If its not megapixels what is it?

[hjulenissen]

I'm saying that if you take an 80MP capture /and downsample it to 12MP/, that there will be a level of detail preserved that is not reflected in a native 12MP capture.  We're comparing two 12MP images in this case - one derived from downsampling, and the other not.  So I am not just saying that 80 > 12 here.

The extra level of detail will be high frequency information appearing near the Nyquist frequency of the 12MP capture.  The loss is occurring around where the slope of the AA filter on the 12MP camera kicks in.  The "more or less" that I referred to involves the extent to which detail in the native 12MP capture might be reconstructed using, for example, deconvolution.

But the "12MP" number is entirely arbitrary. You dont have a fixed-pitch 12 MP display or printer? The reason is that a 12 MP camera does not have the resolution of a hypothetical mental model of a "12 MP" camera that we have in our heads but that have never been physically realized. Once you are able to get rid of that model of how a 12 MP camera should have looked, and rather focus on how it actually works, then you are able to focus on what matters.

The important thing (?) is how stuff really end up looking on screen or on paper. A 80 MP camera could ideally have more details when rendered to some final format than a 12 MP camera. That is all that counts, is it not? The reason why it can, is that 80 > 12. To make things easier, we could call the 12 MP a "oomphaloomph", and the 80 MP a "dingalong". dingalong >= oomphaloomph... :-D The nitty-grittys that limit camera performance (AA etc) affect both the 12 MP and the 80 MP model. One could say that a 200 Gigapixel sensor was not enough because its AA filters and Bayer sensor limited it to a measly 150 Gigapixel equivalent resolution, but so what? As long as it had no perceivable and/or measurable benefits over last years 100 Gigapixel sensor (that was "really" a 75 Gigapixel sensor), who cares?

The Bayer sensor layout, a LCD monitor r-g-b stripes and an inkjet all have different technologies for giving a perceived color image with spatial details. The conversion from one to the other is non-trivial.

-h

[LKaven]

[...] The important thing (?) is how stuff really end up looking on screen or on paper. A 80 MP camera could ideally have more details when rendered to some final format than a 12 MP camera. That is all that counts, is it not? The reason why it can, is that 80 > 12. [...]

I can't tell if you are disagreeing with me or not.  It is more complicated than 80 > 12 as I indicated.

[Jack Flesher]

I know Michael. I guess the bigger is better is just not cutting it. Seriously folks I agree here since we both have shot these a lot already along with Mark and Jack we are just not that freaking nuts. There is a nice difference in this 180 back that i think the 4 of us can completely agree is very special in LOOK.

I think you guys are the nuts for still banging heads with the DXO/D3X zealots that come out of the woodwork everytime a new MF DB (that the four of us happen to like) is released.
 
 

[kers]

My idea for choosing a camera is the subject- what i want to achieve/show- in the endresult.
For some pictures 1 mega pixel is enough for others i might need 100 megapixel.
Sometimes 1 megapixel is even better.

My idea of "better- more beautiful' pictures with MF backs also has a lot to do with the lenses that can be as expensive as the backs.
I can see a difference in my d3x pictures from a cheaper lens or the more expensive PCE lenses. Not many people use those. It is good practice that Nikon and Canon are starting to produce better- more expensive lenses- needed to feed the 24 or more megapixel cameras.

coming back tot the A2 question.
My d3x still has still 243pixels per inch at A2 size prints - i know the camera has a very good dynamic range.
My printer can do 600dpi ( so ti says) - i can see the difference from 300 dpi with good glasses when i am very close to the paper...
but
I wonder if you can see much or any difference with a d3x shot with a good lens or the 80mp back- with always - a good lens.
I think slight differences in working with the digital image has a far greater impact on wich image is better..

( about HIFI- I noticed that even with my bad ears i can often hear the difference especially between cheap and expensive loudspeakers- you do not have to be an expert ear)

[LKaven]

[...]
I wonder if you can see much or any difference with a d3x shot with a good lens or the 80mp back- with always - a good lens.

One place to look is in the high frequency detail that is right up around the nyquist frequency for the output image (at it's final rendered resolution).  This is to say nothing of any kind of tonal fidelity you might gain from the large sample space of the high MP sensor, which is a bit harder to verify in practical terms, even though it might be very real. 

I personally think 24+ MP does wonders for rendering of scenes with high frequency detail, like trees, skin, hair -- any any output resolution.  In portraiture, you have the basic full-frame head shot, which fixes a reference point often used in making judgments about image quality.  The scale that we're operating on here is in the "pixels per face" realm.  The difference between a face with 12MP and a face with 24MP is salient. 

The nicest thing about the D3x is that it enters that realm of having just enough pixels-per-face, as judged by that one kind of standard.  And it is a very good implementation with a very quiet sensor.  I personally think that this is where the D3x earns a seat at this particular table.  With an 80MP back, you can get 24MP pixels-per-face in the context of a scene in which the face is a third to a quarter of the scene.  Now that's another standard to consider.

[The Sony K-5/D7000 design would look awfully good in 24x36 form at 36MP.  Imagine how it would look in 645 form.]

[vjbelle]

for A2 prints and Nikon D3x vs IQ 180 ? no not at all
but YOU know it by your own  I guess

As a matter of fact I don't know for an IQ 180..... I don't have mine yet.  But.... I do know for a P45 vs D3x at A2 and there is a discernible visual difference.  Maybe you should try all of this out for yourself.....

[mikev1]

I think the only way to satisfy everyone is with prints in their hand.  Or at least generate an even hotter discussion.

If someone is willing to prep two files, one with the IQ180 and whatever DSLR is deemed appropriate or at hand I'll print them for free on my Epson 9900 (I also have an IPF8300 but no profile for the paper I propose).

I am willing to donate free prints (A2 let's say?, approx 16 x 24) on a 44" roll of Epson Premium Glossy Photo Paper 250gsm.  You pay the actual shipping.  Shipping in Canada usually runs between $8 to $16 depending on the location and to the US is usually around $16 or so.  Anyone anywhere else I imagine the shipping will be a lot more.

I accidentally bought two rolls of the stuff and seeing how infrequently I get orders for prints on it, it will probably just end up sitting here for a very long time. And once the roll is gone it is gone.

I have no horse in this race.  I own a 1D mkIV and a Leica M9.  I used to own a 5D mkII and a 1DS mkIII as well as other canon cameras.  Oh I forgot I also have an NEX-5.

I own a small photo printing company up here in Canada and this is not some self promotion gimmick.  I have never mentioned my site on these forums and probably never will.  I am busy as I want to be right now.  My real job is raising a two year old and a four year old.  I have a lot of faithful clients both professional and amateur that keep me busy enough.

Why am I offering this?  I guess I'm just weird that way.  The outcome doesn't even really interest me all that much though I have enjoyed following this discussion.  I just don't understand people making claims about a camera they have never shot with or made prints with.  To be up front, I guess I'd have to say that I believe Michael and Mark's comments and would expect to see a better print from the IQ180.  So if that somehow makes me biased pressing Ctrl-P sobeit.

You guys can discuss politely how the files should be handled.  Perhaps a couple of other people can have access to the raw files as well to confirm anything.  I know a couple of well respected professional photographers (published articles, books, gallery, etc)  that I can ask to verify my end.

Again, I promise you this is not part of some great conspiracy to influence the debate one way or another.  I am not sitting on an island right now with Elvis and Osama enjoying a cold one.

 

[Erick Boileau]

As a matter of fact I don't know for an IQ 180..... I don't have mine yet.  But.... I do know for a P45 vs D3x at A2 and there is a discernible visual difference.  Maybe you should try all of this out for yourself.....

As I said before I can compare myself P45 & H1  vs 5d mark II vs 1Ds Mark III, I prefer MF but for other reasons than mega pixels

[hjulenissen]

( about HIFI- I noticed that even with my bad ears i can often hear the difference especially between cheap and expensive loudspeakers- you do not have to be an expert ear)

This has been documented in peer-reviewed papers as well, so it is not disputed.

-h

[hjulenissen]

I can't tell if you are disagreeing with me or not.  It is more complicated than 80 > 12 as I indicated.

That complication is usually irrelevant as I indicated. If you like, I can say that 72 > 9.5. Every problem should be described as detailed as needed, but not more than needed.

I agree that AA filters, the Bayer sensor, this and that limits the spatial detail that can be captured. Unless you can suggest a use-case where the output have to be exactly 12 MP, I simply dont see the relevance (for this line of discussion). You could say that a 12 MP camera is "really" 9.5 MP. And that a 80 MP camera is "really" 72 MP. 72 > 9.5. So what? MP is an abstract term for most people. The whole debate over Foveon vs Bayer and what conversion-factor Sigma should use when comparing the SD-X series of cameras to Bayer cameras shows that it is little more than a marketing term used to fool buyers.

-h

[bcooter]

I don't understand a single one of these comparisons, because to me it's apples and oranges.

I mentioned that my segment of the industry has changed and requires motion in parallel with stills, which requires continuous lighting, which requires higher iso which . . . (well you get the picture, pun intended).

That doesn't mean I don't enjoy or not not use my medium format backs.  It just means I use them less, so investing in a new one makes less financial sense to me than buying a new Epic, or 8 Kobald Brons, or well there is a long list.

On the flip side of this are the ultimate image quality group and I don't doubt for a minute they don't love their detail driven images and the devices they use to make them and most importantly believe they made the right choice,  but most if not all of them work in a different genre than I do.

Lately the word in advertising and editorial, is "real".  If you can show me a photographer that isn't presented with a creative brief every two weeks with the words "real" "approachable" "natural", "spontaneous" then it's probably just Karl Lagerfeld.

I hate to say it but flash rarely looks natural compared to mixing daylight and hmi's.   Slow hold that position photography isn't spontaneous and running back and forth to the digital tech station to check focus on a backlit scene is just another few minutes per shot that can be used elsewhere.

Smaller cameras with more focusing options are better for this type of creative brief and faster which allows more time for other considerations like the motion sessions.

Now if medium format offered the same attributes then I'd probably buy another one, but they don't and in a way I guess I should be relieved that I can hold on to the money or place it in other areas.

Everyone is different, including clients.  Actually for every client that marvels over the detail in a stitch or the fact we can capture a nose hair from a block away, there is another subset of clients that believe over sharp, over detailed images look artificial and digital, so once again, everyone is different.

But please keep in mind, this is just my experience today.  Tomorrow if the wind changes I'll change with it, because that's the only way to have success in this industry.  When the first client tells me they must have a 50 or  80 mpx file,  I'll probably be on the phone looking at hasselblads and working the best deal I can. 

IMO

BC

[gazwas]

Everyone is different, including clients.  Actually for every client that marvels over the detail in a stitch or the fact we can capture a nose hair from a block away, there is another subset of clients that believe over sharp, over detailed images look artificial and digital, so once again, everyone is different.

Like you said, every industry is different. I personally don't feel these new backs are just about more detail. For my interiors I was using a 1DsIII with live view and thought it was the best tool for the job until I tried my current camera a P65+ on an Arca camera. To most, the Arca sounds like a nightmare but with my first job I realised how much easier it was. Yes a little slower to set the camera up but once framed, two captures it was on to the next view. Slight camera movements or two stitched shots to make one wider view without the perspective or barrel distortion of using very wide lenses (Canon 16-35/17mmTSE). There was also so much more visible detail and exposure range in the P65 making it easier with lighting positioning compaired to the Canon. The less additional/artificial lighting needed, the more real the shots look and like you said that is in vogue at the moment. I used to do a 5 shot bracket to be on the safe side but now can be done in two or three captures, not to mention the retouch time after the job.

So with the possibility of more exposure range in the IQ180 and the freedom from a laptop on exterior locations it sound like a winner to me. I'm seriously looking at all the reviews coming in from new customers and considering selling all my Canon gear now to fund the upgrade.

Good times IMO.  

[DaveRichardson]

I am intrigued at the assumptions in this thread that seeing a difference with the newer high resolution backs must be some form placebo effect or snake oil.

If we look at this just from a mathematical view, an A2 print (16.5" x 23.4") at 360 ppi requires just over 50M pixels. This would suggest that 80M pixels is wasted at this print size. However the A2, 360 ppi, print actually requires 50M pixels red , 50M pixels blue and 50M pixels green.

With a bayer pattern of 1 row RGRGRG.. and the next GBGBGB.... etc this means that even with an 80M pixel sensor then captured resolution for the green channel is 40 M pixels the Red is 20M pixels and the Blue 20 M Pixels. Any "resolution" above that, with a single shot sensor and a bayer pattern, is down to software and interpolation. Why then should we be surprised that a difference can be seen with the higher resolution sensor. At A2 360dpi, 80M of bayer pattern pixels is still not enough to resolve each colour without interpolation.

I have to declare that I have not seen prints from the new sensor and that the above is theortical. However it does lead me to trust those that have seen a difference in prints and suggests a reason why such a difference may be visible. It also suggests why comparisons with a 24 MP sensor (6Mp Red, 6Mp Blue and 12MP Green) may be futile.

I am happy to be corrected if my assumptions above are incorrect.

Dave

[hjulenissen]

I am intrigued at the assumptions in this thread that seeing a difference with the newer high resolution backs must be some form placebo effect or snake oil.

I think that many have a desire for good testing methology that better answers: 1)Is there a perceivable difference and 2)What is the probable cause of that difference.

Some tend to jump rather quick to conclusions ala "I perceive a difference between this image taken at 80 MP and this image taken at 25 MP. The reason I perceive a difference must be that 25 MP has too little details". Just like some jump to the conclusion that CD must be a poor format because they prefer listening to a given LP over a given CD of the same release, without considering that the CD could be a different (re)master that was the result of a different artistic intent.

If we look at this just from a mathematical view, an A2 print (16.5" x 23.4") at 360 ppi requires just over 50M pixels. This would suggest that 80M pixels is wasted at this print size. However the A2, 360 ppi, print actually requires 50M pixels red , 50M pixels blue and 50M pixels green.

I think that equating pixels or megapixels at the camera sensor, within photoshop and at the printer output is bound to cause headaches. Luminance lp/mm or lp/ph is probably better.

1. Is there any proof that 360 ppi is really needed?
2. Is there any proof that current printers actually have a brick-wall response all the way up to 360 ppi for each color channel separately?

As printer have only one or a few output "bits" (either spit a drop of ink, or dont), they all (to my knowledge) use heavy dithering to give the appearance of 8-bit or even 16-bit gradations. This means that even if the distance between any two drops of ink is minute, the real spatial detail level is more limited.

The most important information is probably what is going on in the luminance channel. Humans are far more sensitive to spatial luminance than chrominance, and natural scenes tends to contain a lot more luminance details than chrominance.

With a bayer pattern of 1 row RGRGRG.. and the next GBGBGB.... etc this means that even with an 80M pixel sensor then captured resolution for the green channel is 40 M pixels the Red is 20M pixels and the Blue 20 M Pixels. Any "resolution" above that, with a single shot sensor and a bayer pattern, is down to software and interpolation. Why then should we be surprised that a difference can be seen with the higher resolution sensor. At A2 360dpi, 80M of bayer pattern pixels is still not enough to resolve each colour without interpolation.

Any resampling is going to be carried out using interpolation, and as long as the camera pixel grid does not match the minimum dot distance for your printer, you are going to have resampling now matter what you do. I dont see the evil in doing interpolation, the question is what the end-to-end spatial resolution is for luminance and chrominance, and how that looks for a given scene, viewer, print size, vieweing distance etc.

-h

[Fritzer]

My idea for choosing a camera is the subject- what i want to achieve/show- in the endresult.
For some pictures 1 mega pixel is enough for others i might need 100 megapixel.
Sometimes 1 megapixel is even better.

I beg to differ - in my opinion, more and bigger is always better, without exception .
Unlike film, in the digital world one can always create a 'poorer' look without much effort .

Of course, some tools are better suited to particular uses, or just too expensive, that's when one might want to sacrifice the original file's quality .

[bcooter]

1. Is there any proof that 360 ppi is really needed?
2. Is there any proof that current printers actually have a brick-wall response all the way up to 360 ppi for each color channel separately?

I think it's interesting we're talking prints in a digital world that's moving to mobile, handheld and lcd viewing, whether it's in publishing, in-store display, outdoor display or any form of image distribution, still or moving.

8 years ago our studios printing costs in paper and ink went from $45,000 to almost double that when you factor in drives, dvds, cds and hard copy delivery.

Now I doubt seriously if it's a 1/10th of that, probably less for actually more images produced.

Right now were in post production from a project where we are retouching 36 images that will run from outdoor to small web view and cutting 6 videos, (two are already finished).

From initial display, corrections, markups, client changes . . . back and forth, to final delivery there will never be one single print made, as we do it all electronically.

It's not unusual for us to finish a project that has over three dozen web galleries for client review due to post production changes and corrections.

Most of these images will end up in print, though if I polled most ad agencies and clients I would imagine actual high quality print on paper is probably less than 30% of their marketing effort or concern.

The print lasts about 6 months to a year, the online marketing seems to go on for years.

So knowing this I'm camera brand and format agnostic, which is a strange comment from someone that owns phase, canon, nikon, leica, RED, canon video and panasonic.

I've gone through the cutting edge of digital, been an unintended beta tester for a lot of brands and have come to the conclusion that the best camera for our studio is the one that is the most transparent camera, or in other words the system I don't have to think about or worry about, from on set to delivery.

I know a lot of people on this forum either love their brands, or have some relationship with the makers or dealers, which I think is fine because it's a form of commerce for them and I'm all for open commerce.

What I do find is that when it comes to making a purchase decision for my studios, 20% more detail or 5% more lens sharpness doesn't concern me.   Ease of use, a tested system, fast repairs, rentals in most markets and most importantly workflow mean a lot more to me than anything else, including costs.

I don't change still cameras quickly because I don't' have clients asking for any different file sizes, though with motion the 4k buzz is now taking over and that has become a request.

I will change computers quicker than any piece of equipment because nothing effects our day more than a slow computers and most of the software programs have become so heavy it takes a lot of computing power to run them.

BTW:  This is the minimum of what it takes to work with RED footage on set



Anyway, if I was (and may be) buying into another medium format system I'd only look two ways, or make that 1 and 1/2 ways.

I'd look at the IQ series with a Hasselblad H mount, only because of the higher resolution screen and an the fact the H camera is the medium format standard of our industry.  (especially in rentals).

And I'd look at the H4d 40 because I understand it does the cleanest 800 iso of any of the cameras at full rez and good skin tones.

When time permits I'll test both.

IMO

[Dustbak]

Lately the word in advertising and editorial, is "real".  If you can show me a photographer that isn't presented with a creative brief every two weeks with the words "real" "approachable" "natural", "spontaneous" then it's probably just Karl Lagerfeld.

I hate to say it but flash rarely looks natural compared to mixing daylight and hmi's.   Slow hold that position photography isn't spontaneous and running back and forth to the digital tech station to check focus on a backlit scene is just another few minutes per shot that can be used elsewhere.

BC

Funny you mention this. This is exactly what I hear a lot lately. 'Credible' is another one of those terms. Fortunately some people are moving away from the overly blown-out backlit scenes that everybody is currently doing.

Indeed, I get another MBP every 2 years and replace my MacPro every 3 years while continuously upgrading it (I am a small 2 person operation so my scale of operations is much smaller than yours .

BTW, I haven't been printing anything in years (not me myself that is).

[Professional]

The war about MF vs. DSLRs became popular these days more than Canon vs. Nikon.

Now i am just curios about MF itself, how you compare different system with same mp, say LEAF AptusII 12 against IQ180, or IQ140 against P40+ and against H4D-40? or say H4D-60 against P65+ and IQ160, i have H4D-60, and honestly speaking i should be the luckiest to have this camera because i am hobbyist 1000%, i think even some pros don't go that MF higher mp route, so for that i should be happy with it until i can be wealthy as some here to go for larger mp for want not need, and i will wait as my H4 is new, so maybe by next year or after i can have budget to go with newer model will be in the market that time, and hope that time this new IQ180 will not be outdated and/or not sufficient for applications in the future.

[fredjeang]

I don't change still cameras quickly because I don't' have clients asking for any different file sizes, though with motion the 4k buzz is now taking over and that has become a request.

Do you mean that RED is really becoming the "obliged" standart in commercial for clients and we can/should be preapared to through away the current 2k "Canons-and-Panas" in the garbage?

I was expected something like that happening but not so fast and reading your lines it seems that it's there now.

 

[bcooter]

Do you mean that RED is really becoming the "obliged" standart in commercial for clients and we can/should be preapared to through away the current 2k "Canons-and-Panas" in the garbage?

I was expected something like that happening but not so fast and reading your lines it seems that it's there now.

 

No, I didn't mean to imply that about 4k footage.  I just meant that mentioning RED and 4k from either my or the client's side is the only camera or format I've seen mentioned in years.

Now still digital capture is usually just takent for granted that most professional cameras from 22mp on will comfortably do the job.

(and before any medium format guys get their panties in a wad, that doesn't mean you shouldn't or wouldn't use something larger).

The RED is interesting to me because it shoots raw and has a thick file, less interesting to me because it takes more multiple steps in workflow.

But I guess if I have a point, for any image maker that works for commerce it should be understood that medium format still cameras don't just compete against each other or Canon and Nikon.

They compete for all professional spending dollars, like computer systems, the RED, continuous lighting, etc. etc. etc. and all this stuff is not cheap.

Even a 5d/7d outfitted properly can get into 10 grand without glass, so what was once our business of owning a few formats of still cameras and some lights has now moved on to broader but more expensive territory to compete.

IMO

BC