Wednesday, October 22, 2008

Gone to Wordpress

I just got the memo : Wordpress obsoletes Blogger.

Most of the blogs I read regularly moved to Wordpress long ago. For some reason I stuck with Blogger for 114 of my wordy, graphic, painful posts (to write). Blogger’s editor is so rife with bugs that I switched to writing directly in HTML earlier this year. At least post previews sort-of worked.

To make the change all the more dramatic, Wordpress imports your entire Blogger blog automagically – text, formatting, images, categories, comments and all. The conversion was essentially perfect, I had to retouch a handful of posts only because I had abused blockquotes. If Blogger had an import feature, it would be a text box where you can re-type them.

Welcome to Wordpress:

http://ogiroux.wordpress.com/

Idle Daydream


Tuesday, October 21, 2008

Photograph like an engineer.

Analytical Solutions

I first studied mechanical engineering at a school that prides itself in being the deepest into theory out of four major universities that serve the Montreal area. Most of the time I spent on the school bench was devoted to the quest for the holy grail of engineering that is the analytical solution. As I look back, I think engineering scholars suffer from a “physicist complex”, in that deep down they wished they really were physicists and not engineers. Professional engineers get little kick out of playing pretend-physicist.

Numerical solutions are easier to get, sufficient for most of all purposes, and you could teach juniors how to set them up for a vast space of problems. Comparatively few people can set up theoretical arguments with sufficient modeling to come out with more useful precision in the end. Yet the analytical solution remains a source of pride for its author, even when he/she is aware of this.

Why? First, the analytical solution is an achievement that lifts the author above the ordinary. It’s a statement about his or her discipline and sustained concentration. Second, the analytical solution presents itself untainted by crude errors borne in the physical world. The solution’s limits were agreed upon from the outset by the stated simplifying assumptions and thereafter the solution is presumed to be free from error.

Thus the analytical solution is perceived to be more worthwhile, even when it is more painful and/or less useful.

Analytical Photography

Ack! What a blood-curling term we have here. Photography is art!

I don’t know what percentage of the population is made up of art historians, but I don’t think my family is a good barometer for popular interest. Within one degree of separation I count two art historians, yet within even two degrees I am the only engineer that I’m aware of. I mention this cultural bias to prepare you for what I’m going to say next.

Analytical art is not an oxymoron. There were periods in history, particularly during the renaissance, where artists were obsessed with technical minutia, for example harmonics, perspective and human proportions. They have argued about color for thousands of years. Recently (in terms of history) so-called “method” actors formalized the simulation of emotion even, to the delight of millions. Basically, analysis comes into play when art is concerned with depicting something real.

Most art today is not intended to promote analysis of its expression, however, and here is the defense of contemporary art. Displays of art now invite analysis of a hidden message instead. Nevertheless, many aspects of contemporary art involve advanced techniques and technology which must be mastered to make the art appealing and draw the audience to the message. This definitely applies to photography.

Analytical photography, I define now, is meticulous technical photography that would not be considered art today. In the current context, that means photography that has no message or doesn’t fit in a coherent body of work with an overarching message.

Engineer Photographers

There is a deepening interest in analytical photography today. Dozens of discussion forums filled with analytical content, huge vaults of perfect generic images, and a vocal community hungry for technical prowess are evidence of this in my opinion. This community includes professionals and amateurs alike, and for all the effort they pour into photographic analyses there are very few in it that produce artistically meaningful work.

I am part of this community. I do not produce artistically meaningful work.

At this stage I see myself as an “engineer photographer” and I mean it derogatively: a photographer in search of the analytical solution. This is a sort of manifesto for the moment.

My wish is to grow past this stage.

Saturday, October 18, 2008

Moore's Law in Action

The guys at "ifixit" have taken apart one of the MacBooks...


It's really astonishing how small the computer inside the computer is.

Tuesday, October 14, 2008

New MacBook shocker

I don't post about non-photo technology very often, but this is really special. The new MacBooks are as plainly awesome as computers get. I didn't expect this, personally I always liked the "Pro" so much better.


Apple has got to be afraid that MacBooks will steal sales from MacBook Pros now that they're so close. I guess the gamble is that MacBooks will steal more sales from the crappy plastic Uninspirons over yonder. I sure would dump mine in a heartbeat if I could.

Icing on the cake : it's an all-Nvidia line up in this family of notebooks that makes up ~30% of the volume in the USA. Not only that, but graphics is also the main differentiator between the basic model and the "Pro" model. It's a good day for GPUs today.

Discovering Fun

This weekend was Canadian Thanksgiving. Here’s a short visual diary, free from photo-babble.


We packed up the family+dog and drove up there, made good time too with just 5.5 hours.


Vermont was pretty much like this the whole way.


We went apple-picking and the MacIntoshes were phenomenal.


Mommy building up enthusiasm.

In the afternoon we took my daughter to slide for the first time – she’s 9 months now.


She didn't get it the first time.


But after a couple of times... that's pretty cool!


Ok. I’m ready, let’s do it!


That’s my toy.


It's better alone with grandma.

Cheers,

Thursday, October 9, 2008

Op-ed: The Leica M8 would make more sense as a B&W camera

Hello,

My name is Olivier and I am addicted to the idea of owning a Leica M camera.

My problem (trying to service my addiction ;^) is that I can’t justify it. I have no interest in debating the issue of price, except that the non-trivial cost warrants a thoughtful and level-headed justification. At this time in October of 2008, the M8.2 fixed just about every known issue with the camera except the most important one: the less-than-stellar digital imaging system.

There are four basic things that are wrong with this system…
1) IR sensitivity pollutes images with either a magenta or a cyan cast. This distortion of the source data is really permanent damage to the image – correcting the cyan cast inevitably amplifies noise in the red channel.

2) Risk of developing color moiré pattern from high-frequency detail. Again the source data has to be improved for outputs to be free of moiré.

3) Poor performance at sensitivities above ISO-800 compared to the competition today. The benchmark to compare against is not color negative film, that’s a don’t-care, it’s modern professional DSLRs in the same market category (D700/5D-II)(ignoring that they cost ½ as much). The M8 needs to see a 1.5 - 2 stop improvement to be competitive in this environment.

4) While good, resolution is not class-leading. Leica lenses deserve better.
One thing that I think is not wrong with the M8 is the 1.33x crop factor. It offers an excellent balance of quality overall, without skewing the system towards telephotos too much. I imagine the cost would rise with a 1.0x sensor, which would only serve to make justification more difficult. The Leica lens stable is full of good choices for the 1.33x factor too, for example the 28/75 pairing for which many interesting permutations are available.

Entertain this thought experiment if you will… what if the M8 lost its Bayer color filter array? Change nothing else, just strip away the CFA and turn the M8 into a black & white camera.

Let’s see how this would affect the four issues I listed above:
1) IR sensitivity is much less of an issue. Magenta casts appear as luminance variations that are not as distracting, cyan cast appears as light fall-off. Full-resolution IR photography becomes an option using a filter on the lens to block visible light.

2) Color moiré is no longer possible. High-frequency detail is recorded simply as fine detail, in all its glory.

3) Sensitivity improves by 1.5 - 2 stops as the color filtration no longer consumes light; all wavelengths contribute to luminance.

4) Real world resolution improves by 1.5x - 4x depending on situations because no interpolation is necessary to produce image pixels. 10MP is a terrific resolution for a monochrome sensor.
The result of this single modification, a simple ECO for Kodak, would be a stunning black-and-white M8 camera. One that can stand its own in the modern competitive environment, and would now be priced appropriately taking its tradition into account (assuming the price is the same).

Sunday, October 5, 2008

Neat Photokina Blog

Some of the best chronicles of this Photokina 2008 are in David Farkas' blog, over at http://dfarkas.blogspot.com/

If you only have time to read only one of his posts, I would read this one where he recounts a very intersting discussion with Peter Karbe (the Canadian in charge of Leica's optical program for the past few decades). He reveals tasty morcels left and right, but my favorite is the one where he let slip that the 50mm Summilux-M ASPH is an APO design, had been from the start. They never advertised it publicly because they felt it would sound weird for a 50mm to be labeled APO, so they only advertised it on its sister lens the 75mm Summicron-M ASPH.


Notice the different shade of glass elements, indicating different glass chemistry. Click to enlarge. Linked from here.

In another article I found a really neat discovery. Not only did Zeiss display a cut-out 2/100 Makro-Planar T* lens, that's super cool, but that lens was a relative of mine too. The very same person verified its assembly and test.

This isn't all there is on David's blog, so go check it out.

[Apparently David got his trip paid for by his employer store. I'm not saying I endorse the store, but it looks like a good place to get high-end boutique stuff.]

Friday, October 3, 2008

Fun with Canon's interview

DpReview just posted an interview with Canon Camera’s head honcho. It was… unenlightening really, just spin. So I’m going to spice it up by replacing Mr Maeda’s answers with my own decryption of what he said each time.

Disclaimer: what follows is sarcastic humor, spiked with a tinge of truth. The questions are all real, but the answers make up my satire.


It's me, FakeMasaya.
I'm older and Japanified.

DPR: Let's start with DSC and the announcement of the G10. We're seeing a lot of pent-up demand for a better quality compact, perhaps with lower megapixels / bigger pixels or bigger sensors. What our readers want to know - and what we want to know - is this; is there any chance of Canon making high-end compacts that don't have more pixels, but have better pixels, to fill the big gap between DSC and DSLR?

FakeMasaya: Phil – can I call you Phil? – it’s simple, no. We looked at our brand power and we determined we could continue to grind Olympus into dust without listening to customers’ demands. That’s called branding, and our plan is to stick to it.

DPR: Which leads on to our next question. What benefits can we expect to see from the use of CMOS sensors in compact cameras?

FakeMasaya: Hmmm, smell that branding. So good.

DPR: Is this (CMOS sensor compacts) something we're going to see more of in future models?

FakeMasaya: For sure, Phil. We’ll use these “CMOS inside” stickers, here, to indicate to the customer which products have the best margins for us. It’s our way to provide feedback to the customers.

DPR: So at the moment is there still an image quality disadvantage to CMOS in a small sensor?

FakeMasaya: Look – I’ve got this 1Ds Mk-III on the left, and a small-sensor compact on the right. By putting my “CMOS inside” sticker on the compact, here, I’m saying it’s on the level of the 1Ds Mk-III. If it makes no sense to you, it’s because you’re not in the target market for this, it makes perfect sense to them.

DPR: We ask because the standard set by EOS cameras using CMOS - in noise terms - is very high, and there's an expectation - realistic or not - that this will be reflected in compact cameras using CMOS sensors.

FakeMasaya: Exactly! You can expect us to build our marketing campaign around this expectation. We’re calling it “Canon MOS”, that’s what CMOS stands for really, we’re getting the trademark as we speak so nobody else can make this claim.

DPR: Do the compact camera systems use the same proprietary on-chip noise reduction systems as the EOS sensors?

FakeMasaya: That’s none of your business Phil.

DPR: Now we have tiny compact camera sensors with over 14 million pixels are we getting to the point where resolution is being limited by the lens?

FakeMasaya: Not at all Phil, not at all. Our JPEGs always contain 14 million pixels in them, irrespective of the lens quality, so we’ve got quite a bit of headroom still in the G-series lens.

DPR: So our point is, why keep going? If you're already at the point where adding megapixels brings no benefits why do it? As a market leader could Canon not take a stand on this issue…

FakeMasaya: See I’m standing right here, and the issue is over there. That’s how I stand on the issue, well I don’t actually stand on the issue as you can see. Does that answer your question?

DPR: Moving on to SLRs, cameras such as the EOS 50D and 5D Mark II are stretching the capabilities of lenses harder than ever. What are your priorities for lens development?

FakeMasaya: Excellent question Phil, but I find your lack of faith in the “L”s is disturbing. I can reveal that we are working very hard on a new firmware that will deliver more resolution on all our shipping cameras, by increasing the default sharpening from “high” to “crispy crunch”. We’ve been looking at the 5D and D700 reviews out there and we’ve determined that default sharpening is really the biggest advantage of using Canon L lenses instead of Nikkors. Upcoming “L” lenses will communicate to the body to make sure sharpening is applied even to RAW files, and can’t be disabled. That’s a genuine innovation from Canon you won’t see anywhere else.

DPR: Are we likely to see EF-S - and APS-C cameras in general - moving to the entry level, with full frame moving towards the mid-range / EOS 40D/50D sector of the market? Is there any danger that EF-S will be pushed out of the market altogether long term?

FakeMasaya: Phil, as you know, there will always be women and children to which we can market APS-C. (Editorial: I kid you not, he said this) It’s true that Real Men™ as a market resonnate with the larger, bulkier cameras that offer more hubris and higher margins. Our plan is to not disturb this pricing structure too much for as long as possible, basically we’re going to wait for competitors to stick their necks out first.

DPR: Some professional users have expressed concern about the fact that the 5D Mark II offers the same resolution and a more modern processor than the EOS-1Ds Mark III, making the 1Ds Mark III seem a little dated. What are your thoughts on the relative positioning of the two products?

FakeMasaya: Whiners.

DPR: One of the most common complaints we've seen about the 5D Mark II is that it still has the same AF system as the original 5D. Why is this?

FakeMasaya: It’s simple really, I fired the entire auto-focus team last year. Not only did this lower R&D costs overall, but the old technology is cheap to make too. It’s part of my plan to turn the 5D Mk-II into a pricing weapon, a zweihander made of cash, should any competitors stick their necks out like I said earlier.

DPR: Would you ever consider removing the anti alias (low pass) filter - or using a lighter one - on high end, high resolution models such as the EOS 1Ds Mark III, to improve pixel level sharpness, removing any moiré in software (like medium format cameras)?

FakeMasaya: That’s just stupid. No.

DPR: Currently the contrast-detect AF on Canon SLRs is very slow; given that video and live view are now part of the DSLR landscape are you planning to do anything to improve contrast detect AF?

FakeMasaya: I’m glad you asked. After I fired the auto-focus team I decided that in the future, all focusing will be contrast-detect. We hired a bunch of interns and they’re working on it.

DPR: At the moment in live view on the EOS 50D you offer a silent shooting mode that uses an electronic 'first curtain', starting with the mechanical shutter open. Is it technically feasible to offer a fully silent mode that uses a totally electronic shutter - using the sensor alone to produce the exposure?

FakeMasaya: Oh yes. In fact, that will be the key feature of the 60D that will differentiate it from the 50D. That and a video mode built in such a way that it doesn’t threaten the 5D Mk-II at all.

DPR: Will movie capture be working its way down into models further down the range? Is there a reason that the EOS 50D, for example, doesn't have video capture?

FakeMasaya: We are planning to bring this capability to each market segment when we feel we’ve successively extracted all the profit we could from the higher segments. It was important to include it in the 5D Mk-II first because otherwise it would have been easily challenged in the market – whereas now it’s easy to deflect criticism just by mentioning the video. That worked out great and our plan is to do the same with the 60D.

DPR: Are there any issues with the sensor 'heating up' when shooting extended movie clips?

FakeMasaya: I don’t know, I haven’t used it. Vincent tells me his private helicopter’s rotor cooled the camera sufficiently.

DPR: All the buzz at this year's Photokina has been around the concept of the 'mirrorless' interchangeable lens camera following the launch of Micro Four Thirds. Is this an area Canon is interested in?

FakeMasaya: It’s cool. We enjoy it when our competitors take big risks like this, so we don’t have to. If Olympus manages to make money from this, you can be sure we’ll be right there to crush them into the dust.

Fall at the Boston office front-door


The two trees in front of the office have been begging me to shoot them all week. Weather really started cooperating yesterday. It’s rare that weather makes me happy, but yesterday afternoon I had an extra bound in my step from the brisk air and glorious sky.


This morning is also the first time I that I shot my new Zeiss 100mm Makro-Planar T* lens outside for more than a test shot. The last two images are examples of its handiwork.