Joseph S Wisniewski

Lives in United States Detroit, MI, United States
Works as a Speech and sensor scientist
Has a website at
Joined on Jul 12, 2002
About me:

I've been in love with photography for over 40 years.

My favorite form of photography is macro, and much of my gear is of my own design.

I've done professional photography on and off for over 20 years.
Taught 8 years at Midwest Photography Workshops.
Designed 3 scientific digital cameras and 6 lenses.

Author of mImage and (coming soon to the App Store and Google Play) ColorForEveryone.


Total: 681, showing: 341 – 360
« First‹ Previous1617181920Next ›Last »
In reply to:

Dan Tong: Even if you have no interest in buying a Lytro camera you have to be impressed by the technological breaktrough that it represents, as well as how fast it went from the laboratory to an actual commercial product.

It's also a good indicator of the stupidity and arrogance of many of our posters, who smirk and insist on airing their ignorance whenever Lytro is discussed.

It reminds me of the many inane comments about other early digital camera innovations which these like minded "geniuses" dismissed.

It is often associated with the attitude that "If I have no interest in it (fill in whatever) then it is of no interest to anyone else and it is just plain dumb".

I can imagine a not too distant future where Light Field technology is used to make films and the director and editor will be able to change the DOF as well as some of the perspective after the film is shot.



I still have a Nimslo. It's very pretty on a shelf. There was still enough market after they folded, and the technology was simple enough that people rolled their own processing software (like Lentils) and other 3 and 4 shot lenticular cameras.(i3D, Nashika) that the market kept struggling along for 25 years. Heck, Fuji even has a lenticular printing service for their W3 stereo cameras.

Lytro will fail more spectacularly. The algorithms are much more complex, the audience smaller. There's less potential for post-shot money, no prints, and pay-to-host on a cobbled up Lytro gallery site won't fly. When they die, they're staying dead.

Link | Posted on Nov 18, 2012 at 14:27 UTC
In reply to:

Hugo808: Everyone sneers, but in a few years Lytro will have sold this to Canikon and it will be a standard feature of the D9000 and everyone here will love it just as they eventually loved autofocus....

It will never be standard on a DSLR. Read Dr. Ng's dissertation to understand why. The plenoptic lens array has to be set up to decimate a particular maximum aperture down to a particular minimum aperture. For the thesis camera, he went from f4 down to f22, and that took about a 13:1 decimation. The production camera does only 4:1 on an f2 lens, so that you still have some image left (about 1mp) after decimating a 16mp sensor.

Canon DSLRS cover the range from f1.2 to f32. That needs, according to Dr. Ng, 50:1 decimation. Put the plenoptic array in front of a 250mp sensor from the far future, and you get 0.1mp out.

Basically, the technology is not suitable for DSLRs. His own dissertation explains why.

Link | Posted on Nov 18, 2012 at 14:18 UTC
In reply to:

wakaba: So here we have a camera that is really bad at taking pictures and uploads your picture, denies ownership and only lets you see trough a browser window.

What will happen to your pictures in 5 years? Kodak went bust...Flash is getting phased out...Firefox will change...

Raw and JPEG win. For the price of a Litro a very decent SLR can be had.

No go unless your a noveltyseeker drone.

Somebody else drew an analogy to Nimslo as an innovator elsewhere in this thread. So, I'd say the business will go the same way. After Nimslo went under, a few small labs tried to keep up with the processing, but the camera soon became basically an expensive paperweight.

When Lytro goes bust, unless some really ambitious folks try to duplicate what they did with their site or a tablet/smartphone app, the camera becomes a colorful brick.

Link | Posted on Nov 18, 2012 at 14:05 UTC
In reply to:

TotallyFred: I think that the 3D aspect of it is the most compelling aspect for now -- at least it can be grasped by many and has more immediate applications the public/viewers could want to use.

The effect seems limited however.

It would be interesting to understand more about the "shifting" capability. Is it really shifting or rotating or a shift/rotate ?

It's a real shift, just a very small one. A bit under 3/4 the width of the lens. More properly, 3/4 the width of the entrance pupil of the lens. That means, for a constant aperture lens, the amount of shift you can have decreases as you zoom wider. The pupil of the Lytro ranges from 3.8mm at its 43mm normal setting (that's as wide as it gets. The lytro idea breaks down totally on wide angle lenses) to 30mm at its longest telephoto setting.

So, you can have a stereo base of 3mm for normal, 23mm for the longest tele shot. By comparison, the stereobase of human vision is about 65mm, the intrapupil distance of our eyes. So, for actual 3D, the Lytro only works at close up or macro distances. At the normal setting, you get the best 3D at a distance of about 90mm (a bit under 4 inches). In the extreme tele setting, you can go all the way out to 700mm (2.5 feet). Perfect for flowers, or pictures of your food.

Want to take a picture of humans? Get a Fuji W3.

Link | Posted on Nov 18, 2012 at 13:59 UTC
In reply to:

marike6: I'd rather have a RED Epic. RED is the only company I know of that is actually upgrading existing user cameras with new sensors (see Dragon sensor).

It's nice to see a company address the problem of camera obsolescence in a big way.

Yeah, it's amazing how anyone could call BS on Red. They delivered their Monstro series of 35mm, 645, and 617 sensor cameras exactly as promised. Scarlet lived up to all its promises, too, as did their still camera series.

Their ability to deliver the goods is epic. I'm sure this sensor upgrade is every bit as real as the Monstro.

Link | Posted on Nov 2, 2012 at 03:23 UTC
In reply to:

zoranT: why isn t global shutter implemented in all cameras? too expensive tech? what s the disadvantage?

No, all CCD sensors do not have global ("snap" is a better word) shutters. Medium format CCDs don't. Nikon D2X didn't.

Snap shutters have nothing, at all, to do with being CCD or CMOS. There was a time when sensors and cameras had electronics so slow that they took pretty much near a movie frame time (1/24, 1/25, or /130 sec) to read out the image. So, doing a "rolling shutter" could result in the entire image tearing across the frame, from upper left to lower right. The only way to make an acceptable video camera was to add snap shutter circuitry, even if it lowered image quality or increased the size of the chip.

As camera and sensor circuitry got faster, rolling shutters had less impact, and people found the tradeoffs between tear and other image quality issues (blooming, low light noise) to not be so cut and dry, and you had both kinds of cameras.

Link | Posted on Oct 31, 2012 at 17:31 UTC
In reply to:

SHood: So how long until this global shutter sensor shows up in cameras? That is the big question.

If they implemented it in the way Sony described it in their papers, it doubles the size of the chip. So an APS camera like these "Super 35" F cameras has a FF sized sensor.

If you don't mind paying FF prices for an APS camera, you can have this in a still camera, too.

Link | Posted on Oct 30, 2012 at 22:39 UTC
In reply to:

mholdef: I think Canon has just been eclipsed here...suppose DSLR video had its time...

Yeah, a $15k and $25k APS video camera prove that $400-1200 APS DSLRs "had its time".

Sort of the way Panavision has been proving that there's no market for 16mm or 8mm movie gear, or camcorders, for over 1/2 a century.

Link | Posted on Oct 30, 2012 at 22:36 UTC
On article Reviewers offer first look at Microsoft Surface tablet (121 comments in total)
In reply to:

M Lammerse: I do own the ipad (3rd generation) and I bet it can't reach the hardware of the ipad by means of screen and battery life. But if you are not an apple junkie and you are more font of connectivity, compatibility and integration with your (Windows 8) PC - I think even this Windows tablet gives a better choice.

That this is a again written by an Appletologist (almost a scary Scientology level devotion) makes me wonder how much trust I have to put in this preview.

A couple of things: if you want to mention a huge number of apps, don't mention Apple but mention Android.
No it will of course not run IOS nor Android apps. But it will give you the freedom to install what you want and how you want it.

The intel based Surface Pro will. The ARM based RT tablets run app store apps, microsoft's way only. Not really that different from Apple or Android...

Link | Posted on Oct 25, 2012 at 03:39 UTC
On article Reviewers offer first look at Microsoft Surface tablet (121 comments in total)
In reply to:

DougRight: It boils down to this - will we see Lightroom on an ARM? If not, then the surface photographers will want is the Intel version, not this Surface RT.

On a side note, I personally know well the appeal of the tablet set up, but after throwing money away on accessories for my iPad, I've come to the conclusion that for content creation purposes, nothing beats the portability and utility of a notebook out in the field. A Mac Airbook is nearly as thin & light as a tablet, and far far more productive.

I just finished my first RT app. Based on that experience and knowing what I do about raw processors, I'd say Lightroom isn't doable on RT.

Link | Posted on Oct 25, 2012 at 03:32 UTC
In reply to:

PicOne: I wonder if there's anyone working on development of a curved (parabolic?) sensor; isn't the fact that sensors are flat, and can't take lens light from oblique angles, the cause behind why larger sensor compacts can't be made? Ie.. compare to the old Olympus Stylus Epic ("fullframe").

It's in any optical engineering text. I'd recommend Warran J. Smith.

Now, the conversion of an inherently planar projection lens to a curved projection by adding a cylindrical anamorphic lens that's larger and heavier than the original projection lens might fit within your definition of "insanely easy", but I'm tempted to drop both the "ly" and the "easy".

So, what inherent defect of sensor designs is going to be addressed by such a Rube Goldberg contraption?

Link | Posted on Oct 19, 2012 at 22:27 UTC
In reply to:

PicOne: I wonder if there's anyone working on development of a curved (parabolic?) sensor; isn't the fact that sensors are flat, and can't take lens light from oblique angles, the cause behind why larger sensor compacts can't be made? Ie.. compare to the old Olympus Stylus Epic ("fullframe").

You can't remember where you read it, because it's wrong, which means that you didn't read what you think you read.

The central part of the retina is flat. The outer part just is basically used for motion sensing, so it doesn't matter that curvature of field makes it soft. Being spherical is a major optical problem for the eye. It's a side effect of the eye being able to swivel in its socket. It's all about motion and tracking, not "image quality", LOL...

Link | Posted on Oct 18, 2012 at 15:34 UTC
In reply to:

Torkn Photo: In what universe is a 1" sensor even remotely close to 1 inch?

In my day, 1" was 25.4mm, and the so called 1" sensor is only around half that size. Even the diagonal is much smaller than 1".

In your universe, LOL.

The universe where, for a century or more, it's normal to talk about the width of the film, like 35mm, despite the cameras not taking an image that's 35mm diagonal, 35mm wide, or 35mm tall. 35mm covers 24mm tall formats in widths ranging from 18mm (Oly, Yashika, and other half frame) on up to 24x65 (Fuji XT/Xpan). It also includes 28x40 on cameras that used unperforated 35mm film, like Kodak 828, 20.955x15.24 academy motion pictures, and 24.89x18.65 "super 35".

So, video folks did the same thing, and talked about the width of the tube, something you can just whip out a ruler and check. Early digital cameras borrowed video sensors and the video terminology that came with them.

@robjons, forced? Really? I had no idea you were so formidable. I'll be watching eagerly to see how you manage this.

Link | Posted on Oct 17, 2012 at 13:47 UTC
In reply to:

PicOne: I wonder if there's anyone working on development of a curved (parabolic?) sensor; isn't the fact that sensors are flat, and can't take lens light from oblique angles, the cause behind why larger sensor compacts can't be made? Ie.. compare to the old Olympus Stylus Epic ("fullframe").

Nope, not the cause, at all. A curved sensor has limitations with short lenses or fast lenses, and it can only match the exit pupil of a single fixed lens: it's essentially useless on interchangeable lens cameras or zooms.

Plus, optical projection is inherently planar, and lens design for curved sensors is insanely difficult.

But thank you for not mentioning the eye. Usually, when someone asks about curved sensors, they bring up the eye being curved...

Link | Posted on Oct 17, 2012 at 13:19 UTC
In reply to:

Chekr: I would actually give this a try if the images format was an open this stage you can either view the images in the desktop program from Lytro or export them as flash files. Given that flash is well and truly on the decline and Lytro could disappear next year, remind me why I would commit my memories to this technology with proprietary image formats...

What would you do with the image data if it were an one standard? The system is effectively dead, no one else is ever going to write software to deal with the files.

You'd need them not only to open the standard, but to open the source to the software.

Link | Posted on Oct 13, 2012 at 01:56 UTC
In reply to:

CameraLabTester: The greatest challenge for the Lytro LF Camera is to find a really truly genuine USE for it.

The Wankel was great. The Segway was great.

The Lytro LF is great!

...great gadget, users please apply within.


That is not fair. There are some practical uses for the segway, and Wankel engines do have pretty good performance.

Link | Posted on Oct 9, 2012 at 23:10 UTC
In reply to:

facedodge: Heard they have the fastest Autofocus around.


No. Their autofocus is slow, and inaccurate. It shouldn't be: the plenoptic system essentially makes the entire sensor a giant phase-detection autofocus sensor.

I think you're trying to be sarcastic, and make a statement about the lytro not actually needing to focus, which is not true. It only increases DOF about 3-4 stops worth vs wide open, and their lens is fast enough and long enough that wide open, the DOF is pretty shallow. It needs good AF to put the subject within that shallow DOF so you can do at least some "refocusing".

Link | Posted on Oct 9, 2012 at 23:06 UTC
In reply to:

Tom view to a still: Light Field technology??? What isn't light field technology when it comes to photography?

I have heard the hoopla about this for a long time now. They don't explain anything really... bunch of techie double talk that really says nothing. I think the emperor has no clothes.

My guess at the big secret... focus bracketting. Several images taken simultaneously or near simultaneously. Software that uses the images to "slide the focus" (or de-focus) arcoss the multiple images. Give it a cool tech sounding name. You've got a new pet rock.

So far sounds like another red herring, time wasting gizmo. But.. love to see a true critical review. My mind is still open to facts, just not slogans.

If you had checked, like you accuse Tom of not doing, you'd have realized that the amount you can actually shift a POV with a lytro is about 10mm, making that feature useless at anything other than macro distances.

Past that, and you run into exactly the same occlusion problem you have converting a focus stack to 3D: an image that looks extruded rather than actually dimensional.

Focus stacking and lightfield imaging both produce volumetric data. The processing algorithms are similar.

Link | Posted on Oct 9, 2012 at 23:00 UTC
Total: 681, showing: 341 – 360
« First‹ Previous1617181920Next ›Last »