ProfHankD

ProfHankD

Lives in United States Lexington, United States
Works as a Professor
Has a website at http://aggregate.org/hankd/
Joined on Mar 27, 2008
About me:

Plan: to change the way people think about and use cameras by taking advantage of cameras as computing systems; engineering camera systems to provide new abilities and improved quality.

Comments

Total: 1396, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

brn: Like many, I've not heard of them before now. I just did a search on DPR and only found an article on their demise. Did DPR not know about them either? Perhaps DPR was too busy posting articles about NAS and cell phone reviews.

I never heard of them before either. Looking at a couple of issues, it's really impressive. Coffee-table worthy. Unfortunately, I'm not sure the words "financial viability" come easily for an art publication....

Link | Posted on Aug 15, 2017 at 02:00 UTC

Cell phones definitely do a good enough job for smallish online images, and computational photography methods will continue to make them better, but you really can't get past basic physics: you need enough photons. You can collect more by sampling over a larger area (bigger sensor or bigger aperture... or arrays of smaller ones) or for a longer time (as in my TDCI work). Both those tricks can be used to some extent in the cell phone form factor, but a bigger form factor can do better.

As for ILCs being doomed, well, maybe -- but that's because you can tune the camera system better as a whole when the sensor and lens are designed as one system (at the very least, ILC lens designs are becoming more camera-specific). For example, for the upcoming eclipse, I'm having a hard time matching the IQ of a Canon PowerShot SX530 HS zoomed to "1200mm" with my A7RII or 5D IV and the longest lenses I have. The SX530 cost me $130 as a Canon refurb.

Link | Posted on Aug 14, 2017 at 14:05 UTC as 296th comment
In reply to:

BBQue: And the copyright to this video belongs - no doubt - to the tiger, right?!

Nope. The photographer initiated the capture. The tiger simply wasn't very good at taking direction as a model... or maybe it was VERY good at that?

Link | Posted on Aug 13, 2017 at 17:07 UTC
In reply to:

Gimli son of Gloin: Very nice photos.

I would recommend a 1 inch camera instead of mFT since the IQ and resolution are very similar and all 1 inch cameras are more capable and portable than this kit.

Since the difference in IQ is not very big in 1 inch vs mFT a huge step up in IQ would be to FF. But then we would be seeing an increase in size compared to 1 inch, not so much compared to mFT.

Dheorl: "People are people, so why should it be you and I should get along so awfully?" -- Depeche Mode. I think people are generally insecure about their big purchases; they need reassurance that they have made the one best choice.

Thanks to my research, I have the luxury of buying multiple things and using the most appropriate for each task... which explains how I can praise Sony over Canon for sensors and adapted lens abilities while praising Canon over Sony for reprogrammability via CHDK and ML.

MFT is at a bit of a disadvantage because it was based on what turned-out to be wrong guesses about cost of larger sensors and dominance of the 4:3 aspect ratio, but they've really pushed the convenience features and small form factor (especially for lenses) very far. Those just don't happen to be major advantages for most of what I do with cameras. The 4K photo mode was why I bought the GX850; I'm disappointed that's a movie mode and adapted lenses always suffer rolling shutter.

Link | Posted on Aug 13, 2017 at 15:26 UTC
In reply to:

Gimli son of Gloin: Very nice photos.

I would recommend a 1 inch camera instead of mFT since the IQ and resolution are very similar and all 1 inch cameras are more capable and portable than this kit.

Since the difference in IQ is not very big in 1 inch vs mFT a huge step up in IQ would be to FF. But then we would be seeing an increase in size compared to 1 inch, not so much compared to mFT.

String: As a computational photography researcher, I make a lot of measurements (e.g., OOF PSF). I appreciate what DxO has done in openly publishing test results and I have the greatest respect for Imatest. However, I once was a professional photographer and am now an avid amateur; that perspective drives my research and determines which measurements I care about.

A couple hundred cameras are in my personal and research lab collections. I use more Canons than any other brand, mostly PowerShots that I and my students reprogram using CHDK. Other brands: Unibrain, DLink, Sony, Nikon, Olympus, etc. My latest MFT body is a Panasonic GX850, which has been unexpectedly disappointing. I have yet to be seriously disappointed by a Sony; my last five personal workhorses have all been Sonys: A100, A350, A55, NEX-7, and A7II.

Of course, my biased opinions are my own, not necessarily those of my university, and they are subject to change as observations warrant. ;-)

Link | Posted on Aug 13, 2017 at 01:31 UTC
In reply to:

Gimli son of Gloin: Very nice photos.

I would recommend a 1 inch camera instead of mFT since the IQ and resolution are very similar and all 1 inch cameras are more capable and portable than this kit.

Since the difference in IQ is not very big in 1 inch vs mFT a huge step up in IQ would be to FF. But then we would be seeing an increase in size compared to 1 inch, not so much compared to mFT.

The little Sony RX100 V (or one of its older brothers which also have merely 1" sensors) produces IQ that compares well with that of a PEN-F with 12-40. Most DxO measurements are very similar, and I'd sooner take an RX100 IV or V than a PEN-F with just the 12-40. Then again, I'd even sooner take an A6500, A7II, or A7RII -- all of which are enough better to make a qualitative difference -- but I rarely bring less than 3 pounds of camera equipment on a multi-day trip. In other words, there are 1" sensor cameras with comparable IQ, and APS-C and FF with markedly better IQ (I find the big step is MFT to APS-C and your mileage may vary), but the PEN-F is a perfectly reasonable choice.

Link | Posted on Aug 12, 2017 at 22:04 UTC
In reply to:

ProfHankD: Yet another indication that Sigma is still having to reverse-engineer lens interface stuff. That has been a very long-term problem for Sigma....

Ilia Snopchenko: I don't think Zeiss has ever had a relationship with Canon. On the other hand, Canon was the first with fully electronic lens control, so there's been more time to reverse engineer and quite a few folks have pretty much gotten the protocol. Canon still "requests" those who know don't publish it. The catch is newer models may use extensions that were not observable in earlier models, hence were not reverse engineered.

Link | Posted on Aug 11, 2017 at 14:27 UTC
In reply to:

ProfHankD: Yet another indication that Sigma is still having to reverse-engineer lens interface stuff. That has been a very long-term problem for Sigma....

jennajenna: Sigma is a family-owned business, but claims to be the largest independent lens manufacturer -- they don't seem to collaborate with any camera makers (they do make cameras of their own). However, many 3rd-party lensmakers have close relationships with camera makers. My understanding is that Tamron in particular has had a number of deals where they made lenses sold under camera-maker brands, so they got the info needed.

For example, while Sigma has a long history of Sony incompatibilities, neither Tamron nor Zeiss do -- but they are both 3rd parties who have been involved in producing Sony-branded lenses. Such collaborative deals are common, and that's often how the bigger 3rd parties get access to specifications.

In sum, it's common that smaller 3rd-party companies (and 3D-print makers like me) must reverse engineer, but it's quite odd that a lensmaker as important as Sigma still must....

PS: Sony's E specs are mechanical only, and available only to approved companies.

Link | Posted on Aug 11, 2017 at 03:16 UTC
In reply to:

J A C S: I love those wavy light rays in the diagram. This is what happens if you are too close to a black hole.

Yeah, not even then. The waves also change color on the way. Apparently, the Crystal Clear Lens also has some odd stair-step pattern going to the hole in the middle... wait. Hole in the middle? Maybe they should let an engineer look at the illustration after the artist has made it and before they advertise it?

Link | Posted on Aug 11, 2017 at 01:01 UTC

Yet another indication that Sigma is still having to reverse-engineer lens interface stuff. That has been a very long-term problem for Sigma....

Link | Posted on Aug 11, 2017 at 00:50 UTC as 14th comment | 9 replies

Sounds like a modern re-make of the old 60mm macros from Mamiya/Sekor, Yashinon, and Tomioka -- except they were f/2.8 and went to 1:1. Note sure this lens made the right choice on that tradeoff. The old 60mm usually run about 1/3 the price of this new, so the price isn't too bad for this.

Link | Posted on Aug 8, 2017 at 12:29 UTC as 44th comment
In reply to:

OliverGlass: "(But Camera companies are not your friends)..."
With the exception of FUJI which has been generous with firmware updates that add more features and are never afterthoughts. They listen to their users too.

cosinaphile: Fuji didn't fix things within a 45-day window -- ignoring early reports (many of those centered on the black holes, which were even worse but were repaired by a firmware update in 12/2011), there were LOTS of white orb complaints by 10/2011, the ineffective firmware fix was 2/25/2012, and the sensor replacement program started in 4/2012. I can't imagine Fuji didn't know the problems before the first unit was shipped, but it was still at least six months to a real response to customer complaints. Even after that, they continued selling the original sensor version side-by-side with the better (not good) one, without distinction, until product end-of-life.

In sum, I see the X10 sensor as a commendably bold, but clearly failed, experiment... that Fuji kept selling anyway. The lesson I think they learned is that product functionality doesn't matter as much as the company appearing to be responsive to customers. Is that good? I'm not sure....

Link | Posted on Aug 8, 2017 at 00:31 UTC
In reply to:

OliverGlass: "(But Camera companies are not your friends)..."
With the exception of FUJI which has been generous with firmware updates that add more features and are never afterthoughts. They listen to their users too.

cosinaphile: A respectable company would not have sold a known-defective product, not lied about the firmware fix, and done a full refund/replace, not offer major surgery on your new camera. I'm sorry, but there is just no getting around the fact that Fuji handled this exceptionally badly.

Ironically, the repair algorithm in DeOrbIt, which I made freely available, was feasible for them to have used for a firmware fix -- and usually produced better results than their revised sensor (when that became available). I don't know why Fuji didn't put the DeOrbIt algorithm in either a firmware update to fix all cameras or in their software so existing photos could also be repaired.

Link | Posted on Aug 7, 2017 at 20:05 UTC
In reply to:

OliverGlass: "(But Camera companies are not your friends)..."
With the exception of FUJI which has been generous with firmware updates that add more features and are never afterthoughts. They listen to their users too.

Short memory. Fuji shipped known bad cameras (X10 and X-S1 "black holes" and "white orbs" problems), released a completely ineffective firmware fix, and finally only offered a service fix to owners that complained -- while continuing to sell the defective version with no consumer-visible differentiation on the packaging between the defective and revised (much less defective) version. I suspect they would never have done anything if DPReview hadn't made such a stink (or perhaps if I hadn't released the free DeOrbIt software to computationally repair the defects -- they announced a service fix a week after DeOrbIt's release and started offering repair the following month). In any case, good if they have learned not to do that again....

Link | Posted on Aug 7, 2017 at 03:04 UTC

Well, I think Canon's lack of 4K is largely due to an older computational infrastructure in their cameras. Basically, there is what engineering can do, what manufacturing says it will cost to produce in quantity, and how marketing thinks it can be profitably sold. Most companies fall into patterns -- winning formulas -- for how to trade-off these different forces, and that's a lot of the difference we see between brands.

Link | Posted on Aug 6, 2017 at 18:51 UTC as 99th comment
In reply to:

cgarrard: Man we need a lot more info on the net like this. The internet has become a fantasy land for so many.

I think phrase now is alt-reality... but that's probably fake news.

Link | Posted on Aug 6, 2017 at 15:56 UTC
In reply to:

joe6pack: Impressive! I'm actually quite surprised the fire works out so well with all the flames instead of just 1 big blur like light painting. Now I want to play with fire.

That surprised me too. Only a few hints of horizontal movement....

Link | Posted on Aug 6, 2017 at 15:43 UTC
On article Photo of the week: The Purple Hour (121 comments in total)
In reply to:

entoman: A classic photo Erez. I love sub-arctic and arctic landscapes (provided that the HDR isn't overdone - a trap many fall into), but it's too cold for my liking - I'll stick to temperate and tropical climates!

If I can be permitted a minor criticism - I don't like the starburst effect on the moon, it seems quite unnatural. Starburst works well sometimes with the sun, but not the moon. It is of course a well known feature of the 16-35mm Canon optic, but not one that I admire.

I agree about the starburst... but then, I generally see starbursts as the annoyingly structured flare that they really are.... ;-)

Link | Posted on Aug 5, 2017 at 17:14 UTC

Another solid entry from Laowa. Not as impressive as some of their others, but 15mm field of view, f/2, and it takes a filter is a very friendly combo and MFT users generally prefer friendly over awkward and a little higher IQ.

Link | Posted on Aug 5, 2017 at 13:58 UTC as 16th comment

I'm a big proponent, and developer, of free software. I put most of my software out as open source with fewer usage constraints than a copyleft. I also post my 3D printing designs for free. However, Unsplash is a hard thing to get used to.

First off, I don't see what their business model is -- and there should be one, because it is a .com, not a .org. Second, I don't quite get the verbiage about they will handle license issues, because I don't see where they'd get money to pay for any court battles. Third, I'm honestly uncomfortable giving blanket permission for use of my photos... which is strange, given how comfortable I am giving away so much of my work product that is more valuable, but I guess they're often "personal" enough that it would bother me to see them used in certain ways.

In sum, I have to think about this....

Link | Posted on Aug 5, 2017 at 02:03 UTC as 45th comment
Total: 1396, showing: 1 – 20
« First‹ Previous12345Next ›Last »