Jon Stern

Lives in United States Mountain View, CA, United States
Works as a Electronics Engineer
Joined on Apr 21, 2003

Comments

Total: 144, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

mgblack74: What a crock of $hit. This is the photographic version of the emperor's new clothes.

theongu, if you want to image and separate different illuminants the filters need to be on the lights, not on the lens.

I don't think they do have "full spectrum". I suspect the CRI will be extremely low. Too low to get good color reproduction, but enough so that different-colored surfaces don't have big errors in reflectance (in luma).

If they're not doing what I'm saying, then their system will have very little utility.

Link | Posted on Dec 5, 2017 at 06:28 UTC
In reply to:

mgblack74: What a crock of $hit. This is the photographic version of the emperor's new clothes.

Marcus, you could, but then the different reflectivity of colored surfaces would interfere with the effect.

For example a red surface would only be illuminated by one of the lights, and so the lighting adjustment for that part of the photo would be limited to changing the intensity of that one light.

Link | Posted on Dec 4, 2017 at 07:07 UTC
In reply to:

mgblack74: What a crock of $hit. This is the photographic version of the emperor's new clothes.

No. See my post above. After initially thinking "no way", I can see how they're doing this and what the limitations are.

Link | Posted on Dec 1, 2017 at 06:50 UTC
In reply to:

ProfHankD: Clever idea. It looks like they're just letting you do shaped masking of the individual light sources, not arbitrary relighting based on recovering a 3D model... but that's probably good enough for many photographic purposes. In theory, they should be able to recover and relight color images, but what they're doing with B&W is a piece of cake computationally and far less likely to show bad artifacting. I can see a few easy ways to implement this....

Basically, it's a hack on the usual structured light 3D capture approaches (which I'm familiar with -- I have a patent in that field). What they do nicely leverages conventional lighting practices of photographers, and it's computationally simpler than building the 3D model. It still has the same problem with surfaces that have unfortunate reflective properties. It seems they're using color to encode the different light sources, so they'll also have problems with very colorful scenes.

I suspect they're creating white lights with different, spiky spectral power distributions. They can then be extracted using custom color correction matrices that create per-illuminant channels.

While the lights can be white (as perceived by the photopic eye), spiky spectra result in poor color reproduction, and make chromatic aberration look worse.

The system also won't work well with a lot of (black body) ambient light.

Link | Posted on Dec 1, 2017 at 06:48 UTC
In reply to:

marcio_napoli: Disclaimer: I'm at my day job right now, I've watched the video without audio (this may affect my comment?).

Impressive alien tech. As a feat of technology, it's absolutely impressive.

Changing light after the shot could be the single most impressive advance in photography after digital took over 2 decades ago.

OTOH...

From a photography POV, it takes away of the last things an actual photographer could claim as a child of his/her talent, experience and hard work: understanding light, and the ability to do it well, on the fly, under pressure, on real time under the watch of an impatient client.

You take that away, and congratulations, everyone deserves a star, yay.

Another reason why pro photography is dead. Yaay.

With things like this, fake DOF on cellphones, etc, the coffin is already running out of space for another nail.

Amazing how you can kill a wonderful profession that has endured decades (centuries?) is less than 20 years.

theongu, their lights don't have to be colored. They can all have the same "white" color temperature, but consist of different narrow spectral spikes.

As I posted above, the issue is that spiky spectra result in bad color reproduction, (which is why they're B&W only), and worse-looking chromatic aberration.

Link | Posted on Dec 1, 2017 at 06:42 UTC

It took me a couple of mins, but I can see how this works. They're using narrow passband filters so that each light has it's own spectral characteristics.

White light of a particular color temperature can be reproduced with a wide range of spectral power distributions (SPDs), including very spiky light sources (with a low CRI). Those spikes can be detected in the spectral response of the image sensor and digitally separated (using custom CCMs).

There are two drawbacks of spiky spectra. The first is chromatic aberration can be more noticeable (as now you're imaging spatially discrete edges, rather than the smooth continuum that a black body spectrum produces), and secondly color reproduction errors get worse.

This second reason explains why they're only supporting black and white images.

The other issue I can see is this system will only work when there's little ambient light, as this could drown out the clear signature of their spiky SPD. This wouldn't work well as a fill light.

Link | Posted on Dec 1, 2017 at 06:35 UTC as 14th comment | 6 replies

Another Bay Area start-up solving a problem with a sledgehammer. This will go the way of Juicero.

Link | Posted on Nov 18, 2017 at 23:58 UTC as 50th comment
In reply to:

jhinkey: Decreased operating costs - mostly likely the big layoffs they had a while back finally having an effect.

That's one strategy a company sometimes has to take: develop a new product, which can take years, get it into production, then layoff a lot of employees to trim back down, especially if the next new product won't be out for a while. Do they have a head count vs. quarter buried in there somewhere?

@jhinkey, GoPro did NOT layoff the people who were working on these products. They laid off most of their loss-making media team, and other people who were working on things that had become distractions from their core business of consumer cameras.

Link | Posted on Nov 5, 2017 at 05:19 UTC
In reply to:

entoman: The way things are heading, within a few years this sort of thing will be automated and built into camera software. And then the *!"%?ing camera will decide what can and what can't be included in the picture!

The sooner we all go back to manual metering, manual focusing, film cameras and actually putting some thought into taking our photographs, the better.

Cartier-Bresson knew what the word Photography really meant.

@entoman This tool is still selectively applied by the human photographer. The artist decides what elements to remove and this tool simply makes that task faster and easier. The decision is still under human control.

Art is not the same as technical skill - that's craft. Art is all about the intention behind the piece.

Link | Posted on Oct 22, 2017 at 18:27 UTC
In reply to:

entoman: The way things are heading, within a few years this sort of thing will be automated and built into camera software. And then the *!"%?ing camera will decide what can and what can't be included in the picture!

The sooner we all go back to manual metering, manual focusing, film cameras and actually putting some thought into taking our photographs, the better.

Cartier-Bresson knew what the word Photography really meant.

This expands the definition of photography. The old meaning does not go away (even if there is the issue of knowing if an image is authentic, or not), but this moves photography closer to painting, where the artist would do exactly the same thing: decide what elements to include and leave out in a painting.

Link | Posted on Oct 21, 2017 at 18:57 UTC
In reply to:

AKH: This is for tech nerds not photograhers.

You really think so? You don't see that this is moving photographer towards traditional painting, where the artist would "make a more perfect scene", rather than just recording a literal representation of what he or she saw?

Do you think that if Claude Monte was painting that arch, he would include those people in the picture?

Photography doesn't have to be one thing. There is space for capturing the event (journalism) and for capture the feeling of an event. While it does create issues of undermining the authenticity of the former, images without an expression of the artist's interpretation is not art.

Link | Posted on Oct 21, 2017 at 18:54 UTC

A surreal bit of life imitating art, with Kumail Nanjiani (an actor who plays a programmer in the show Silicon Valley) being there on stage.

It took me a minute to realize who it was. For a while I was thinking it was an actual guy from Adobe I was chatting with when we were both with our toddlers at a local park (I'm based in the Bay Area).

Link | Posted on Oct 21, 2017 at 18:45 UTC as 62nd comment
On article Here's why your beloved film SLR is never going digital (289 comments in total)
In reply to:

Cornu: Leica DMR. That was a digital back for the Leica R9 film camera.

One interesting fact that most people don't know, is that Silicon Film originally collaborated with Leica on the DMR, and proved to them that the concept was viable. After SFT went bankrupt, Leica found another partner to complete the project.

The follow up project to the DMR was always intended to be a digital M-series camera.

Link | Posted on Oct 14, 2017 at 05:41 UTC
On article Here's why your beloved film SLR is never going digital (289 comments in total)
In reply to:

Jacob the Photographer: Somewhere early 2000's I still had my Nikon F3 cameras. Silicon Film came with a very promising digital 'film' solution. I joined them as a prospect buyer. (Crowdfunding did not exist in those days) but one day the news drip from their side died... if you look at the mechanics of a 35 mm camera , the film transport , the film guides , the shutter so close to the film plane etc. then one has to realise that bringing the imaging sensor in the film plane is near impossible and as said firing the shutter(s) simultane ....
BUT:
There is a camera that has none of those limitations and I often wondered why a HighTech developer has never come up with at least an attempt to make it digital: The TLR Rollei / Yashica / Mamiya cameras. There the shutter is far from the film plane , the shutter mechanism is very accessible (Read: Easy to modify or even to replace) , the two large 120 mm film compartments allow for plenty of electronics and battery space.
C'mon HighTecHs GO for it and surprise us !!

@Jacob the Photographer, we always felt that (e)film was a limited opportunity product. I don't think the most optimistic of us expected there to be much of a market for it by 2017.

If anything, the DSLRs, and now mirrorless cameras, advanced quicker, and fell in price faster, than anyone could have predicted, making the serviceable market in 2017 even smaller.

There is still a market for it, but it's too small to support a viable business. Firstly, it would take several million dollars to re-start development and get something decent to market. Then the low volumes for the product would mean that the unit price would have to be too high for what it offers (compared to a DSLR, or MILC).

I certainly wouldn't quit my job at a well-known camera company, to join an attempt to resurrect the concept. The world has changed, and there are more-exciting frontiers in imaging today.

Link | Posted on Oct 14, 2017 at 05:35 UTC
In reply to:

ozturert: This looks so cool, unlike GoPro and it's copies. If price is right, Casio may be my next buy.

What do you mean? Do mean the features or the industrial design look cool?

Link | Posted on Oct 14, 2017 at 05:18 UTC
On article Here's why your beloved film SLR is never going digital (289 comments in total)
In reply to:

Jacob the Photographer: Somewhere early 2000's I still had my Nikon F3 cameras. Silicon Film came with a very promising digital 'film' solution. I joined them as a prospect buyer. (Crowdfunding did not exist in those days) but one day the news drip from their side died... if you look at the mechanics of a 35 mm camera , the film transport , the film guides , the shutter so close to the film plane etc. then one has to realise that bringing the imaging sensor in the film plane is near impossible and as said firing the shutter(s) simultane ....
BUT:
There is a camera that has none of those limitations and I often wondered why a HighTech developer has never come up with at least an attempt to make it digital: The TLR Rollei / Yashica / Mamiya cameras. There the shutter is far from the film plane , the shutter mechanism is very accessible (Read: Easy to modify or even to replace) , the two large 120 mm film compartments allow for plenty of electronics and battery space.
C'mon HighTecHs GO for it and surprise us !!

@Roland, yes it did have a small sensor. That's why it was EFS-1. We already had the working samples of a larger sensor (similar size to APS-H) that would have gone in to EFS-2.

The sensor package visible in the package above was actually designed for that larger sensor. The black mask that you can see on sensor cover glass, that was eliminated for packages with the larger die (so the size was pretty much the size of that black region).

You skepticism is understandable and fine. I'm familiar with it by now.

Link | Posted on Oct 13, 2017 at 01:32 UTC
On article Here's why your beloved film SLR is never going digital (289 comments in total)
In reply to:

Jacob the Photographer: Somewhere early 2000's I still had my Nikon F3 cameras. Silicon Film came with a very promising digital 'film' solution. I joined them as a prospect buyer. (Crowdfunding did not exist in those days) but one day the news drip from their side died... if you look at the mechanics of a 35 mm camera , the film transport , the film guides , the shutter so close to the film plane etc. then one has to realise that bringing the imaging sensor in the film plane is near impossible and as said firing the shutter(s) simultane ....
BUT:
There is a camera that has none of those limitations and I often wondered why a HighTech developer has never come up with at least an attempt to make it digital: The TLR Rollei / Yashica / Mamiya cameras. There the shutter is far from the film plane , the shutter mechanism is very accessible (Read: Easy to modify or even to replace) , the two large 120 mm film compartments allow for plenty of electronics and battery space.
C'mon HighTecHs GO for it and surprise us !!

It was possible to bring the sensor close to the film plane. This is helped because the sensor cover glass shifts the plane backwards by 1/3rd of it's thickness. At Silicon FIlm, I co-designed a custom, thin ceramic package that worked.

The film transport wasn't an issue because all SLRs would allow the camera to operate without a film loaded. Detection of a loaded film cartridge used switches in that were usually combined with the DX-coding contacts. If you look at that photo of the Silicon Film (e)film cartidge above, you can just about make out a depression (surrounding the gold download contacts) in the film can region that means the film load switches don't get depressed. The camera will therefore operate as though there is no film loaded (which is full operation intended to allow practice shooting).

Silicon Film failed because we ran out of money after the dot-com crash. Not because of any fundamental engineering issues.

Link | Posted on Oct 12, 2017 at 01:39 UTC
In reply to:

TIMagic: Still no optional date and time stamp display like many of their overseas competitors offer? I have no idea why GoPro has ignored this suggestion I have mentioned in numerous communications (email, app, telephone) over the past 2 years on their cameras.

If you're wanting to overlay the date and time on the video, that is available under the Gauges in Quik for Desktop.

Look at the "INFO" gauge on this page: https://gopro.com/help/articles/Block/How-to-Use-Gauges-in-Quik-for-Desktop

Is that what you want?

Link | Posted on Sep 29, 2017 at 17:55 UTC
In reply to:

Lawson Raider: I hope this isn't what the Hero 6 will look like, it is twice the size of the Hero 5 and will be a bulky monster.

It's Fusion. GoPro's new spherical capture camera.

https://www.wired.com/story/gopro-fusion-360-hands-on/

Link | Posted on Aug 8, 2017 at 15:26 UTC
In reply to:

Alberto Tanikawa: A curved sensor would not reduce vignetting, or improve peripheral illumination from lenses. That is dictated by the lens design itself. Light fall-off towards the edges will be exactly the same whether on a flat sensor, or elevated by a few millimeters on a curved sensor. More modern lens designs mitigate the need for such sensors. Not to say they are useless, far from it. Matching a telescope to a specific curved sensor should make coma a thing of the past.

@Enginel, polarization is not easy to achieve at the pixel level. It is an area of research though.

It's difficult to do better than the standard Bayer pattern. Attempts to do so have rarely gone far.

The two green diagonal pixels with a red and blue ends up being a good approximation for the human photopic response. If you try to add more colors you have a lower spatial sampling frequency in red, green and blue. Which would result in more moire, not less.

Link | Posted on Jul 17, 2017 at 07:07 UTC
Total: 144, showing: 1 – 20
« First‹ Previous12345Next ›Last »