Everything Matters. It's All About The "Small Details" by Mark Dubovoy Jan 2012

[photodan]

This is in support of Mark Dubovoy’s article:  I am restricting my comments to those technical aspects of his article regarding the audiophile analogy and the photo comparison.  When I read Mark’s articles I keep in mind that he is, in my opinion, a perfectionist with a technical bent. Not everyone’s cup of tea orientation of course, but I can relate to it so therefore I like to read his articles.

Although I found the article interesting and informative, others obviously have not, and some even have called for future articles to be heavily edited or censored altogether. This is a serious over-reaction in my opinion. Surely there is room for a variety of world-views, so to speak, in the articles on this site. If  you don’t like an article, or think it’s full of hogwash- fine, just ignore the articles in the future and/or criticize them at that time as is being done now. But elimination of them – way over the top!

One of his major points is his opinion to not rely on engineering specs (DxO specs for example, as related to photography). Test a product and look at the final results. Regardless of specs and test lab results in restricted circumstances, the results in real-world conditions (your shooting conditions) may yield different conclusions as to how ‘good’ a lens is (at least as it relates to each photographer).  I heartily agree. Does anyone really disagree with this? 

Test results can be a good guide as to things to look for, but I have encountered many instances where I prefer a lens that didn’t test out in published lab test as well as another one – it could relate to the bokeh, or micro-contrast or whatever, for which there may not be adequate tests (not to mention just pure personal preference as to various tradeoffs in lens qualities).

As to his photo comparison I do think it would have been better to use a small sensor camera that shoots in raw, something that has similar DxO specs to the medium format back used in the test (dynamic range, color sensitivity etc.). Then, try to get the ‘best’ out of each, using a raw processor such as Adobe Camera Raw, or LR. Then present the results. Not layers and such - just simple raw processor adjustments such as black/white points, highlights recovery, etc.  Not an objective test, but I believe it would be a more practical and valid comparison. It might show that DxO is right on the money, or it might show that the specs don’t tell enough of the story (as Mark asserts).

Mark used audio as an analogy for contrasting actual experience vs. specs. I think it’s a useful analogy on an absolute basis. However on a relative basis- meaning percent improvement vs. cost, perhaps not that universally applicable (and so far most of the comments on this forum indicate that is an understatement).  While his power cable portion of the analogy may be a bit much for me, I  support much of what Mark said, via my own personal experience.

I have a fair amount of lay-person experience in music – I play several instruments, my mother was a professional singer, and my late father was a mechanical engineer an amateur recording engineering using old-time pro equipment – 15ips tape, tube equipment, high end (for their time) speakers etc.   I have attended many live music performances of small and large groups; gone to many audiophile shows, performed extensive listening sessions in various showrooms, and have tried out a large amount of mid to high end equipment in my home (and some years ago purchased some low-end audiophile equipment that has given me much enjoyment to this day).

 I have heard live music vs. almost immediate recorded playback, analog via tape vs. live, vs. CD, DAT, LP, direct-to-disc, etc.  As the result of my experience I do believe that the best analog is superior, in some way or ways, to the digital I’ve heard. And very generally speaking, expensive equipment often sounds a lot better than inexpensive equipment.  I don’t feel that expensive cables and such are worth much to me, and the differences in sound can perhaps be replicated in other ways. But that doesn’t negate the other aspects that relate to tubes vs. transistors, and distortions in the digital recording and playback chain (that have more deleterious effects in some ways than the analog distortions).
 
The 1st generation of digital, marketed as perfect sound, sounded dead and unmusical to me. Succeeding generations of digital is trending much better. Yes digital has perhaps more headroom/dynamic range, less noise, less harmonic and non-harmonic distortion, and better frequency response in some respects than analog. But there is something about analog that is more real and more natural sounding to me. I have yet to see specs that relate to this. 

Some will say that I’m deluded, or that I prefer the large euphonic harmonic distortion of analog (tube pre-amps / amps for example). Well, whatever – fine, I’ll take what sounds more realistic and truly musical to me.   I know how music sounds, and I know that digital sounds inferior to me in some regards.

OTOH, as a practical matter, most music is released on CD, not on LPs, and I don’t want to deal with having to make sure my stylus is aligned perfectly and that I keep the records totally clean etc -Just like I gave up shooting large format film with all its maintenance requirements, not to mention the processing. Plus, digital audio is getting better, with improvements in jitter reduction, digital to analog conversion etc.  Now if only most audio would be released in a widely available format beyond the regular CD :-)

Most of the population seems to be involved in the convenience of downloading compressed music and video, and playing with the so-called art filters on photos from their smart phones, rather than obtaining better sounding fidelity such as from CD and better video such as from Blu-ray. 

And, yes - art matters more than technical quality to most people, but what is wrong in pursuing both,  or at least discussing both?

The topics discussed on LL don’t concern most of the population, as they are too esoteric and quality driven.  Mark’s technical quality objectives seem to be another cut beyond most photographers objectives.  I’m not saying that Mark's objectives are important or should be to others,  or should not be to others. I’m saying that they are worthy in their own right, and since they are applicable to at least  some people beyond just Mark, that they deserve a hearing.
 
I look forward to Mark's next installment.

[PierreVandevenne]

There has to be a reality check at some point.

Michael does a nice, well documented, field test comparing a Sony and a Leica (at least I found it a nice test). Since those things are always a bit subective, he gets some (undeserved imho) flak and gracefully returns to the topic. Now, that is what makes this site a great source of info.

Can we agree on that?

Mark, for the fourth time or so, tackles is pet topic, which is that MFDBs are better than absolutely everything else and that only poor ignoramuses would fail to understand that. Assuming he knows how to write, he purposefully adopts a condescending tone: he is on a mission to educate the unwashed masses. We get the usual round of flawed analogies, mind blowing claims, culminating this time with a magical power cord (but that is somehow not magical enough that it works at once, it needs to be broken in). The whole things sounds a bit childish, as if the main purpose of the article was to remind everyone that he has access to all the toys he wants.  All of this in an article whose main point is that "details do matter", which in itself, is generally consensual and hardly worth an "essay".

Now, assume we get that point, don't "details" matter in the way an article is written?

If the lesson is that we should worry about picoseconds, last yard power transmission, Stradivarius and so on, shouldn't we expect better than a branch, some fog and an amazingly crappy flower bouquet as illustrations?

And to make matter worse, if one of the file actually came from an iPhone, this sentence is borderline misleading.

The images received no processing adjustments.  I exposed them as carefully as possible, did a direct conversion from RAW to TIFF with no adjustments and since both images had a lot more pixels than 600x800, I reduced the image size in Photoshop using the standard "bi-cubic sharper" algorithm.

Yes, a lawyer would argue that Mark did not say "two direct conversions..." But still, it is obviously confusing...

In short: condescending and patronizing overtones don't mix well with imprecise (yes, I am charitable) sentences and outlandish unrelated claims.

I imagine the flak would even be worse if people were unaware of Mark's position in Lula...

Anyway, ultimately, we are given a rationale: "MFDBs have better pixels". Well, the hard truth here is that they do not: they are different, in the sense that CCD is different from CMOS, but being of an older generation, in a smaller market, they unfortunately aren't "better" in the absolute sense of the term. No magic cable will ever coerce them into producing a better dynamic range than their engineering specs suggest. FWC and read noise are what they are. Modern CMOS sensors have lower read noise and better dynamic range.

I am sure that if Mark said

"I love MFDBs and I will try to show you why in this wonderfully illustrated article" instead of "some barbarians are still unaware that there is one camera to rule them all..."

he would be extremely well received.

PS: I do prefer CCD pixels to CMOS pixels myself, but wouldn't claim they are "better".

[PierreVandevenne]

One of his major points is his opinion to not rely on engineering specs (DxO specs for example, as related to photography). Test a product and look at the final results. Regardless of specs and test lab results in restricted circumstances, the results in real-world conditions (your shooting conditions) may yield different conclusions as to how ‘good’ a lens is (at least as it relates to each photographer).  I heartily agree. Does anyone really disagree with this? 

I certainly do agree. This being said, no system (especially in the field of image sensors) can exceed its engineering specs. Specs are not the last word, and obviously you can design a bad system around a good sensor, and a very good system around a so-so sensor.

But one doesn't need to make outlandish points (such as 6dB of additional dynamic range) to make that point.

As to his photo comparison I do think it would have been better to use a small sensor camera that shoots in raw, something that has similar DxO specs to the medium format back used in the test (dynamic range, color sensitivity etc.). Then, try to get the ‘best’ out of each, using a raw processor such as Adobe Camera Raw, or LR. Then present the results. Not layers and such - just simple raw processor adjustments such as black/white points, highlights recovery, etc.  Not an objective test, but I believe it would be a more practical and valid comparison. It might show that DxO is right on the money, or it might show that the specs don’t tell enough of the story (as Mark asserts).

Yes. That would also dismiss the impression that the "test" was rigged and ambiguously described.

Mark used audio as an analogy for contrasting actual experience vs. specs.

Audition in general has been proven to be highly subjective. How we hear a Hi-Fi system is a valid analogy for how we look at and appreciate a picture.

But don't drag sensors and pixels there: unlike audition or or personal vision, they are the epitome of objectivity. They are an essential, reliable and objective tool in most of the hard science and hard tech done since 1980. From Hubble do the LHC, from medical imaging to DNA sequencing, they are everywhere and don't display any kind of magical properties.

Some will say that I’m deluded, or that I prefer the large euphonic harmonic distortion of analog (tube pre-amps / amps for example).

Essentially, that's a subjective preference (which I share to some extent) I own tube amps and old classic speakers.

The ultimate demonstration of that audio subjectivity must be this

http://www.npr.org/blogs/deceptivecadence/2012/01/02/144482863/double-blind-violin-test-can-you-pick-the-strad

We all would respect Mark's choice if he preferred MFDB. The problem is the justification attempts on bogus claims. He doesn't trust specs, yet invents his own specs to support is preference. That's why he is poorly received imho.

I look forward to Mark's next installment.

so do I

[Schewe]

Please explain to me why *any* of that should be mentioned in an article on photography?

Ever heard of an analogy? He was trying to get a point across (apparently lost on you).

[kwalsh]

(I kinda have some experience in this phenomena).

With all due respect Jeff, you've never written anything half as ill reasoned or poorly articulated as what Mark is writing here

Oh, wait...you didn't PAY any money to read the article did ya?

Sorry, give it a rest on this one Jeff.  You know better than that. I know you do because I've read most all of your work - including your "free" work - and its quality demonstrates you don't subscribe to this kind of thought in your work.  This is the standard excuse for crap writing, and that line of reasoning leads to more crap writing.  You have an audience.  It isn't your audience's money that matters - it is the time and attention they lend you that matters.  Write with that in mind, that someone is taking time from their life to evaluate your writing and you'll continue to have an audience.  Write with the idea that what you spew is free and a gift to all who read it and you end up with what Mark wrote.  I've never seen you write in such a way, which is why I value what you write.  Do Mark and the rest of the "free" writers on the internet a favor by holding them to a standard comparable to your own work and don't coddle their egos with an off-hand "its free" remark.  I don't think it becomes you, and really it doesn't serve them either.  Even if they are a friend or colleague.

Similarly, there is a lot of garbage on the net and sites like LL retain their readership by applying editorial controls to what goes on them.  I think it is important for Michael to have feedback when his readership finds what he allows on the site to be a waste of their time.  That's why I'm bothering to spend more time writing about this awful article.  So that the value of LL's editorial content is retained in the future.  If it just becomes a sounding board for under-baked concepts and poor writing I won't bother visiting anymore, and from the sounds of this thread there are a lot of people coming out of the wood work to express their opinion on that - pretty consistently.

Ken

[DaFu]

Goodness gracious! What a lot of vituperation and commotion!

I rather liked Mark's article and was about to post some thoughts on it but I found that photodan's comment said just about all I wanted to say.

Dave

[Schewe]

With all due respect Jeff, you've never written anything half as ill reasoned or poorly articulated as what Mark is writing here

Thanks...I think :~)

In point of fact, while I read Mark's article it didn't really present anything I didn't already know. And knowing Mark, I think he sometimes struggles to get his passion and love of the craft across.  On the other hand, I've learned not to bother watching Fox News...the tone and tenor just purely pisses me off. So, I've deleted that channel from my TiVo lineup so I never have to get pissed off again.

So, are you gonna read part two?

:~)

[bjanes]

Thanks...I think :~)

In point of fact, while I read Mark's article it didn't really present anything I didn't already know. And knowing Mark, I think he sometimes struggles to get his passion and love of the craft across.

Mark discounts DXO, but yet DXO refers to the IQ180 as the king of the sensors. Does anyone see a disconnect here? I haven't had the opportunity to work with the IQ180, but it is undoubtedly a very fine instrument and I don't see why he needs to make outrageous claims about 6 additional stops of DR or its marked per pixel superiority to a dSLR.

Getting his passion and love of the craft across is a nice thing to do, but there might be a slight conflict of interest as suggested by a Google search for "podas workshops dubovoy". Those all expense paid trips to some of the most photogenic sites in the world and possible honoraria must also be nice.



Regards,

Bill

[Schewe]

Getting his passion and love of the craft across is a nice thing to do, but there might be a slight conflict of interest as suggested by a Google search for "podas workshops dubovoy". Those all expense paid trips to some of the most photogenic sites in the world and possible honoraria must also be nice.

You painting me with that same brush bud?

Do you really presume Mark's opinions are available for sale?

Are yours...or maybe it pisses you off that nobody places any particular value on your knowledge and ability to teach...really dooode, that's way off base and below your usual low standards.

I suggest you offer an apology to Mark for casting aspersions on Mark's reputation and go stand in the friggin' corner for a while (I also expect this particular thread to not last too long with that kind crap being slung about).

If you knew Mark you wouldn't have said that...and the odds you'll ever get to meet him just got a lot lower.

(BTW, if you don't know, Mark has a PHD in nuclear physics and has made enough money in VC so as to not have a particular care about financing–hence the tendency to being an audiophile and a lover of great wine. So to postulate that his opinion could be bought for material good is really an extreme sign of ignorance).

Criticize the concepts or writing but leave the ad hominem attacks at the door (assuming you want to continue having a LuLa subscription).

[ErikKaffehr]

Hi,

I'd suggest that Mark's passion for perfection is absolutely real. So he uses medium format with the best stuff he can get.

In my view that all adds up.

- Careful work
- Exact focusing
- Camera shimmed to perfection
- The best lenses available
- The best MFDB available

So I'd suggest he gets better performance out of that platform than most shooters. I'm glad that he tells about his experience.

On the other hand, in the article there are a lot of esoteric references. And he makes the statement that the advantage of the superior equipment will be visible in small prints. Well, fact is that photography is a processing pipeline with printing as the final step. When we print small we discard most of the information at hand. Resolution is limited by the printers native resolution and dithering algorithm. Most of the color space is also lost.

My experience is that doing correct tests is very problematic, partly because the processing pipeline does include steps where we do signal reconstruction. At some stage of the process we apply sharpening which is highly subjective, than we apply color matrix and curves and do individual tuning for colors. With all variables we never can say if the superior results are a consequence of the processing or the underlying signals.

I would say that the great 2006 MFDB shootout here on LuLa went to lengths to make a correct comparison of sharpness. At that time the article got a lot of flak that 4x5" inch Velvia was scanned at to low resolution. I guess that the reason for that was that it was exactly the resolution the authors used in their normal work.

The 2006 test is here: http://www.luminous-landscape.com/essays/back-testing.shtml

A very good test was published recently, this time comparing large format film with digital: http://www.onlandscape.co.uk/2011/12/big-camera-comparison/

The above tests were 'properly made'. Incidentally, I also made some film vs. digital testing recently and I come to a different conclusion, partly because I look at somewhat different things and partly because I use lesser equipment.

I can also mention that Mark seems to have a lot of respect for DxO, even if it is not obvious from his writing.

Regarding the 6 stop advantage in DR he sees I have some suggestions:

1) He is talking about DR involving texture, so it is a combination of sensor DR and lens MTF
2) Flare will normally also affect DR
3) Mark is comparing MF at low ISO with his Canon EOS 1Ds (mark something) at base ISO where, comparing with a state of the art Nikon (D3X) or Pentax (KT) he would find less difference.

My article http://echophoto.dnsalias.net/ekr/index.php/photoarticles/38-observations-on-leica-s2-raw-images compares Leica S2 with Nikon D3X based on raw images kindly made available by Lloyd Chambers. My focus at the time was moiré but the article offers some other insight in differences between the two systems and I don't think it is much biased.

Best regards
Erik

Ps. I'd suggest that Mark works with PODAS because he loves MF digital, not the other way around! I also had some direct communication with Mark about specifics of his articles and I got very well written responses promptly.

Mark discounts DXO, but yet DXO refers to the IQ180 as the king of the sensors. Does anyone see a disconnect here? I haven't had the opportunity to work with the IQ180, but it is undoubtedly a very fine instrument and I don't see why he needs to make outrageous claims about 6 additional stops of DR or its marked per pixel superiority to a dSLR.

Getting his passion and love of the craft across is a nice thing to do, but there might be a slight conflict of interest as suggested by a Google search for "podas workshops dubovoy". Those all expense paid trips to some of the most photogenic sites in the world and possible honoraria must also be nice.



Regards,

Bill

[degrub]

When we get smart enough as engineers to create a functional human brain we will have figured out how to understand human perception. What Mark's writing points out to me is that we have a long way to go and there is a gap between them. We don't understand perception well enough to explain it and what we perceive doesn't always match up with the science that we do know and can build systems from.

Frank

[bjanes]

You painting me with that same brush bud?

No, you have not made any assertions that violate the basic laws of physics. You write with clarity and style and I have enjoyed and learned much from several of your books that I have purchased.

Do you really presume Mark's opinions are available for sale?

If you knew Mark you wouldn't have said that...and the odds you'll ever get to meet him just got a lot lower.

(BTW, if you don't know, Mark has a PHD in nuclear physics and has made enough money in VC so as to not have a particular care about financing–hence the tendency to being an audiophile and a lover of great wine. So to postulate that his opinion could be bought for material good is really an extreme sign of ignorance).

I don' really understand Mark's departure from reality when he is discussing MFDB and making statements such as 6 additional stops of DR. That is not the type of statement that I would expect a PhD in physical science to make. He should really state possible conflicts of interest and his relationship to Phase One, much as Micheal has done. Maybe he is just carried away with the quality of his setup, but why the need for such hyperbole? It only detracts from his post which has not been well received by the majority of those who have commented.

As for blowing my chances of becoming a fan boy of Mark, I can live with the consequences. At least, I know that my knowledge of the science of photography is sufficient to see through his preposterous assertions that are made in the absence of any data or scientific rationale.

Regards,

Bill

[Schewe]

He should really state possible conflicts of interest and his relationship to Phase One, much as Micheal has done.

He gets paid to teach...much like Mike and myself as well as Art Wolfe, Andy Biggs, Bill Atkinson, James Martin, & Peter Eastway among others. The fact that a Google search turns up such urls is of little or no consequences (and to be expected)...to try to make something out of it is petty and demeaning...it really tells far more about yourself than it does about Mark...

[bartfrassee]

Actually it's not...I wrote a comparison between an iPhone 3, Rebel, 1DsMIII and a P65+ in my book Real World Image Sharpening...

I still believe it is (at least the way Mark Dubovoy did it): Mobile phone camera JPEG images tend to mimic post card pictures, i.e. seem to be processed in camera for high contrast and high saturation, so naturally they tend to blow out channels. To really compare sensors, the images should be carefully processed from raw files.

Jeff, I actually own your (excellent) book (see, I pay for well researched information) and remember that (well done) comparison very well. That's why I want Mark to redo his comparison in a serious and sound way (and perhaps use a Nikon V1 ore some m4/3 camera instead of an iPhone, because he should use raw data). I think he will have a hard time to really prove his claim with such a small output image size in a fair comparison.

If you don't like what Mark wrote, unread it. Either that or demand your money back. Oh, wait...you didn't PAY any money to read the article did ya?

Well, time is money 

[BlackSmith]

Schewe,
In one sense it is noble to come to the defense of a friend, but are you underestimating him? I think he is a big boy. Any good scientist can admit they are wrong when logical arguments are presented to them.
bjanes,
Its clear to most everyone that you are correct in most of what you are saying. But dude - why poke the bear?

My take home message from this article is a reaffirmation that the LuLa forums consist of quite remarkable readers and contributors.

And here's hoping the follow up article is a surprise gocha!
(because... well come on!
Even though us PhD Chemical Engineers look down on PhD physicists, I wouldn't expect this from a BA. I'm kidding. The point is that it's silly to bring sheep skins into the argument. )

[jeremyrh]

My kudos to Mark for this article!  Why so many are attacking his presentation with such vengeance is beyond me.

There may be some hyperbole involved -- he may have a tendency to present personal opinions as gospel -- but I see this merely as a matter of style and an aid to get his point across effectively and with clarity.  Works well for me.

Overall, I found the essay to be wonderful food for thought.

Not so much hyperbole as deception, IMO - passing off an iPhone shot as something else, even by omission, is intellectually dishonest in my book.

Oh - and since little things matter, kudos means "glory", so you don't have any you can give to Mark, it's not transferable :-)

[Rob C]

Interesting, no such thing as empirical observation in your world?  Everything is just words, your words or the words of someone else?  No facts, no observables, just varieties of "opinion" and "wisdom".  Very telling as to why you are having trouble communicating with some in this thread Ken

That, Ken, is the nature of life; I am no politician and have no need to try for mass appeal. In fact, I have no need to try for any sort of appeal. I am what and who I am, and I'm perfectly happy that some will accept that and others not. I tell it how I see it and am perfectly aware that other's see things (or not) in their own manner, which as long as it does not harm me or mine, seems perectly reasonable to me.

Rob C

[dchew]

Article quote in the Summary:
- Everything matters.  It is all about the small details.  Celebrate these details. They are what makes photography so exciting.
- Do not ignore details. Even the smallest ones can be crucial. The weakest link in a chain will always determine the ultimate quality of the entire chain.
- Do not rely on specifications, measurements or marketing claims.  Trust only your eye/brain system.
- Search for the unseen. The first step in creating a great image is to show something heretofore unseen.
- It is not only the subject matter that contains the unseen.  It can be a special angle, a special view, unusual lighting, a distinct vantage point, Hyper-Reality or something else.
- Prints and screen images from larger format captures always look better, regardless of the size of the print or the screen image.
- Avoid myths. This is why I busted the first one and will bust a few more in the second part of this essay.
-end quote

So let's see... The devil is in the details, test equipment yourself, search for the unseen, prints from bigger formats look better and avoid myths.

If you strip away the emotions flowing from this thread, I think most of us would agree with the list in that sentence.  People might take issue with Mark's style, but personally I like it; I think it is kinda fun.

Ciao,
Dave

[bartfrassee]

So let's see... The devil is in the details, test equipment yourself, search for the unseen, prints from bigger formats look better and avoid myths.

If you strip away the emotions flowing from this thread, I think most of us would agree with the list in that sentence.

Sure, but like in school exams, if you've got the right answer but your proof, explanation or derivation is (partially) wrong, you won't get full marks.

[bjanes]

Are yours...or maybe it pisses you off that nobody places any particular value on your knowledge and ability to teach...really dooode, that's way off base and below your usual low standards.

Criticize the concepts or writing but leave the ad hominem attacks at the door (assuming you want to continue having a LuLa subscription).

Jeff,

Why don't you follow your own advice?

Regards,

Bill