Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs β and Large Sensor Photography => Topic started by: geesbert on October 08, 2010, 05:05:09 pm
-
I am a commercial photographer specializing in food advertising. I shoot national and international campaigns for well known companies. I am working with Canon, but every few months the MF bug bites me. I got burned quite heftily with the Leaf AFi, which lost me a lot of money and a few pissed off clients, still i can't get me hands off those huge cameras.
This time I did a test comparing the Canon 1DS mk3 & 90mmTS-E with the Hasselblad H3D-50 with the HTS1,5 and the 80mm.
both lenses were tilted at their max, no shift.
The subject matter is just my groceries, but I tried to duplicate a couple of problems I am facing in daily shooting.
I am not telling how I postprocessed them, not do I care that people will moan that one cannot compare resized sRGB jpgs. I am happily and comfortably supporting me and my family with my work which nearly all end up either in CMYK on billboards or in magazines or on the web, so no one sees them in 100% on calibrateded Eizos, usually not even my clients who pay for them.
I had big troubles focussing the H. Live view is the joke we all know from MF, and with food one cannot run the model lights too high, so I got out a hotlight and lit the ceilling of my studio when I tried to focus that thing. Still the plane of focus is not equal in these shots, so beware when you compare sharpness.
Phocus is nice to work with, I just could get going without any delay, very well done. much nicer than that shitty EOS capture I am stuck with.
My server is seeing a lot of traffic recently, so I put those files (27MB) on rapidshare.
please let me know what you think, I will post the solution in a couple of days
DOWNLOAD FILES HERE (http://rapidshare.com/files/423915529/Canasselblad.zip)
-
I like games... ;D
-
Ahh, some fun.
I bet hassy = A.D.E
;D ;D
edit: quote:I am happily and comfortably supporting me and my family with my work which nearly all end up either in CMYK on billboards or in magazines or on the web, so no one sees them in 100% on calibrateded Eizos
So true!!
-
ADF
-
ADF
cool, now on my large monitor it's BDF for me.
It's the monitor that matters, not the camera :-)
-
I voted ADE, looking at the out of focus areas, noise characteristics, and detail in the focused areas did not look like Canon files to me. All at 100% of course...
-
If everything is shot the same.
A = mfd, better highlight detail and out of focus area is cleaner
D = mfd better overal detail
E = mfd, This was the hardest image to see any difference. I thought the highlights looked better.
-
Funny, I think it's BCF....
Ben
-
to keep DOF similar i closed the Hassy down on stop, so when the canon needed f5.6 the hassy went f8
-
I have no clue and have voted for, I don't know and I don't care :) I am pretty sure most of us will have it wrong anyway ;)
-
For this to be a meaningful test it would have to contain fine detail that could only be resolved by MF.
-
For this to be a meaningful test it would have to contain fine detail that could only be resolved by MF.
Those three setups were chosen to tackle common problems i encounter when doing my work:
A/B shows a lot of fine detail in the dirt of the mushrooms, the pepper, the wood and the napkin. certain elements in the picture need to be in focus, while others don't. how much more fine detail do you need? surface structure is extremely important in my kind of work.
C/D is aiming shadow details and chromatic aberations
E/F goes for colors and focus, when the subject doesn't have hard edges like I encounter with fruits or soup.
attached you see the setup
-
ADE for me. The colors seem more clean, the bokeh sharply more in compliance with the idea than I make of one MF. But I have maybe any forgery. I like the photo of the table-top in your last message.
PdF
-
My guess is A C E
-
My guess BDE
To me they all look very noisy and oversharpened and the levels are probably adjusted. If this quality is fine for your clients I guess you can go with either system.
IMO most of the IQ advantages are lost because of the settings/processing.
-
A
D
E
-
Why not... BCF.
-
My guess is ACF.
-
Ok, I think the main point has been made : identifying at those sizes, a 21 MP DSRL and a 50 MP MF is far from being an obvious affair, for this particular kind of subject . A good lens, good light, DOF placement possibilities and postprocessing are far more important than your choice of system (for these type of shots, at those sizes).
Anyway I would have been curious to see a comparison between the "best" of each system, for a "fuller picture" :) : a Hartblei 120 Makro-Planar Super-Rotator on the 1ds3 and the 100 HC with the tilt-shift adapter (it doesn't work with the 120 Macro) on the Hassy, would have (probably) made an even more interesting comparison. Oh , well :) ...
Anyway, hopefully the OP will now reveal the answers .... :)
-
B C E - for me. More noise, harsh bokeh, hot pixels, all attributes of digital back ;D
-
BDF for MF
-
A C F
-
Hasselblad files: B C F
Canon files: A D E
what does that mean to me? I really try to get that MF thing working for me, I really like big cameras, but whenever I try it either costs me a lot of money (remember my Leaf ordeal (http://www.luminous-landscape.com/forum/index.php?topic=37147.0)?) or it doesn't give me any more value than what I am getting out of my current system. Even more, I don't think I could comfortably work with what's on offer in the MF world.
On my Eizo the hasselblad files clearly show more resolution and I am probably more used to working with canon files, so maybe I messed up the H files, but still, up to today no one has proven to me that the hassle and expense of going MF would help me produce better images.
Often people say MF slows them down and thus makes them think. that doesn't help me in my kind of work when my herbs are wilting cause I cannot find the right focus. The Hasselblad previews appear nearly instantly on my screen, but then it takes nearly 13 seconds to get a 100% view. With the canon it is 7 seconds, which often feels way too long when you try to catch sauce running of a spoon.
The Hasselblad files take up about 5 times the space on my harddrive. People say storage is cheap, but still I prefer handling a shoot with 10 GB to 50GB.
Handholding the H, especially in portrait, with the HTS is impossible for a whimp like me, but it's quite easy with canon 1ds, with the 5d mk2 I do that 10 hours in a row.
than there is the question of backup. I think this little game has proven that a 5dmk2 is a proper backup to any MF, but when I promise and charge my clients to get them those yummy 100MB files I cannot use a canon if the main camera decides to cave in.
don't even get me started about cost...
If it works for you, I am really happy for you. For me it certainly doesn't.
-
;D
-
8)
-
...up to today no one has proven to me that the hassle and expense of going MF would help me produce better images.
Well, I got 'em right, so feel free to send me all your MF gear... ;)
FWIW, I thought the difference in these examples was subtle, yet fairly obvious at the pixel level. The fine detail is, well, finer... more structured/organized... something. I'm not sure exactly how to describe it, but I can see it.
That said, I think your earlier point about "no one sees them in 100% on calibrated Eizos" is absolutely true. Just because the MF files are better, that doesn't mean the Canon files are automatically inadequate for your work.
-
I am shocked, I chose ADE. With the recent posts showing 1Ds Mark III vs. Hasselblad, I am changing an opinion I've had. MFD is not better in the studio (unless you need the resolution) for any reasonable business need. However, I still hold the opinion that MFD is better for distance objects shot outside. If you do extensive retouching MFD seems to be easier as well. I don't know... Annie shot the Queen using a 1Ds Mark III and a 24-70L. Why can't we all just get along? Those of you with MFD, have fun with them, they are neat, I like the DOF control and so on. 35 users have nothing to fear. Seems to me we can all just use what we like and be happy.
-
I was right on 100%, being wrong that is. :(
I will admit I only looked on my laptop. I now need to go back and look on my other system. I have never worked with a canon camera so trying to judge the look of Canon files against the Hasselblad was definitely a blind test for me. I have been a Leaf Aptus user and also shoot with a Nikon D700. I tested the Hasselblad h4-50 a few months ago, and was ready to buy the system if I liked the images, and can say I did not like the look of the files.
-
For me this test confirms that the canon system is right for the type of commercial work (melting ice cream, running gravy, wilting lettuce etc.) shown. But it's not any kind of a test that could be applied to landscape work.
I couldn't see an immediate or glaring difference in the test images but I wouldn't take that to mean that a canon full frame image is close to a medium format system image for landscape work. That comparison would truly fall apart in a large (32x40 inch) print.
So I guess I'll take the test results for what they are. For me they show that different tools are needed (or at least better suited) for different jobs.
Claude Fiddler
wildernesslight.com
-
what did you all expect from a canon camera? to only produce s**t?
of course it is a very good camera.
it seems that some of you guys are looking for reasons to put something bad on digital medium format cameras.
but why?
-
I hadn't noticed that you had put the answer before I chose B C F as I thought the other files ran to their limit very fast, whereas the MF files could go much further if one still had to blow them up or manipulate them...
-
I hadn't noticed that you had put the answer before I chose B C F as I thought the other files ran to their limit very fast, whereas the MF files could go much further if one still had to blow them up or manipulate them...
Being able to manipulate the images much further than 35mm is one of the reasons for going MF. But I feel that the MF pictures look more natural the 35mm has a 'over sharpen' feeling to them.
The Canon delivers beautiful pictures and if you don't have to manipulate them, crop them insanely or make huge posters it will be the best choice because the Canon has much more features.
-
there is a visible MF advantage if you know where to look at! The war between MF and DSLR has always been done on details and pixel peeping but this is the last thing to look at.
On the attached file you can see the different level of Hasselblad in the creamy tone passage. On C and D files there is a clearly focusing error on Hasselblad shot.
In these days I have my HY6 e75 broken and have to use Canon (1DS MKII) ... the files look toys comparing to Sinar's.
-
If you need to 'know what you are looking at', if most photographers who use the equipment involved still got it wrong - why would we think the client can tell the difference? Surely the difference for the client cannot be enough of a selling point if only photographers can tell? It must be other factors such as enlargement, ability to work the files, skin tones etc. Or am I wrong?
-
On the attached file you can see the different level of Hasselblad in the creamy tone passage.
The fine detail is, well, finer... more structured/organized... something.
there is a visible MF advantage if you know where to look at!
that is exactly the BS we are hearing for years now, but hardly anyone really sees it, and no one does once in print. It's why I keep on testing this stuff. And it is why I keep on doing my commercial work with my Canon.
-
...and why he runs counterclockwise
-
that is exactly the BS we are hearing for years now, but hardly anyone really sees it, and no one does once in print. It's why I keep on testing this stuff. And it is why I keep on doing my commercial work with my Canon.
Wait, the quality that I used to correctly identify all the MF files is BS? How does that work?
-
How about plain luck?
why did 92,5% of participants choose something different? 15% chose the canon files, that doesn't mean they are better, it only means the difference is hardly discernable and me personally, I'd rather spend the money and the time I save by shooting a 35mm solution on holidays with my family.
-
Why luck? It was obvious the files were quite different, you can't justify giving your client somethng that you know in truth is of lower quality even if your client is happy.
That he doesn't pay you more for giving higher quality that is another question.
The day will come when you will regret having given out a job that you knew could and should have been of a higher standard being a professional...
-
Why luck? It was obvious the files were quite different, you can't justify giving your client somethng that you know in truth is of lower quality even if your client is happy.
That he doesn't pay you more for giving higher quality that is another question.
The day will come when you will regret having given out a job that you knew could and should have been of a higher standard being a professional...
Well Said !
-
why is it that important for you that canon/nikon cameras are close to mf...?
i think the answer is that you do only small or medium jobs where you cannot put the rental for a digital medium format system on the calculation, or even buy one for yourself. so it must be very important that your 35mm system is close to that.
next time when i have a hasselblad in my hand i do a comparison with a canon 5d mark 2 and use a focal lenght between 20 and 35mm (on canon) to a equal hassi lens. there will surely be no doubt which pictures belong to which system.
-
why is it that important for you that canon/nikon cameras are close to mf...?
i think the answer is that you do only small or medium jobs where you cannot put the rental for a digital medium format system on the calculation, or even buy one for yourself. so it must be very important that your 35mm system is close to that.
next time when i have a hasselblad in my hand i do a comparison with a canon 5d mark 2 and use a focal lenght between 20 and 35mm (on canon) to a equal hassi lens. there will surely be no doubt which pictures belong to which system.
Well Michael, if you go on that lands...MF has got nothing to do with the level of the jobs involved.
There are tons of masters with worldwide reputation that hardly use MF gear and tons of MF gear users that shoot "their garden trees"...
-
Why luck? It was obvious the files were quite different...
Gee... how was it "obvious" when only 7.5% guessed it correctly!? The only thing that is obvious is that the difference is so negligible (at least for this type of example) that only an extreme minority (if any) knew it, while most guessed it, some of which was then randomly correct.
-
Maybe it is just a reflection of the level of most users on this forum. I still say it was obvious...
-
Maybe it is just a reflection of the level of most users on this forum. I still say it was obvious...
Is that natural, or does the smugness come bundled with the MFDB?
-
How about plain luck?
why did 92,5% of participants choose something different? 15% chose the canon files, that doesn't mean they are better, it only means the difference is hardly discernable and me personally, I'd rather spend the money and the time I save by shooting a 35mm solution on holidays with my family.
How about I knew what to look for and I found it?
92.5% of participants choose something different because you stacked the deck. You wanted to prove that the MF advantage doesn't really show through under certain output conditions, and you did. Not sure why that would have ever been in doubt in the first place, though.
And if we're speaking personally, I don't shoot MF; hell, I don't even shoot 35mm. I'm not an evangelist here. I have no doubt that your decision to shoot 35mm is a valid one. That said, one CAN tell the difference, even if it is vanishingly small in these examples.
-
Why luck? It was obvious the files were quite different, you can't justify giving your client somethng that you know in truth is of lower quality even if your client is happy.
That he doesn't pay you more for giving higher quality that is another question.
The day will come when you will regret having given out a job that you knew could and should have been of a higher standard being a professional...
What a load of bunk, please tell the forum when you buy that new 80mp back, anything less is to dishonor your customer. Sorry to be to the point... it is a public forum.
-
Hi everyone,
In my view this is an interesting comparison. On the other hand the images are scaled down JPEGs, so the resolution advantage of larger sensor is eliminated. Would the MFDBs have an advantage in DR, it would not be possible to see it in JPEG. This is really an example of good enough.
So this demo puts MFDBs in a perspective, but doesn't in any way demonstrate their full potential. For that we would need actual pixel images, preferably in RAW.
Best regards
Erik
Gee... how was it "obvious" when only 7.5% guessed it correctly!? The only thing that is obvious is that the difference is so negligible (at least for this type of example) that only an extreme minority (if any) knew it, while most guessed it, some of which was then randomly correct.
-
... the images are scaled down JPEGs, so the resolution advantage of larger sensor is eliminated.
Perhaps the resolution advantage is eliminated, but there remain advantages to greater downsampling (averaging of noise to create a higher signal to noise ratio, for example).
What is interesting to me is that this "elimination" of MF's alleged advantage did not produce an approximately equal distribution of selections between medium and small formats. Assuming the math was correct, self-selected respondents selected the small format images overwhelmingly at ~92%. It would be interesting to know how the respective files were processed, as clearly something had an effect to skew the selection.
Said plainly, if tiny JPEGs are too small to see the "superior" image quality of MF, then why were the Canon images overwhelmingly favored at a rate in excess of 11:1?
Since the outcome between approximately equivalent choices (JBerardi and ced excepted, of course ;)) should essentially be random, it suggests one or more additional biasing factor(s) were in play.
-
What a load of bunk, please tell the forum when you buy that new 80mp back, anything less is to dishonor your customer. Sorry to be to the point... it is a public forum.
No disrespect to you but I trust ced to know about MFDB (and digital imaging in general) more than all of us put together, from developing it to using it in real life. I've personally learned a lot from the guy and if he says this is obvious then it is!
Yair
-
although I have an idea who ced might be, I am not trusting anyone's opinion who doesn't disclose their name.
thankfully I don't have to be paid by ced or the like, but by companies who are used to commission professional work. The whole production environment really makes the difference between professional and amateur work. and they trust me 100% in choosing the right equipment necessary for the job.
-
Said plainly, if tiny JPEGs are too small to see the "superior" image quality of MF, then why were the Canon images overwhelmingly favored at a rate in excess of 11:1?
Are you serious? The reason that you got higher contrast at 11:1 is that sharpening is added to the Canon e.g. on the mushroom in the upper left corner on A and B there is a dark edge around head at the transition to the steem on A - that is not how mushrooms look like.
-
The files fooled MF daily users, not dslr's only users. Yes they where downsampled and as Erik pointed, a Raw comparaison would have probably make the differences easier to spot.
If some experienced, trustable and above-the-crowd-users have reached the obvious solution, is that the differences where obvious.
If only 2 or 3 people could see this obviousness, is that the differences are unsignificant in most of the commercial assignments.
And, the normally loud crowd of dealers in this forum stayed curiously very silent. Don't they like games ?...or, maybe they where not sure 100% to get the right files?
-
The files fooled MF daily users, not dslr's only users. Yes they where downsampled and as Erik pointed, a Raw comparaison would have probably make the differences easier to spot.
If some experienced, trustable and above-the-crowd-users have reached the obvious solution, is that the differences where obvious.
If only 2 or 3 people could see this obviousness, is that the differences are unsignificant in most of the commercial assignments.
And, the normally loud crowd of dealers in this forum stayed curiously very silent. Don't they like games ?...or, maybe they where not sure 100% to get the right files?
However, this completely misses on how the files stand up to post production... shooting.
Yesterday at an event in London, I spoke to a photographer who bought his first H3D39 second hand a year or so ago. Much preferred it for fashion work.. felt he wasn't worried about losing highlights... shadow detail and so on.
Also found it much easier to work with for retouching.
All of those things can't be seen by a simple side by side comparison.
-
Sorry, I seem to have closed this thread accidentally by browsing it on my phone. this thread has gained me a lot of insight, so I hope we can keep on with it for a bit.
-
I think no one and definitely not me stated that DSLR files are as good (whatever that means) as MF files, my point is that as a commercial shootier in my field this quality advantage doesn't show.
What shows is that DSLR solutions are easier and quicker to work with, critical focus is easier achieved, compositional fine-tuning is way easier, financial and technical burden is way less and hardly anyone seems to notice the difference.
there is a lot of stuff that really improved my photography in recent years, e.g. using a dedicated 26" screen only for live view, or using those insanly expensive Fisso clamps to prop my subjects or using a second assistant just to keep the clients happy or swallowing part of the studio rental fee, just because both me and my clients are more comfortable and many things more. the tedious workflow of MF certainly work in the other direction.
The argument, that MF files leave more room for manipulation doesn't really count for me. My shooting environment is very controlled, it just can't happen that I blow my highlights or kill the black.
Maybe I should disclose my workflow for these files:
Hasselblad: shot into Phocus, edited there to get the look I liked, exported to 16bit tiff, into lightroom.
Canon: shot into eos capture, edited in lightroom.
then I tweaked the colors to equalise them. I only used global adjustment and very little of that. downsampling and color space conversion by lightroom.
using Hasselblad DNGs in lightroom wasn't successful.
-
... why were the Canon images overwhelmingly favored at a rate in excess of 11:1?... it suggests one or more additional biasing factor(s) were in play.
Additional biasing factor = Canon files look better? Just saying ;)
-
Are you serious? The reason that you got higher contrast at 11:1 is that sharpening is added to the Canon
Yes I'm serious--I'm just putting the question out there, as it seems to have been missed in the discussion. FWIW, I happen to agree with your answer.
-
Additional biasing factor = Canon files look better? Just saying ;)
:)
-
Yes I'm serious--I'm just putting the question out there, as it seems to have been missed in the discussion. FWIW, I happen to agree with your answer.
So boiling it down (and blatantly putting words in your mouth - sorry)
- If an equivalent post production (e.g. sharpening) was applied to the MF picture, there is a good chance that it would have been lost in the downsampling.
- It seems that there are two ways the pictures has been evaluated 1. The prettiest picture with the highest impact which favors Canon 2. The most natural picture that preserves tonality and fine low contrast detail which favors Hasselblad
The argument, that MF files leave more room for manipulation doesn't really count for me. My shooting environment is very controlled, it just can't happen that I blow my highlights or kill the black.
If you don't need to adjust you pictures that much and the resolution from the Canon is good enough then I agree that MF is overkill. But being able to shot straight into the sun and still get something meaningful out of the rest of the picture makes it fun to shoot MF
Maybe I should disclose my workflow for these files:
Hasselblad: shot into Phocus, edited there to get the look I liked, exported to 16bit tiff, into lightroom.
Canon: shot into eos capture, edited in lightroom.
then I tweaked the colors to equalise them. I only used global adjustment and very little of that. downsampling and color space conversion by lightroom.
I think it would be more fair to let Lightroom do all the processing as it can read Hasselblad 3fr and Canon raw files
-
... It seems that there are two ways the pictures has been evaluated 1. The prettiest picture with the highest impact which favors Canon 2. The most natural picture that preserves tonality and fine low contrast detail which favors Hasselblad...
Just as side note: Kodak, for ten years, insisted on a similar logic against Velvia vs. Kodachromes and Ektachromes: the latter were simply more "natural", vs. Velvia's "unnaturally" exaggerated contrast and saturation. Kodak was of course right, colorimetric-wise, but wrong marketing-wise, as tons of people favored the "unnatural" look.
-
My favourite photographer here β now sadly AWOL β uses Canon, but I could care less, heβs just very good and would be very good whatever he used.
I'm not AWOL Keith, shooting with Canon and Hasselblad.
Nick-T
-
I correctly grouped the frames into [ADE] and [BCF]. But I attributed each group to the wrong camera!
FWIW here were my observations as I examined the frames:
----------------
A: more colour-free details in the cleaver scratches; less blown knife highlights; whiter oof diffraction patterns
B: hot pixel in yellow background; chromatics on oof diffraction patterns
C: chroma noise, chromatics in oof diffraction patterns
D: speckly lumi noise, whiter oof diffraction patterns
E: more colour-free details in the sweets highlights; zero colour in white tablecloth texture; speckly lumi noise
F: more colour in oof diffraction patterns, sweets highlights & tablecloth texture
-> Hassy 50D= A, D, E ? [Wrong!!]
----------------
The interesting thing is that the Canon's optics had better chromatic correction (this shows regardless of the downsampling, sharpening etc.).
Sensor-wise, the H50D seems to have less luminance noise but more chroma noise (colour moire may explain this in the sharply focussed details, but it can't explain it in the oof regions).
Ray
-
The interesting thing is that the Canon's optics had better chromatic correction (this shows regardless of the downsampling, sharpening etc.).
I wasn't impressed by the HTS at all. the whole thing feels like one of those aftermarket solution (Mirex, Zoerk, Cambo) I tried with my Canon. It is much more wobbly than what I want for a lens. I think Hasselblad would have done better with a construction with dedicated lenses like the Arcbody.
The 90mm Tse is one of Canon's oldest EF lenses, it has been introduced nearly 20 years ago in 1991. It is still ok, but could be much better, seeing what the new 17Tse and 24Tse are capable of. I really hope a new one comes around, I expect it with the introduction of the 1DSmk4 next year.