Is the difference of DR on MFDB vs 35mm dslr discernible on print?

[cmi]

Mine is better than yours--that is your reasoning, not mine. I've tried to use reason and analysis to explain what we might or might not be seeing. Thus far to no avail and answered only with ad hominem attacks, which are the resort of those who are losing their side of the argument. Whether or not I have taken any pictures is irrelevant, but I expect my analysis to be answered with reason, not invective. Instead, you ignore my findings and I suspect that you can not rebut them. Thus far, no analysis from you but only unsubstantiated assertions that you think we should believe since you have high end equipment.



I don't doubt that you have the equipment that you claim, but if the results are really so much better why don't you demonstrate that? Thus, far your shot of clouds is uninspired and proves nothing. Personally, I prefer Bernard's images, whose quality has been demonstrated in multiple posts and is widely admired and your denigration of his work is out of place on this forum as are your ad hominem attacks. The quality difference you may be seeing in your clouds is likely explained by considerations other than dynamic range or perhaps your inability to process images from your Canon, which I am sure of capturing the limited dynamic range in your cloud shots.

Assuming Jack Flesher is right it would not be his problem to prove it. I mean this is no rocket science. Anyone who shoots regulary with P65 and D3X/1ds3 could tell easily. Frankly I trust more in the experience of a pro who knows what these systems do rather than theoreticans with raw graphs and no access to these cameras. Nothing against technical analysis! But it has to relate to the experience of a working photographer and not to an engineering standpoint. In practice the final image is relevant nothing else.

And after all, a proper explanation has to stand on it's own and be graspable by common means. But what I see here is different. Only technical rebuttals of single sentences, not really own standpoints. I mean IF there is no difference between DR of MFBD and 35mm I expect an explanation wich makes sense of the issue as a whole in understandable words. I want it explained fully then, not these silly raw graph snippets with partial explanations. This are not serious explanations or proofs. A proper explanations has to make sense of all the facts from all sides.

At the other hand as an outsider the fact that this thread goes for so long makes me wonder if there is really such a significant difference. I mean after all if the difference is big it should be really obvious it should be confirmable by a plethora of shooters here, no? Or maybe just nobody views threads with too many pages and certain topic titles?

[Guillermo Luijk]

I have a problem with your method; averaging the intensity in the channels does not yield anything useful regarding the camera. I don't know with which ISO the image was created and how it was processed, but this blotchiness is not characteristic to any mid-range DSLR at -8 to -9 EV with lower ISOs, except perhaps the A900 with the built-in noise reduction.

Can you make this raw file available to me, or is it too private?

I agree the method is flawed; I averaged just to get a quick 'luminance' to locate each zone someway. The correct way to do it is consider each undemosaiced RAW channel separately, but I understood this would have been too heavy for a thread where trained eyes account. I consider my approach valid enough for the purpose.

It was ISO100. Luckily the original file is still in the download server: HASSELBLAD DNG DOWNLOAD.

Would you be able to tell us what software you used to process the RAW file and where it came from?   I still contend that any fair analysis will include the manufacturers supplied software and files will be evaluated post processing not prior.

I used DCRAW with the standard AHD demosaic algorithm, because is the software I always use so it's the more adequate for me to compare. However if you wait for a while, surely Gabor (Panopeeper) will do his analyse on the undemosaiced RAW data.

I trust more in the experience of a pro who knows what these systems do rather than theoreticans with raw graphs and no access to these cameras.

You are free to think that way, but assuming that being a professional photographer means knowing how to take gear to its best, is not only not necessarily true, it is in fact usually false. Quality is rarely aligned with efficiency.

BR

[Phil Indeblanc]

Are you unable to get the results you are striving for?

If that is the case, rent one, and end your uncertainty.

If you don't know and are grasping at straws of a question, then take the mind set of a perfectionist and see the areas where your quality is lacking. Try different lens'. Try different apertures, try different times of the day.  Spend the money on great glass. Use a adapter and try a different make glass if you are not sure. Try a SLR without a AA. all these will be around $1k-1500 in price.  They are the expense of a learning shooter that is striving to be pro or perfect his/her craft/art.
When you see something missing, as you mix print into the thing. Try and see what another shooter that execeeds your abilities is doing.  you might be surprised to learn how much time and effort goes into a fine art print. Maybe you are already making them, and just have run out of work or are so well off that you want to find that silver bullet.  

When you have a reference target of a certain quality, you know what you are after. If your own quality is not in question, who's stomach is empty?

If you have used film on LF, then you will know what meat you want to eat. Then you will know what quality you are after.  science can only measure what definitions and laws that are already in place. Some things are not defined, yet. A lot of what we do is based on calculations. But when you see beauty, and you have a vision in your head, you will use calculations, but if you dont have vision in your head, the calculations will limit you.

if you want to do a scientific test to see the difference, that is great, but that is a little different from your OG post.

I also think the subject matter that you shoot is somewhat of the utmost important, and I highly doubt everyone here is shooting the same subject.  I am sure there are event shooters here that have no need to discuss this, and even in landscape the very demanding landscape shooters...the subject is so organic, it is harder to make such apple to apple tests.

The answer will also get plugged up sometimes as so many have poured so much time and money to discover the "sweet spot", that people dont feel like sharing it, and I dont blame them. You go spend years of your time and hard earned money to strive for the best that you seek, and then dump it online.  Well when you have people fueling and flaming each other like this, perhaps the water will boil sooner :-)  Michael's LL testing should be updated and he should supply a new CD/DVD perhaps, with the latest canon nikon and PhaseOne sensors, as he would be the person to correctly execute with the apples to apples gear in the apples to apple situation and raw data.  How about it Michael ?

[Phil Indeblanc]

I never made any assertions about the "usable" DR of prints. I simply pointed out that your excuse for why the D3x comes out on top in the DxO test was inaccurate.

Setting aside for a moment the definition of usable DR, I can actually find it quite believable that the MFDB have some advantage in DR. However, I think the difference is much smaller than many MFDB owners claim. The conventional wisdom that MFDB's have a huge advantage in DR may have been true a few years ago; but the CMOS sensors in DSLR's have been improving at a faster pace than the MF CCD's.

14bit vs 16bit. each bit is acounted for about 1000 levels in each color?  I dont know the exact #s, but I know this to make a difference

[Panopeeper]

14bit vs 16bit. each bit is acounted for about 1000 levels in each color?  I dont know the exact #s, but I know this to make a difference

The 16bit is a myth. Although the actual raw data is 16bit wide, no more than 14bits carry useful information. In fact, in some cases (for example P45+) using the least significant two bits results in awful image; for example the "centerfold line" becomes visible.

However, this does not take away anything from the high quality of some DBs.

(Note: the bits are not accounted for 1000 levels per bit. One bit accounts for only two values; the combination of the bits yields the high possible number of levels.)

[TMARK]

Honestly, who cares if Irving Penn and Paolo Roversi and who ever gives a shit or not about this subject? This is meant for those, who actually understand what it is about; for those, it is about ohotography. For the digitally illiterate, it is nerd talk. I see no problem with that.

If you don't understand it, get over it. There are many threads I don't understand much of, but I don't post in each and every thread, that the others are nerds, because they know something I don't.

Because, "Gabor", this is Digital backs and Large Sensor Photography.  This inane line of name calling may involve digital backs and large sensors, it has nothing to do with photography.  Isn't there another section for this type of "discussion"?  That's my question, can you take it somewhere else, perhaps away from the digitally illiterate? Or at least stop condescending to people?

[Ray]

Don't know what all the fuss is about. Jack claims his P65+ has better DR than his 1Ds3, which enables him to get smoother gradations in the bright area of clouds without introducing unwanted noise in the darker parts of the image, and DXOMark test results indicate he has good reason to hold such a view.

At a normalised image size of 8mp, the P65+ has a whole stop better DR than the 1Ds3, according to DXO. In my experience, a whole stop of DR difference is significant.

At the same normalised size of 8x12", the Nikon D700 has about the same DR as the 1Ds3, according to DXOMark. Here's a shot I took a few months ago with my D700 demonstrating some cloud effects. The camera's pointing directly at the sun which is behind the clouds but almost peeking through, causing a few very small patches of blown highlights.

The foreground in this image was quite dark and needed lifting during processing. It's interesting to note, if I'd used a D3X for this scene I would have had an additional 1.67 stops of DR available. Instead of the exposure of 1,000th that I used for this shot, I could have used a 3,200th of a sec on a D3X, at the same aperture and ISO, without compromising the quality of the foreground.

[attachment=16687:0251_crop.jpg]

[Panopeeper]

This is the per-pixel SNR seen in a 100% crop in the deep shadows. Like on any high end DSLR, noise appears at -7EV (i.e. in the 8th stop from saturation), and is very visible at -8EV (the 9th stop from RAW saturation). That provides about 9 f-stops of usable DR for photographic applications.

Guillermo,

I took a look at the raw image. The noise in the crop you posted comes exclusively from the red channel, which is at -9.65 EV, i.e. more than 10.5 stops.

Btw, the noise of the Hasselblad H3D II-31 is between 1/3 EV and 1/2 EV less than that of the 5D2, with ISO 100 (i.e. its dynamic range is so much higher).

[EricWHiss]

Guillermo,

I took a look at the raw image. The noise in the crop you posted comes exclusively from the red channel, which is at -9.65 EV, i.e. more than 10.5 stops.

Btw, the noise of the Hasselblad H3D II-31 is between 1/3 EV and 1/2 EV less than that of the 5D2, with ISO 100 (i.e. its dynamic range is so much higher).

Great so is this the end then? We all finally agree that MFDB has more DR than DSLR's?  Right?  

[Jim2]

When performing a 12 images stitch with a d3x (3 rows of 4 images in vertical orientation), I am de facto simulating a sensor that is approximately 85x65 mm, and the enlargement ratio for a given print size will therefore be about 1.5 times smaller than that of - say - a P65+.

Cheers,
Bernard

Hi Bernard, what gear do you use to do 3x4 stitching? Special tripod heads, etc? How long does it take to execute it, and in the case of taking photos of quickly changing scene (sunset) will that be practical? I have always been a single frame shooter and have only tried stitching once or twice with a very basic tripod head. Perhaps with the right gear, stitching would be easy and quick to do.

The technical camera's ability to shift the back sensor to allow stitching seems appealing to me as it seems very easy and very quick to do. This can be achieved with either MFDB or 35mm DSLR.

I guess ideally for me is to just get an MFDB + tilt / shift body and try it myself to see which is more practical, but nobody rents it out here where I live. I guess I need to go to Sydney to rent the gear, and the rent is expensive last I looked the MFDB itself costs $600/day + I need to rent the body and lens.. and probably tripod head and a bigger tripod lol. Add that to the expense of flying there + car rental + hotel etc it would add up to about 2K-ish.

Furthermore it will be hard for me to do a fair comparison since I am used to using my 1ds3 but I have totally no idea how to use the MFDB + its lens + the technical camera. It would take me some time to get used to working with it and probably a lifetime 'mastering' it. Furthermore, there's that article that says that there are big variations amongst the copies of the same product, and usually the rental gear is even worse presumably due to abuse or lack of TLC?

I guess at the end of the day, ideally I'd like to know from several people who are used to using both P65+ with a technical camera and 35mm dslr (1ds3 or d3x) to come out and comment about the comparison of equal size prints (e.g. 30" x 40" or bigger) of the same scene taken at the same time - using P65+ (or other kinds of MFDB) vs 35mm (perhaps being stitched to make it 'fair' on resolution).

I know that several people have voiced their impressions although they cannot explain in technical speak (which is probably not necessary?) and I thank them for that.

So far I haven't seen any opinion to say that a 35mm dslr (stitched or otherwise) print looks better or even of the same quality / beauty (for the lack of a better term) than MFDB on large prints - in case if I missed it, please kindly correct me.

Thanks for all the posts so far. It has been very educational to me.

[Jim2]

Can't imagine where you get that from, but it sounds like you need to shoot a P65+ and a D3X for a weekend so you can see how things come out when the rubber hits the road.

Doug

Doug Peterson
__________________
Head of Technical Services, Capture Integration
Phase One, Leaf, Canon, Apple, Profoto, Eizo & More
National: 877.217.9870  |  Cell: 740.707.2183
Newsletter: Read Latest or Sign Up
RSS Feed: Subscribe

Doug, Has anyone at Capture Integration ever done such comparison (perhaps with 1ds3 if you guys don't have D3X)? You guys would be the ideal people to do this having familiarity with both systems. The end result to compare is on a large print, viewed from a decent distance (not pixel peeping) as one would enjoy a beautiful photograph on the wall.

[Panopeeper]

Because, "Gabor", this is Digital backs and Large Sensor Photography.  This inane line of name calling may involve digital backs and large sensors, it has nothing to do with photography.  Isn't there another section for this type of "discussion"?  That's my question, can you take it somewhere else, perhaps away from the digitally illiterate? Or at least stop condescending to people?

Tim, apart from the fact that I have not started this thread, it is about MFDBs as well, so I don't see why the OP should not have posted it here. Anyway, I don't understand your problem. I don't even look into the vast majority of the threads. I have no problem with ignoring such threads, which either I don't know enough of, or I don't find interesting or I dislike for any reason; I wonder why you and some others can't do the same.

[Jim2]

Here is a shot with in similar bright sunlight I took earlier this year in Moab (actualy this is a stretch of the Petrified Forest) with my P45+.  Again, in my humble opinion only, these clouds look a lot more like real clouds with similar exposure values for our main foreground subjects and nice, clean detailed shadows. (And the sky is the proper blue.):

Jack, what (kind) of camera did you use for the P45+?

Edit: Never mind I have finally caught up with this thread and saw your camera bag. You use the Phaseone 645 / mamiya body

[Panopeeper]

Great so is this the end then? We all finally agree that MFDB has more DR than DSLR's?

I don't know of any make or model called MFDB, nor DSLR. Do you?

[Jim2]

I downloaded the cloud image and used the threshold in Photoshop to see where the lowest cloud luminance was located. No clouds had a pixel value below 93. Clipping of these clouds would not occur even with a P&S camera! A pixel value of 93 in a gamma 2.2 space corresponds to a pixel value of 28 (in 8 bit notation) in the raw file. The corresponding value in a 14 bit raw file from the Canon would be 1781. This is very far above clipping. If you really see a difference, you should look elsewhere for the explanation.

[attachment=16663:FlesherImage.png]

bjanes, would you care to explain to me a bit more on this please?

The part that I don't understand is how does the 93 in gamma 2.2 becomes a pixel value of 28 in 8 bit?

I would assume that to convert 28 in 8 bit into the equivalent value in 14 bit would be: 28 / 2^8 * 2^14 =  1792... how did you get 1781?

I'm probably going to sound very dumb here: "very far above clipping" sounds weird to me because I thought clipping means values (i.e. luminance?) above certain points would be 'clipped' i.e. assigned the maximum value be it 255 for 8 bit, or 2^14-1 for 14 bit data. Don't you mean "very far below clipping" ?

I think I need to really sit and think about all this to understand the whole situation about clipping / raw processing etc...

blah all I want is to take beautiful photos to print and hang on the wall! My maths (discrete or calculus or whatever) have all evaporated off my brain many many years ago and I'm not even that old lol

Is there a page that explain this whole clipping thing that you demonstrated with photoshop? I'd like to read it to learn more.

[TMARK]

Because every thread you are a part of degenerates into some bullshit name calling.  Why?  Because you are obnoxious.  Get it?  Its not civil.  I look at this thread because I want to know what photographers who routinely shoot with a d3x/ds3 and/or a MFDB think about how their chosen tool and how it affects their Photography.  Instead we have acrimony and graphs.  Again, partially because you present yourself as an obnoxious, acrimonious prick.  That's all.

Tim, apart from the fact that I have not started this thread, it is about MFDBs as well, so I don't see why the OP should not have posted it here. Anyway, I don't understand your problem. I don't even look into the vast majority of the threads. I have no problem with ignoring such threads, which either I don't know enough of, or I don't find interesting or I dislike for any reason; I wonder why you and some others can't do the same.

[Luis Argerich]

If the truth is not civil then I will happyly abandon civilization...

Because every thread you are a part of degenerates into some bullshit name calling.  Why?  Because you are obnoxious.  Get it?  Its not civil.  I look at this thread because I want to know what photographers who routinely shoot with a d3x/ds3 and/or a MFDB think about how their chosen tool and how it affects their Photography.  Instead we have acrimony and graphs.  Again, partially because you present yourself as an obnoxious, acrimonious prick.  That's all.

[ErikKaffehr]

Hi Jim2,

I really appreciate your attitude. To my understanding Christopher Hauser may be kind enough to help us with some comparison shots in a few weeks. Regarding the technique proposed by Bernard it's really practical. I normally don't shoot multirow panoramas, but single row is fast and easy. I have two small write ups on panos:

The first one is here:

http://83.177.178.241/ekr/index.php/photoa...a-and-stitching

the other one is here:

http://83.177.178.241/ekr/index.php/photoa...s-quick-a-dirty

The second article is intended to demonstrate that it's possible to get decent panos "quick and dirty". It was shot with 12mm lens on full frame DSLR.  Both generated quite a few comments on this forums. Some additional hints:

1) It's my guess that it's better to use specialised tools like Autopano Pro or PtGui instead of Photoshop, but I may be wrong.
2) At lest Autopano Pro doesn't transfer color profile from originals

I would essentially try the technique and invest if it's found usable.

Best regards
Erik

Hi Bernard, what gear do you use to do 3x4 stitching? Special tripod heads, etc? How long does it take to execute it, and in the case of taking photos of quickly changing scene (sunset) will that be practical? I have always been a single frame shooter and have only tried stitching once or twice with a very basic tripod head. Perhaps with the right gear, stitching would be easy and quick to do.

The technical camera's ability to shift the back sensor to allow stitching seems appealing to me as it seems very easy and very quick to do. This can be achieved with either MFDB or 35mm DSLR.

I guess ideally for me is to just get an MFDB + tilt / shift body and try it myself to see which is more practical, but nobody rents it out here where I live. I guess I need to go to Sydney to rent the gear, and the rent is expensive last I looked the MFDB itself costs $600/day + I need to rent the body and lens.. and probably tripod head and a bigger tripod lol. Add that to the expense of flying there + car rental + hotel etc it would add up to about 2K-ish.

Furthermore it will be hard for me to do a fair comparison since I am used to using my 1ds3 but I have totally no idea how to use the MFDB + its lens + the technical camera. It would take me some time to get used to working with it and probably a lifetime 'mastering' it. Furthermore, there's that article that says that there are big variations amongst the copies of the same product, and usually the rental gear is even worse presumably due to abuse or lack of TLC?

I guess at the end of the day, ideally I'd like to know from several people who are used to using both P65+ with a technical camera and 35mm dslr (1ds3 or d3x) to come out and comment about the comparison of equal size prints (e.g. 30" x 40" or bigger) of the same scene taken at the same time - using P65+ (or other kinds of MFDB) vs 35mm (perhaps being stitched to make it 'fair' on resolution).

I know that several people have voiced their impressions although they cannot explain in technical speak (which is probably not necessary?) and I thank them for that.

So far I haven't seen any opinion to say that a 35mm dslr (stitched or otherwise) print looks better or even of the same quality / beauty (for the lack of a better term) than MFDB on large prints - in case if I missed it, please kindly correct me.

Thanks for all the posts so far. It has been very educational to me.

[Ray]

Also if they are down-sampled to 0 megapixels they will have identical dynamic range. At this reproduction size they also benefit from a absolute zero level of artifacts.

Yes. Very amusing. But what amount is zero millons? Is 999, 999 pixels zero megapixels? I presume you mean zero pixels. If that's what you mean then at zero pixels, dynamic range can be both equal and unequal, or anything you like, because it doesn't enter into the picture   . Just being pedantic.

However, I agree you have a valid point. It's a bit slack of me to presume that a D3X image will have 2/3rds of a stop DR advantage over the P65+ whatever the print size, just because at the normalised size of 8x12 it has a 2/3rds stop advantage.

A more correct deduction would be that the D3X image, whatever the print size, will have at least marginally better DR, even if it's only 0.25EV better. It will at least be no worse, according to the test data presented by DXOMark.

For example, the P65+ sensor has 2.5x more pixels than the D3X sensor. The D3X pixel is the same size as the P65+ pixel and has 1.67EV greater DR, according to DXO. 1.67EV greater DR translates to over 3x the amount of light, all else being equal. The P65+ gathers only 2.5x the amount of light for any scene of equal FOV at equal exposure.

Is there a flaw in my reasoning. There could be. I'm not as technical as some of the posters on this site.

[Jim2]

Yes. Very amusing. But what amount is zero millons? Is 999, 999 pixels zero megapixels? I presume you mean zero pixels. If that's what you mean then at zero pixels, dynamic range can be both equal and unequal, or anything you like, because it doesn't enter into the picture   . Just being pedantic.

However, I agree you have a valid point. It's a bit slack of me to presume that a D3X image will have 2/3rds of a stop DR advantage over the P65+ whatever the print size, just because at the normalised size of 8x12 it has a 2/3rds stop advantage.

A more correct deduction would be that the D3X image, whatever the print size, will have at least marginally better DR, even if it's only 0.25EV better. It will at least be no worse, according to the test data presented by DXOMark.

For example, the P65+ sensor has 2.5x more pixels than the D3X sensor. The D3X pixel is the same size as the P65+ pixel and has 1.67EV greater DR, according to DXO. 1.67EV greater DR translates to over 3x the amount of light, all else being equal. The P65+ gathers only 2.5x the amount of light for any scene of equal FOV at equal exposure.

Is there a flaw in my reasoning. There could be. I'm not as technical as some of the posters on this site.

Interesting. If D3X is better than P65+ in DR, in your opinion is there any reason that you can think of, to use P65+ at all?