ending discussions of this topic are to help people sort out where they fall, and why one or the other method is or not a good match for them.


But that is a virtually bogus distinction, because the number of pixels has nothing to do with RAW vs JPEG, nor for that matter the amount of data encoded!


That is an interesting point. With film it was relatively difficult for the average person to get a chance to "zoom in"! They looked at the small prints from the processor, and only very occasionally decided to have even one of them enlarged. Today... people can and do zoom in on any image they like! It's not a bad thing!


Why would anyone assume either is better in every way or for every purpose? But in fact most of the people arguing that JPEG is "good enough" are assuming that, yet very little of the discussion they provide is based on when JPEG is in fact appropriate.

The point is still that for any budding photographer wanting to learn to make better images (e.g. more artistic or more creative), whether for web viewing or for printing, the use of JPEG is extremely limiting compared to post processing RAW.

Keep in mind that all digital images start as raw sensor data, and they are all processed post shutter release. The processing done in the camera necessarily has to be simplified for speed, it also has to be configured before the shutter is released, and the options or preset configurations are all determine by the engineers at the manufacturer. When shooting JPEG the trick is to find a potential image that matches what the engineer chose for processing! And there is very little adjustment available after the fact.

It works. It's fairly similar to shooting film and sending it out for processing.

But the degree of creativity and artistic expression is limited. Processing by inspection, particularly with the added option of much finer configuration granularity, means producing what you want rather accepting what you got.

As noted, pixel count is not the significance. The number of colors, another commonly raised issue, is not significant either. Dynamic range along with tonal expansion or contraction, particularly as they relate to various kinds of "noise", are the issues that make a difference.

A JPEG image has only 11 distinct levels in the 7th fstop down from maximum white. At any fewer than 8 levels per fstop our eyes see it as posterization or banding. Technically that is quantization distortion, which is a particularly ugly form of noise. With just 11 levels to start with, it simply means you cannot pull up the shadows and see meaningful detail. Compare that to a 14 bit RAW file with 15 levels at 10 stops down from maximum white. Or to the 127 levels available at the same 7th stop down. With RAW a huge adjustment (3 stops) can be made, compared to a very small change (1/3rd of a stop) with a JPEG file. That limitation is most obvious when trying to pull out the shadows, but it affects sharpening, any contrast adjustment, and white balance just as it does brightness.

I specifically pointed out the posted image cannot be used to judge sharpness. The issue of sharpness was specifically addressed with a reference to selling a 16x20 print. Given a print made with only 116 pixels per inch, it is very clear that the original had to be fairly sharp.

(BTW, your broad statement that sharpness cannot be determined on a computer monitor is absolutely incorrect. A 100% crop works very well, and is not what I posted.)


Frankly, being a "published photographer" has zero weight. Northrup doesn't compare to Roger Cicala as an expert on lens testing. However... just as with your previous article the comments you make are not supported by facts, or even in this case by your cited reference!

The video did not discuss prints. What it did was compare the Tamron 150-600mm zoom with one specific fixed focal length 400mm Canon lens (specifically pointing out that no equivalent lens is available for Nikon or Sony users, so that part hardly applies to this thread about use on a Nikon D800), and only very briefly compared the Tamron to roughly equivalent Canon and Nikon zooms. It is not exactly a big surprise to find a 400mm fixed focal length lens is sharper. (I use a Nikon 600mm f/4 if a fixed focal length is preferred...)

But your cited video very clearly said they preferred the Tamron 150-600mm over either the Canon 100-400mm zoom or the Nikon 80-400mm zoom.

I just don't see a reason to post bogus innuendo about what non-specified "professional photographers" supposedly say, when it is easy to find that real reviews by very expert reviewers do not say what you suggest. Backing that up with a review that absolutely says the opposite of what you originally claimed, but suggesting it says something (about prints) that it doesn't, is non-productive.

Note that my previous article clearly addressed head on the issues raised about printing and cropping. It's a bit difficult to deny what was demonstrated. What can be debated is how applicable the features of any of these competing lenses are for any given type of job. For example, the 80mm shorter end of the zoom range on the Nikon 80-400mm is very significant for certain types of work. I do a small amount of sports photography and a significant amount of event photography, and the Tamron 150-600mm may be great for wildlife but it just doesn't replace the 80-400mm for the other work I do.

First, slides are analog and not "8 bit". By changing the color of the enlarger light with a color head it was possible to make any color balance correction without risking quantization distortion (which is nonexistent with slides and other analog images).

There is of course a lot more recorded, both with slide film and with 14 bit RAW, than "the eye can see". In producing a JPEG formatted image everything not necessary to see the image as is, is discarded. Any editing to change it, such as white balance adjustment, is severely restricted because of that. Not so with either the slide or a RAW digital.

Editing a JPEG may produce something seen as "good enough". And also the attempt at editing might clearly have limits on how much can be done while still getting something "good enough". But there is no question that editing the raw sensor data allows greater manipulation and better results.

The upshot is that with only a JPEG, you edit and get something that does the job. But if you had the RAW file and it is appropriately edited there is rarely an instance where comparing the two results side by side would not show the RAW edit is significantly better.

It's a matter of whether you want or care about "significantly better". There is no reason that anyone necessarily has to have results good enough to print poster sized images! Some to, some don't.

"at 400mm the Tamron 150-600 VC and Canon 100-400 IS are virtually identical, with the Tamron 200-500 and Sigma 50-500 a bit behind. Let’s keep it in perspective, though, both the Sigma and Tamron 200-500 are still very good at 400mm."

That is from Roger Cicala at LensRentals.com after doing MTF 50 testing on the listed lenses. Lenses less sharp than the 150-600mm are "very good".

He also said the "Nikon 80-400 AF-S zoom is, as near as we can tell, equivalent to the Canon 100-400 IS as far as resolution goes".

The point of course is that for sharpness the Tamron, Canon, and Nikon lenses are all equal, and other considerations are more important in choosing which to buy.

What in fact has generally been agreed is that the Tamron loses sharpness as it approaches 600mm. Of course every telephoto zoom with a similar range also becomes less sharp at the longest focal length.

Because the others have maximum zoom ranges at 400mm and 500mm the relevant point is that the Tamron is as sharp or sharper than any of the others at 400mm or 500mm.

It isn't as sharp at 600mm, but the others don't go to 600mm. With any of these lenses backing off from maximum zoom is necessary. Note that, as the Exif data shows, the image posted was shot at 500mm.

Cicala's graphs suggest that staying at 550mm or less will produce the best results (equal to the Canon and Nikon lenses) with the Tamron 150-600mm.


The image posted is a 2336x1869 crop from an 7424x4924 original. It was hand held too.

That image of course was resampled and processed as a 900x800 JPEG for web posting, and it is impossible to accurately judge sharpness with something that small.

The real "product", printed from the 2336x1869 crop, was a 20x16 canvas print (sold to a local business). That is only about 116 pixels per inch, so rather clearly the original had to have been very sharp to begin with or it would not have been printable at that size.


Actually it gives you 560mm of reach, and isn't more versatile or as sharp with the TC as the Tamron. The Tamron at maximum zoom is f/6.3, and the Nikkor 80-400mm + 1.4 TC is f/8.

The Nikkor is of known high build quality, and has known compatibility (these characteristics and the comparison too are all equally applicable to the excellent Canon 100-400mm lens). The Tamron's build and compatibility appear good, but it remains to be seen how it holds up over time.

The killer with this comparison though is the price tag. At $1070 the others can't touch the value of the Tamron.

Here's another image with the Tamron 150-600mm. This is a composite made from three images (shot with a Nikon D4 though). Clearly tracking birds in flight (an Arctic Tern in this case) works well. And again these were hand held. ISO was 100, aperture was wide open at f/6.3, shutter speed at 1/320 and the focal length 550mm. (Again, the actual product was a 20x16 canvas print sold locally.)

http://apaflo.com/gallery3/d4a_1212.s.jpg

Correct.



Ouch!

A JPEG is suitable for encoding a display image. It is not suitable as an intermediate format which will under go further editing.

But the display image can be any pixel dimension, and thus a print can be any size. (There may be some restriction in the standard, but if so it is beyond what we normally ever print.)

The distinction between 8 and 16 bit data doesn't really exist for making prints. In either case the file is converted to an 8-bit dithered bitmap and printed one entire horizontal line at a time rather than as unique pixels.

However, just how the workflow is managed does affect the results.

If the image file, regardless of its format, is not the correct pixel dimensions for the size specified it absolutely will be re-sampled (without additional sharpening) before printing. If the software/firmware used is good, and the size difference small, it will not have a significant effect. If the difference is large (for example a 2x enlargement) the image will almost certainly suffer from not having exactly the right sharpening. (If it's a good picture, it won't be noticed. But if also printed with correct sharpening it is also true that many people will easily see the difference when comparing the two side by side.)

That leaves a lot of room!

Currently I'm experimenting with a new 150-600mm Tamron, and at least on the D800 it seems to be pretty good. On a D4 is tends to search for focus, and lock up at one extreme. A slight twist on the focus collar and it recovers, but that is just not the way to work.

The Nikkor 80-400mm AF-S lens is nice, but 400mm is not really long enough for most bird shots.

Zoom is a great convenience. And the relative light weight of those two lenses compared to 400mm, 500mm or 600mm fixed focal length lenses is significant. I use the zoom for "casual" photography, where I'm not sure what to expect in the way of subjects and/or physical problems.

I'm often running around on a 4-wheel ATV on the Arctic tundra, so packing a large fixed focal length can often be too limiting. But that said, a 600mm f/4 on a tripod produces awesome images! Your mileage may differ vastly, and that is what you have to consider.

http://apaflo.com/gallery3/d8a_0082.s.jpg