We've just produced a pair of 40x60" prints comparing the D800E with Nikon 24mm PC-E and the IQ180 and the Schneider 35mm. The results are very interesting indeed - if you can make the hangout at 4pm Irish time, 11am Eastern, you'd find it worthwhile I think.
Add Neil McShane to your circles on Google Plus to receive the invite (https://plus.google.com/116143938603101885158).
For those of us that cannot watch, please post a written summary here for thread posterity.
Thanks in advance,
"Perfection is not attainable. But if we chase perfection, we can catch excellence."7 Member(s) liked this post
Yup, I'm not a Google plus user but would be very interested to hear your findings!
1 Member(s) liked this post
What is Google Plus? a new photo website.
Hi guys -
Sorry for the brief responses to this thread, but as you might imagine we've been incredibly busy with this test. We started yesterday morning at 9am, finished at around midnight and have been at it for about the same amount of time today.
We had the Google+ hangout and it was well attended, sorry if you tried to get on but found it was full.
I'll let you know the basic conclusion now, and will provide more details in a few days when we have the video live and the detailed article written up.
Essentially, at 40x60", the D800E with the 24mm PC-E lens is extremely close to the IQ180 with the 35mm Schneider on the Arca-Swiss. Close enough that at a 'normal' viewing distance of 4-5 feet, you can't tell which is which.
At close viewing, less than about 2 feet, the difference becomes apparent. But it is nowhere close to what you would expect.
In 20x30" prints, the difference is also apparent, but you have to be less than a foot from the print to tell.
I'm gobsmacked, to say the least.
Dynamic range is, as far as I can tell, at least equal between the cameras. Colour rendition appears to be better on the IQ180.
Bear in mind, the IQ180 we were testing is my own. We borrowed the D800 and D800E from friends and colleagues. So I'm giving this result as an IQ180 owner.
Presumably at 20x30, the IQ is being down-rezzed, and the D800E up-rezzed, whereas at 40x60, both are being up-rezzed.
How were the 40x60 files created?
More interesting (to me), is the relative lens performance. Anyone know how the Nikon 24 PC-E compares to the Canon 24 TS-E II? (wondering how an IQ180 on a HCam would perform here).
1 Member(s) liked this post
As a rule of thumb I usually think that going below 200 ppi for fine art prints is undesirable since one may start seeing pixel structures when nosing the print. If nosing quality is not needed (=wanted) we can quit doing high resolution photography right now. 10 megapixel is enough to make any picture good for "viewing distance".
Like most items with high quality, one really have to look closely to see any difference.
What would be interesting to know is if the D800 with PC-E 24 delivers sharp pixels all the way to the corner, and if IQ180 with the Schneider does the same.
Say if the Schneider is outresolved by the 80 megapixel IQ180 so the real resolution therefore is close to the 36 megapixel D800E that would be a shocking result. But if both deliver sharp pixels for their resolution then it is just up to the user to decide how many megapixels they need.
It is quite easy for any of us to test if we think we need 80 megapixels rather than 36, just do a crop test print of a imagined print size resulting in a certain ppi, look at it and see if we think it is sharp enough for the size. I've done that myself in great detail. What I find harder to test and find results for is if system XYZ can deliver sharp pixels corner to corner. It is one thing to have X megapixel sensor, another to have a lens that can make all those megapixels sharp so we get full use of them in a print. I was hoping to get that kind of questions answered from this test.
My guess would be that the PC-E 24 would not hold up 100% to the corners of the D800, and also the Schneider would have some problems on the IQ180 (therefore I suggested rodenstock digaron-w lenses).
1 Member(s) liked this post
For me, the D800 file for high frequency is limited to 250ppi for upsampling which means that the outer limit for high frequency images is 30 inches in the long direction..... that's it for me! This is image specific..... so if I were shooting architectural detail - likes windows or doors - I might be able to print to 40 inches, but the subject matter is lower frequency which now takes up most of the image. Anyway..... my two cents.
They will post a video here The Circle of Confusion
but it's not up there yet.
Zeiss Tessar design, super centre sharpness. I think the comparison shows both that at a technical level a $3000 camera can hold its own against $11000 of gear. It also shows that 18 vs 36 MP is not a great difference, and of course 36 vs 80 MP is not a great difference so by logical inference, 18MP is roughly equivalent to 80MP. It's not so ridiculous because a 36MP sensor can print at 60x40, an M8 can look great at 20x30 or even larger.
My point is that roughly the same argument has been going on since the 30's and reaching a fever pitch again, why has a 24 -> 36 MP change caused such a change? I guess Nikon realise Leica's clever marketing through social media and this is a full on assault using the same means. They just need a celebrity, Nikon equivalent of Seal and they're done.
Sorry but clearly it's all hyperbole to sell more cameras in market where many people are giving up entirely, using instigram on iDroid to upload pictures of their cat to facebook or switching to things like the Nex7/XPro1.
It's a shame Nikon failed to put the D4 sensor in the D800. That's a real upgrade I could have used. A D4 will also look pretty respectable at 60x40.
2 Member(s) liked this post
While I mostly agree with you, there are situations where every pixel counts. I made some very interesting observations during a job for a client recently. This is slightly but absolutely related, so I post it on this thread:
The client wanted some very large (200-240 cm longest side, that's a round 90") prints of landscapes in Norway (not my photos). They had very specific requirements with regards to locations, colours and style. That, and time as well as budget constraints, limited the selection of useable photos considerably. Most of the photos that were looked at initially (before I became a part of the process) were taken with DSLR cameras between 12 and 18MP.
Since I couldn't dismiss the selected photos without showing why, I made, together with the print shop, samples of how they looked at that kind of enlargements. What I noticed with all of them was that, while it's possible to stretch digital files rather far, when they "fall apart", it seems to happen rather abruptly. This confirms previous experiences with files from all kinds of digital cameras, including MF, but it's the first time that I've had the opportunity to test it out on several photos together with a very qualified printer.
If that is correct, it means that if one really needs those extra 10, 20 or 30 centimeters in a very large format (in this case to fill a certain wall space), 50% extra pixels may be very useful, be it a traditional DSLR or an MF DB. Have anybody else had similar experiences?
As an aside (off off topic):
The client had already bought a 35mm slide (Provia, probably 100 or 100F) for too much money which they insisted on using in a space where 220cm width was needed. I didn't really have much hope succeeding with that one, but the printer was of a different opinion. They scanned it to the required size, 240ppi at 220cm length, and printed it on matt paper. While the result doesn't hold up when studied from up close, it didn't "fall apart" like the digital files, and from 3 meters, which is the closest normal viewing distance for this particular photo, it looks very nice. Not perfect, at least not from a photographer's point of view, but the client is very happy.
>While the result doesn't hold up when studied from up close, it didn't "fall apart" like the digital files
Interesting. How did you upsample the digital files?
Editor&Owner of Digital Outback Photo
Canon has already come out with its version II of the 24mm and it is drastically improved.
Here is it's corner performance compared to the Nikon 24mm PC-E.
Nikon 24mm PC-E
Canon TS-E 24mm f/3.5 L Version I
Canon TS-E 24mm f/3.5 L Version II
The progress being made is very significant.
I wonder how the test would have gone with the D800 with the Canon TS-E 24mm on it. Maybe a match for the IQ180 or at least even closer.
Nikons next releases of TS lenses will most likely be on par with Canons current offerings.
These are test shots with the lenses wide open.
Last edited by FredBGG; 28th May 2012 at 11:21.
Is it possible to use Canon TS-E 24mm f/3.5 L Version II on D800 with adapter?
Last edited by subrata1965; 27th May 2012 at 22:18.
They did predict the outcome in beforehand, for film as well as for digital. They also run the best photolab in Bangkok btw., and do all my film processing. Their work is so good that I don't really consider doing that myself, even if it would save me a fair amount of money.
As an aside, going from the examples above, the Nikon PC-E seems squarely in between performance wise compared to Canon's MKI and MKII offerings. I wouldn't call it horrible, either, just average...
To keep this relevant to MF, how do those that own the 80MP backs feel about the findings here? Jack?
It'd be amazing to see some scans of the prints at scale too, if you have time, Peter.
Does anyone out there use Perfect Resize 7 to upres photos? It used to be called Genuine Fractals, I think. I wonder how using such a program would influence a test like this.
Digital files does not have that charm at all, pixel structures are not nice, and I'm not particular fond of the look of fractal upscalers and the like either. Therefore I personally think that with digital the need of high resolution and overall high technical quality is higher than with film.
1 Member(s) liked this post
But I think your point raises another, perhaps more general, question.
This is a test of systems, not just of sensors. Was the best available lens actually used on both the D800 and the IQ180?
Personally, I'd love to see a test between the following:
Best c15mm glass possible for the Nikon on the D800.
23HR on a tech cam on the IQ180
24mm Canon TS-E on the HCam on the IQ180.
1 Member(s) liked this post
Results?The IQ180 will be used both on an Arca-Swiss Rm3d with a 35mm Schneider APO-Digitar XL and a 645AF with 35mm Phase One f/3.5 lens.
+1 for perfect resize. The most common mistake I see (especially regarding Lightroom) is oversharpening the images before upsizing. I recommend Nik Sharpener. It's adaptive sharpening algos work pretty well.
It is also very close in resolution to the IQ140 and similar products. That little bump in megapixels has made comparison with MF systems much more interesting than before. We may also come to a tipping point where a large part would-be MF users find DSLRs to have high enough resolution. Time will tell, but the D800 could really be a revolutionizing camera... a little bit like when photographers shooting 4x5" film eventually went digital because it reached a tipping point when it became good enough (it was somewhere around P45+ for many it seems, i e similar resolution to the D800), but this time around it is MF digital shooters that may move down to DSLRs.
Photography is all about experimentation and without it you will never learn art.
www.guymancusophotography.com2 Member(s) liked this post
But of course, more megapixels is a good thing and important for some. I would be more than happy to own one of those beasts (MF DB or D800), but do I need it to take acceptable photos?.
I don't care what gear I have.
Things I sell: http://www.shutterstock.com/sets/413...html?rid=611051 Member(s) liked this post
I am trying to get excited about the D800E, but I'm not quite there as most of those here. Don't get me wrong, it is the best 35mm DSLR I have used to date, and I have had 1DSIII, 5DII, 5DIII, M9 etc.
I am looking at files out of the IQ180 w/80LS compared to D800E with 50/1.4G, and the IQ files look better, even on screen. The skin tones, and general tonality have a smoothness that I can only describe as magical.
The D800E is hands down the more convenient tool to use, and I want it to replace the IQ for this reason alone, but I cannot see it doing that just yet.
Perhaps it's the larger format. For now, both are going to be hanging around.
I do see the Nikon replacing my Canon kit. The Nikon flash system, as far as I can tell so far, is superior to Canon as well. But I have so many Canon lenses and accessories that I am not sure this is going to happen either.
For now, I am going to use the D800E with 14-24, 24-70, 70-200 as a convenient landscape kit when I don't feel like lugging the IQ around, or when I'm in locations such as my kayak where I don't want to risk dunking my phase kit. YMMV.
1 Member(s) liked this post
I'm still a strong supporter of MF and the IQ from it. That is just not going to go away. I just have to make business decisions that I really sometimes don't enjoy. I like the Nikon a lot it does the job and it does it well so I'm not throwing fits on this decision as it is fun to shoot and it's doing a very nice job on the IQ level and it's good enough in many ways. I can see people in almost any decision on this as relevant as the Nikons certainly are damn good now.
One big issue that plagues the Nikons that I can't find a way to get around is wide angle and movements. Here the tech cam smokes it but than my again my 200 F2 smokes the tech cam. This makes it not a easy decision for a lot of people on there needs. It's really a personally decision on what's going to work the best for your type of shooting. So I can see a lot of tech and MF shooters buying in to fill in the gaps and vise versa. Not easy decisions folks. I'm here to help if anyone needs it but first I'm off to go play golf. ROTFLMAO
Enjoy the day and Happy Memorial day!!!
Lol I went and bought new irons yesterday so I'm off to go play a round with my daughter, her Boyfriend and his dad. I realized my clubs where 8 years old no wonder I can't get my handicap down. We talk about the tech here in photography but the tech in golf is amazing as well.
Have fun let me know if Luke's swing technique worked. Lol
I agree with this as well, and I see this as a printer. Digital files have a higher inherent sharpness due to their lack of noise, but the very noise that obscures the fine detail in film is also what makes it pleasant to look at when it is printed large. It hides those little imperfections better...lens softness and aberrations etc, and it does not have the digital artifacts that makes digital more problematic at 100% (or greater) -- the aliasing, bayer miscalculations, more severe chromatic aberrations and color fringing etc. As far I as am concerned, that all else being equal, film is better for really stretching enlargements, but digital usually looks sharper and cleaner at moderate sizes. The fractal enlargement programs will increase the sharpness at larger sizes, but I don't think it results in a better looking enlargement on the paper...quite the opposite, actually. I think a better technique is actually to do something to mask the errors, rather than try to increase the sharpness. I find that I can enlarge the S2 to 1.5m on the long side without a problem (about 120ppi), but if I were going to go bigger, I would probably add a bit of noise...sacrifice that last bit of detail for a more pleasant character.
You also mentioned the importance of lenses earlier, and I think that is something that gets mentioned a lot, but cannot be overstated. The higher the resolutions get, the more important it is for the lenses to outdo the sensors. That 36-40mp+ will reveal ever little problem with your lenses all the more clearly. If you are going to see a real advantage from it you need a lens that will max out your sensor.
A quick note on the 2m panoramas, however...I think 35mm would be quite a stretch for that! I just did a bunch of 2.5m panoramas (also for a Norwegian!), but they were scanned from 6x17...they hold up extremely well, but 35mm for a 2m enlargement...well, it might have an interesting look, but it would be essentially all fuzzy grain clouds...at least until you were about 5 feet away.
I would have thought the Mamiya 300/2.8 would give a similar look to the Nikon. It and the Nikon 200/2 have very similar physical aperture sizes and fields of view on their respective sensors.
I've quoted these previous bits:
First off, some background. I know Peter and we both went to MF, and ultimately to the same back and the same tech cam for the same reasons. I emailed him before his test and told him what I suspected he'd find --- that conversation went something like this:
Me before the test: "I am really looking forward to your conclusions too. I of course did some tests for myself, but honestly wanted some other confirmation before going public with them. The IQ180 is net superior on detail and color, but for most shooters probably not nearly enough of a gain to justify the 7 or 8 times greater investment in gear and the 4x more complex LCC capture and processing routines. The D800 has slightly superior DR also, which is impressive on its own."
Peter after his test: "Yes, we came up with the same results, although from our testing I'd say the two cameras are equivalent on DR. The margin on image quality and colour is very close for all practical considerations."
IMHO Guy hit it on the head when he said the D800 hit a tipping point in resolution that made it a viable MF alternative.
My more generalized answer would currently be something like this:
1) At a 32x43 inch (80x105 CM) print, the IQ180 file can be printed native at 240 PPI. The D800 file needs a marginal (and easily accomplished) 140% linear uprez to print that same size and resolution. Net result is the IQ file is going to have visibly smoother tonality, but -- and this is a big but -- you'll have to have your nose in the print and to see it; viewed from normal viewing distances, they will be surprisingly similar and equally good.
2) If you now go to a 60x80 inch (150x200 CM) print, the IQ will look notably smoother at any viewing distance less than 3 feet (1M), but still probably not all that visibly superior at normal viewing distances.
3) At print sizes 24x32 inches (60x80 CM) or less, you will not see significant differences unless you put a loupe on the prints to compare them.
I will add a few final comments.
DR: I stand by my claim of slightly superior DR out of the D800. I am not talking strict engineering definitions of DR, but practical extractable DR from the files; the D800 is so superior on noise that the shadows can be pumped a *lot* before becoming unusable. To be sure, the IQ180 shadows can be pumped impressively well too, but the D800 goes to the same level with less noise.
Noise: Per the above, the noise characteristics of the D800 are impressive. We've all believed since the beginning of digital history that bigger, fatter pixels will always be superior on noise than smaller, tightly packed sensors can manage -- and that's been mostly true up until now. Here, both the IQ180 and most especially the D800 prove that old belief flat out wrong.
Ease of use: No question, the D800 is a dream for rapid capture and quick, easy processing.
Cropability: No question, the IQ180 has a huge amount of room for after capture cropping to desired composition -- you can crop it 50% and still have a 40MP file to work with.
Color: For those needing exacting color accuracy, like for product imaging or art-reproduction, the IQ180 will win by some margin in most critical applications. This can perhaps be improved for the D800 with better profiles to the point it's irrelevant, but only time will tell.
And there you have it, my simple two-cents worth.
I've done some DR testing on my own, and my conclusion there is that the MF backs are good enough - the extra DR gain you get from D800 and other sony exmor sensors are into more or less irrelevant shadows where photon shot noise is a bit high anyway. What one would want now is higher full well capacity, that is increase dynamic range into the highlights so we can start dropping gradient filters / HDR for real.
If you're for example a fine art landscape shooter I don't find the tech cam be much more complex than my Canon system. What makes things complex is making fine compositions and capturing them in high resolution. Maybe I'm a bit slow to decide about the composition but I find that the LCC adds about 10 seconds to a 20 minute workflow, and gives you the added bonus of dust removal. In really bad conditions like very hard wind or poor tripod footing or shooting above my head or very low light I prefer the Canon though.
The real problem with MFDBs as I see it (from a landscape shooter's perspective) is not so much the performance or complexity, it is simply the crazy high cost of the backs themselves. It is just too much for too little. The cameras and lenses are expensive too yes, but not crazy expensive, it is more like "you get what you pay for" feeling there. MFDBs are now very close to be worse at everything except resolution, and still cost many times more. It won't work forever.
If it really is true that MFDBs cannot be made cheaper due to large chips and low volumes I think that some of the companies may go out of business the coming years.
1 Member(s) liked this post
Newer and newer DSLR is getting better and better. I owned Sony A100,A700, A900 to A77 (APSC which is actually better than the older FF model) and I can see the IQ is improving.
It wont be long before DSLR will reach to a point where you can't justify the price-performance ratio of a DB anymore. Unless DB comes with something new, amazing DR perhaps We also need a new body with great autofocus/image stabilizer for starter.
I hope the comparison will give us input on DR, clarity, details, color, 3D and depth, file latitude, etc.
Maybe in a few years I have to sell my IQ
However, I respectfully disagree with you 100% on this comment and stick by my original statement: You need to get your face in the prints side by side to see the differences. Sorry, and I know it's going to be an unpopular comment for many MFDB owners, but I have done the print comparisons myself with both processed and optimally printed, and found the difference is minor even on close inspection, advancing to undetectable at normal viewing distances. I will go further and state for most photographers printing 40 inches or smaller, the price-performance ratio only makes sense if you need absolute correct color and then have a client to bill appropriately for it... One of my oldest and most basic business axioms is, "Sometimes 'good enough' is."240ppi, which is equivalent to a 43 inch print, produces a stunning print which cannot be matched by a 36MP file..... although at normal viewing distances they will be similar.
"Perfection is not attainable. But if we chase perfection, we can catch excellence."1 Member(s) liked this post
I agree almost 100% with you. Mind you that my comparisons were done with downsampled IQ180 files and then printed to 40 inches and compared to the original 80MP file. I'm looking at high frequency city-scape stuff that shows a difference but not until you are very close to the print..... but I'm real, real picky. So, I can only apply my standards to me and certainly not anyone else. Certainly there is never going to be a price-performance ratio that makes any sense..... I believe the MFDB model is completely broken and the beast is destined for extinction.
Last edited by vjbelle; 28th May 2012 at 10:43.
I've done tests on the D7000 in terms of noise which per pixel is almost exactly the same as the D800, my results here:
An image sensor noise test
Did not have an IQ180 to compare with though, just my old Aptus 75... yes I have looked at RAW files and done very careful comparisons with exact exposures, and yes the exmor sensor is noticably less noisy.
*But* to make any real difference it must be so much less noisy that the extra dynamic range actually changes something in the workflow, and if you used gradient filters in backlit scenes on your MFDB before you will most likely not stop doing it with the D800. It is not that much better. Getting rid of almost all read noise is great, but shot noise is still there. the only way to improve that further is to get better full well capacity - i e allowing for longer shutter speeds without clipping.
To truly make gradient filters / HDR obsolete for us landscape photographers I would say that you need the noise performance of the D800 pixels plus two more stops up to saturation, so you can push with less photon shot noise.