"Shape, Shadow, Form, Color" - Light L16 photos

Godfrey

Godfrey

somewhat colored
Local time
6:07 PM
Joined
Dec 15, 2011
Messages
13,557
Location
Silly Valley, California, USA
View My Gallery
I'm never quite sure where to post about the Light L16. As a camera, it crosses various boundaries being a multi-camera, computational photography digital camera. So I'll just post here in Photography General Interest and see what people have to say. 🙂

---

It'd been a long while since I used it: Since my arm is healing but still very weak I needed something small and light to try taking some photographs on a walk. Yesterday I pulled my Light L16 out of its box for the first time in a year and a half, charged it up, and reinitialized it, installed the most recent version of Lumen on macOS Catalina, and went for a walk with it.

It still makes me chuckle that this little camera in its case ...
50141863536_da35431e5a_b.jpg
... is a 50Mpixel camera with a zoom FF equivalent of 28 to 150mm. It's only a little bigger than a plus-sized iPhone 8.

I found 14 exposures from my walk yesterday that fit a theme of "Shape, Form, Shadow, Color" that I liked. Here are four of them:


Turret Windows


Beware of Dog


Leaves Akimbo


Pavement Marking​

For the rest, please visit the album on Flickr.com : Shape, Form, Shadow, Color - 2020

It's interesting to compare the shooting experience of the Light L16 with the Leica CL, and the imaging qualities with that camera and the Hasselblad 907x. I'll think about that and write it up at another time. 😀

Enjoy!
G
 
I have problem with digital, I just don't like high contrast (low dynamic range) . Of course, there are many programs that adjust this, but it is a boring adventure. Do you think that maybe the digital (linear) company's are just trying to get us use their problems?. Or is it a digital camera industry tegument?

You did a good conformation with these images.
 
charjohncarter said:
I have problem with digital, I just don't like high contrast (low dynamic range) . Of course, there are many programs that adjust this, but it is a boring adventure. Do you think that maybe the digital (linear) company's are just trying to get us use their problems?. Or is it a digital camera industry tegument?

You did a good conformation with these images.
Click to expand...

I'm not entirely sure I understand, John.

  • ".. maybe the digital (linear) company's are just trying to get us use their problems?"
  • "Or is it a digital camera industry tegument?"
  • "You did a good conformation with these images."
I'm not sure what these questions and statements mean.

The Light L16 is essentially a collection of small sensor cameras arrayed together and used simultaneously, and whose individual outputs are coalesced into a single image file according to Light.co's proprietary algorithms. As such, it is effectively a 13 bit quantization capable of achieving 8192 levels of gray per picture element. Other cameras in my cabinet achieve 14 (16384 levels) and 16 bit quantization (65536 levels) so the L16 is without a doubt a bit on the skinny side when it comes to dynamic range as a baseline.

Exactly what one does when rendering the image files produced by any digital camera affects the end result profoundly. It's difficult to pose sweeping generalizations about the dynamic range of all digital cameras when even in the small set of three cameras I own and use, I can obtain a DR range between them that varies by up to a factor of 8x in their raw capture form. And I can adjust all of them to cover as soft a contrast curve or as hard a contrast curve as I might want by manipulating the camera calibration profiles and higher level image processing adjustments in whatever image processing software I want, up to their individual quantization limits.

On that basis, whether you like hard or soft contrasts in a photograph is up to your sense of aesthetics and your skills at rendering the image. The analog to these operations in image processing when you consider film and chemical photography is how you manipulate and use the exposure index, which chemistry at what dilution, how you agitate, how much time you process, and then how you print the results running against the baseline response curve of whatever film you've chosen to work with. The implementation and means to the rendering results are very different, but the models, and end results, are quite similar.

But I'm just speaking to the questions and statements you made there as best I can understand them, which likely isn't particularly well. The real question I have for you is:

Is there some problem you're seeing with the contrast or dynamic range in these photos? Please articulate that for me, if there is.

thx!
G
 
I guess I left out the the word 'to' in your first comment. Any new product is, just by being new, is going to present change that the public isn't use to. So, either they just keep making that product and assume that the public will get use to it. Or the try to fix it. If the fix doesn't work then they pretend to cover it up.

You certainly have been able to get what you want out of digital, I still struggle and have to go to programs for a compromised consolation prize.

I have no problem with what you do, in fact, you have always been a master at digital. I remember you from your Pentax (on the Pentax forum) days, and was truly amazed then.
 
charjohncarter said:
I guess I left out the the word 'to' in your first comment. Any new product is, just by being new, is going to present change that the public isn't use to. So, either they just keep making that product and assume that the public will get use to it. Or the try to fix it. If the fix doesn't work then they pretend to cover it up.

You certainly have been able to get what you want out of digital, I still struggle and have to go to programs for a compromised consolation prize.

I have no problem with what you do, in fact, you have always been a master at digital. I remember you from your Pentax (on the Pentax forum) days, and was truly amazed then.
Click to expand...

Thank you for the compliments! 😀
Digital imaging is so much a matter of mathematics rather than chemistry that it falls into my brain very easily. I was always better at math and physics than chemistry. LOL!

Frankly, I think the companies keep doing what they do with their standard in-camera processing and the default camera calibration specs for raw conversion because most of the buying users are relatively insensitive to these nuances of the imaging and don't have a context of a 'standard film camera' history to lead them otherwise. The modern aesthetic does seem pointed more at photos with a lot of contrast and saturation, regardless of what the underlying capture technology is. Remember that the audience of picture takers who buy cameras has basically exploded way way beyond anything that existed in the film-only era, and the demographic of those new photographers is on the youngish side ... Olde Fahrtes like me are aging out of the equations.

I think this also accounts for why there is a niche renaissance of film users amongst the younger folks happening: it's a- something new/different, b- has a different look, and c- has a different feeling of "involvement" compared to what they've now grown up on. Never mind that film camera technology and processing/rendering has been highly refined and developed for many years.... Film images have a very sweet and pleasing look, the end result of many decades of development work across the entire system from camera to lens to film to processing to photofinishing. Digital cameras and imaging are still only teenagers, at best, in the consumer space by comparison.

There's a long way to go. I've been at it for longer than a lot of folks since I got my start in digital imaging systems and rendering when I was working at NASA in the 1980s, but there are only a few who have my particular history with these things. For me, it has always been an expansion of opportunity and capability. But that expansion means I've had to be willing to work at learning how to get what I wanted, and it hasn't always been easy (or cheap).

Stay well, stay sane, keep on going!

G
 
charjohncarter said:
I wish I had your skill and expertise, I still think about your forum portfolio at Pentax. Stay safe too. Throw some more stay at home Polaroids at us.
Click to expand...

Coming up soon on the Polaroids!

I had shoulder surgery a month and a week ago: it's just within the past couple of days that my arm is now becoming just barely capable of holding an eye level camera up to my eye again. I'll soon be out and about with the SX-70 variants again. 😀

G
 
I've never seen a camera with so many lenses. Perhaps this is the future of photography?

Do all these lenses/sensors "add up" to a larger or full frame sensor equivalent, with the computational photography involved? It certainly is more convenient than carrying around a large full frame sensor camera which needs a large lens to cover the large sensor.

lightreviewfeat-800x420.jpg
 
Forest_rain said:
Do all these lenses/sensors "add up" to a larger or full frame sensor equivalent, with the computational photography involved? It certainly is more convenient than carrying around a large full frame sensor camera which needs a large lens to cover the large sensor.
Click to expand...

What is interesting is that there are also several different focal lengths between them all too. For the samples I've looked at it certainly gives more than just a single small sensor but doesn't really hit full frame performance either because the quality varies around the frame. Where there is a lot of overlap between sensors/lenses the quality is increased, where there is little overlap the quality noticeably drops. For example look at this image at 100%

https://1.img-dpreview.com/files/p/sample_galleries/3604419316/6089270467.jpg

The edges are all smeary (beyond just DOF) but the center rock varies a bit too. There are areas with lots of details and then areas where it drops off and then picks back up again.

Likewise the focal length you choose to shoot at also affects this as that alters overlap.

Shawn
 
The L16 has a 16 camera array, cut approximately into thirds with 28mm, 35mm, and 75mm equivalent focal length lenses. For every exposure, given a simulated focal length chosen by the photographer, ten of the cameras are chosen to make an exposure simultaneously. The ten cameras' exposures are then coalesced with processing to obtain a single image frame.

If you look at a map of the pixel resolution vs focal length, the total resolution is highest at the 28, 35, and 75mm focal length settings, with approximately 52 mpixel being created at those speeds. It falls between those settings due to interpolation and the number of pixels that fit the overall virtual frame. At 150mm, the total resolution is down to 12 mpixel due to the limited number of pixel sources.

Aside from the pixel resolution, the actual cameras/lenses being used are sharpest in the middle range of focal lengths, from 35 to 75mm, where you have the most possible overlap. Wide views down in the 28mm range (such as the one shown in that link) lose out at corners and edges because there are only so many 28mm lenses. Longer telephoto settings tend to be pretty darn sharp since it's coalescing the more center portions of the 35mm and 75mm eqFOV cameras, but total pixel count is diminishing above 75mm.

Another aspect to the game is the fact that all those cameras' lenses are generally capturing at near fully open (if not actually fully open) aperture and the effect of stopping down is simulated. When you adjust the DoF in the in-camera processing engine, or in Lumen on your desktop, it's doing the calculations of how to fit all the different exposures together and simulating where the out of focus occurs based on the modeling of all the lenses and cameras focusing behavior. It's not like a Lytro camera ... you don't have all the vector information captured ... so you MUST obtain good critical focus on your primary subject to obtain maximum performance, there's no shifting the focus after the fact, only reprocessing within the range of the lenses' modeled focusing range.

So: No, it's not a replacement for a FF or MFD 50mpixel sensor in terms of overall acutance and resolution capability. It's also not a challenger to those in terms of total dynamic range since all those little camera sensors are capturing at most 12 bits of tonal information, vs any good modern APS-C, FF, or MFD sensor at 14 to 16 bits.

What it is is a very clever way to build something incredibly small and light that can give overall performance that's "close enough" to a FF camera for a great deal of subjects in a reasonably broad range of lighting.. and provide a wide to decent tele FoV range in the process. Not bad for something that you can basically slip in a jacket pocket and carry everywhere.

What I find in using it is that, like using an iPhone for 'real' photography, if you lock it down on a tripod, focus carefully, and keep the ISO as low as possible, you can get some terrific photos out of it.

Here are two albums with a bunch more Light L16 photos I've posted. Some of them are full resolution duplicates (I often set those to Private and share the links to them separately, that's why) that can give you some idea of what you can get out of the camera in a variety of shooting circumstances.

https://www.flickr.com/gp/gdgphoto/P69515
https://www.flickr.com/gp/gdgphoto/6m9k59

It's a fun camera. I bought in on it because of the form factor and because of my fascination which this form of computational photographic workflow. I've had a really good time using it ... what you see above is only a small percentage of all the satisfying photos I"ve made with it. 😀

G
 
I must say that I'm quite impressed by the quality. You could tell me that you took those pictures with an ASP-C camera or even full frame for some of them and I'd believe you.

The effect looks pretty close to larger sensor, sharpness is there, but I guess you've figured out how to use it as well.

Some of those pictures look like you added a film effect, which is nice, and they are indistinguishable from film pictures to me.

Godfrey said:
The L16 has a 16 camera array, cut approximately into thirds with 28mm, 35mm, and 75mm equivalent focal length lenses. For every exposure, given a simulated focal length chosen by the photographer, ten of the cameras are chosen to make an exposure simultaneously. The ten cameras' exposures are then coalesced with processing to obtain a single image frame.

If you look at a map of the pixel resolution vs focal length, the total resolution is highest at the 28, 35, and 75mm focal length settings, with approximately 52 mpixel being created at those speeds. It falls between those settings due to interpolation and the number of pixels that fit the overall virtual frame. At 150mm, the total resolution is down to 12 mpixel due to the limited number of pixel sources.

Aside from the pixel resolution, the actual cameras/lenses being used are sharpest in the middle range of focal lengths, from 35 to 75mm, where you have the most possible overlap. Wide views down in the 28mm range (such as the one shown in that link) lose out at corners and edges because there are only so many 28mm lenses. Longer telephoto settings tend to be pretty darn sharp since it's coalescing the more center portions of the 35mm and 75mm eqFOV cameras, but total pixel count is diminishing above 75mm.

Another aspect to the game is the fact that all those cameras' lenses are generally capturing at near fully open (if not actually fully open) aperture and the effect of stopping down is simulated. When you adjust the DoF in the in-camera processing engine, or in Lumen on your desktop, it's doing the calculations of how to fit all the different exposures together and simulating where the out of focus occurs based on the modeling of all the lenses and cameras focusing behavior. It's not like a Lytro camera ... you don't have all the vector information captured ... so you MUST obtain good critical focus on your primary subject to obtain maximum performance, there's no shifting the focus after the fact, only reprocessing within the range of the lenses' modeled focusing range.

So: No, it's not a replacement for a FF or MFD 50mpixel sensor in terms of overall acutance and resolution capability. It's also not a challenger to those in terms of total dynamic range since all those little camera sensors are capturing at most 12 bits of tonal information, vs any good modern APS-C, FF, or MFD sensor at 14 to 16 bits.

What it is is a very clever way to build something incredibly small and light that can give overall performance that's "close enough" to a FF camera for a great deal of subjects in a reasonably broad range of lighting.. and provide a wide to decent tele FoV range in the process. Not bad for something that you can basically slip in a jacket pocket and carry everywhere.

What I find in using it is that, like using an iPhone for 'real' photography, if you lock it down on a tripod, focus carefully, and keep the ISO as low as possible, you can get some terrific photos out of it.

Here are two albums with a bunch more Light L16 photos I've posted. Some of them are full resolution duplicates (I often set those to Private and share the links to them separately, that's why) that can give you some idea of what you can get out of the camera in a variety of shooting circumstances.

https://www.flickr.com/gp/gdgphoto/P69515
https://www.flickr.com/gp/gdgphoto/6m9k59

It's a fun camera. I bought in on it because of the form factor and because of my fascination which this form of computational photographic workflow. I've had a really good time using it ... what you see above is only a small percentage of all the satisfying photos I"ve made with it. 😀

G
Click to expand...
 
Forest_rain said:
I must say that I'm quite impressed by the quality. You could tell me that you took those pictures with an ASP-C camera or even full frame for some of them and I'd believe you.

The effect looks pretty close to larger sensor, sharpness is there, but I guess you've figured out how to use it as well.

Some of those pictures look like you added a film effect, which is nice, and they are indistinguishable from film pictures to me.
Click to expand...

Thank you!

All the ones in the album from the Caribbean trip with the film effect added were actually previews .. I had the camera process them to a JPEG at about half rez, then transferred them to my iPad while still traveling, then finished up with them in SnapSeed and posted them to Flickr. The full resolution images processed at home in the Lumen app and then finished in Lightroom are even better quality, by a long shot.

G
 
Not included in the thread's title set, but two more photos from my walk that day...


Orange Leaves - Santa Clara 2020
Light L16
ISO 200 @ f/16 @ 1/60 @ 55mm


White Flowers - Santa Clara 2020
Light L16
ISO 100 @ f/16 @ 1/90 @ 35mm​

Enjoy! G
 
I've not been posting much, and been off the bicycle, for the past few weeks... various reasons, but I've continued my daily walks/photo sessions around the neighborhood as time and energy permit. And I've been processing/rendering a lot of exposures as well.

Even walks the past week or two have been limited due to heat and (now) wildfire smoke. Sheesh, what a year!

Here are six from Sunday's short walk around the neighborhood.


Blossoms


Leaf


Japanese Maple


Crashed Fruit


Window and Rose


Flag

All with Light L16.​

More to come soon... 🙂
Enjoy!

G
 
I was on my usual Saturday Morning Ride to breakfast and such yesterday. As I reached my turnaround point at the south end of the airport, I thought, "hmm, this would be a good place to get a couple of new shots of me and the bike..." And then I saw this curious graffiti. So I made a few exposures ...


Curious Graffiti
Light L16
ISO 100 @ f/15 @ 1/500 @ 35mm[/I]


GDG and SBG #1
Light L16
ISO 100 @ f/3.5 @ 1/2700 @ 28mm


GDG and SBG #2
Light L16
ISO 100 @ f/2 @ 1/700 @ 28mm​

Not the sharpest self-portraits since there was nothing for me to set the focus on ahead of working the shutter and letting the self timer run 10 seconds, but acceptable at this size rendering. 🙂

enjoy!
G
 
You must log in or register to reply here.
Back
Top Bottom