matthiasbasler

matthiasbasler

Lives in Germany Germany
Works as a GIS developer, hobby photographer
Has a website at baslerphotos.de
Joined on Nov 25, 2012

Comments

Total: 36, showing: 1 – 20
« First‹ Previous12Next ›Last »
On article Lytro Desktop 5.0 introduces depth-based image editing (41 comments in total)

Now the new version is out and I have done some tests here is what I found:
- On the plus side, being finally able to give proper names to images is nice.
- I don't care about the new DepthFX features, thus didn't test them.
- The promised performance improvements ... well, I didn't find any. At least performance didn't get worse either.
- None of the bugs I have posted to Lytro seems to have been fixed. Within two hours of working with the application I had my first main screen freeze. (It would continue to show one and the same image, no matter what thumbnail I chose.)

And by the way:
Although there has never been an official statement it seems obvious to me that 500px has de facto ceased support for displaying Lytro images, although the Lytro software still offers to upload images to them.
Even Lytro's own profile looks rather barren now: https://500px.com/lytro
If 500px really had an interest in it they had fixed this issue within the last two months, I am sure.

Direct link | Posted on Dec 15, 2015 at 18:11 UTC as 2nd comment
In reply to:

luisflorit: Why do I always see lots of artifacts in the blended (stacked) final results??
I've never seen a satisfactory HDR or panorama automatic implementation, always poorly blended. And the results here indicate to me that this focus stacking has the same problems (man face around the nose, and just everywhere in the church image). Manual blending or using dedicate software like Hugin always gives me MUCH better results, although with extra work of course.

I believe this kind of soft halos is not a problem of this (or any other) camera, but rather inherent to how image stacking software works. Think about it: If you have an image with sharp background your foreground will be blurry, which means that a twig (for instance) in the foreground will appear softer and thus wider than it is. It covers quite a bit of background. The software would need to replace this larger section but in the foreground image the twig is crisp and narrow. Thus, the halos stem from information not available in any of the original images.

Since the Lytro literally looks around objects it doesn't have this problem to this extent.

I wonder if you really can circumvent this issue by stacking manually... ?

Direct link | Posted on Nov 27, 2015 at 18:03 UTC
In reply to:

luisflorit: Why do I always see lots of artifacts in the blended (stacked) final results??
I've never seen a satisfactory HDR or panorama automatic implementation, always poorly blended. And the results here indicate to me that this focus stacking has the same problems (man face around the nose, and just everywhere in the church image). Manual blending or using dedicate software like Hugin always gives me MUCH better results, although with extra work of course.

Thanks, now I see what you mean. The issue is directly left of the man's nose and continues down to its upper lip. Clearly not perfect but nothing I would depend a pro or contra for this camera on.

Direct link | Posted on Nov 27, 2015 at 17:36 UTC
On article Lytro Desktop 5.0 introduces depth-based image editing (41 comments in total)
In reply to:

forpetessake: Even though the foreground is very close to the camera and background is virtually in infinity, the automatic separation is pretty poor and defects are visible even on that small picture. The PS mask would do a better job in not much more time. And, of course, in real work creating a mask is the simplest of the tasks. Getting the lighting and colors consistent is more difficult.

Looks like Lytro is still looking for an application for their product and have hard time finding any.

> I wonder how much of this is due to the inherent low resolution of spatial capture?

Not much I believe. As far as I understand the process the depth image is derived by comparing the images taken from different angles and identifying objects in them. This means that on parts of the image with little or no contrast the camera has no chance of separating foreground and background. My typical issue: a green flower stalk in front of a similarly green meadow. A radar sensor would see the stalk, but a camera cannot. No problem on a 2D image, but in a 3D scene the result is a flower hovering in mid-air. (Detection has somewhat improved though.)

Direct link | Posted on Nov 25, 2015 at 18:28 UTC
On article Lytro Desktop 5.0 introduces depth-based image editing (41 comments in total)
In reply to:

LF Photography: Here is what the internet taught me about Lytro over the years:

- They invented a revolutionary tech that allows choosing focus in PP
- The came out with an expensive gadget camera that didn't sell
- They came out with another, even more expensive camera, that was better, but still bad.
This one also didn't sell.
- They closed up shop and turned to VR
- They just came out with an software update for their camera system.

That is all.
I'm sure there's more to the story, but at this point, I just don't care anymore. No disrespect intended, as I'm sure Lytro is a great company with great people, but Light Field Tech, just as VR tech (ironically), has been around for a long time, and hasn't caught on for a reason; It's not good enough. At least not yet. And once it becomes good enough, it will have to earn the consumer's trust all over again, and prove itself as something more than just a hip gadget.

After having used the Lytro Illum for half a year I also wonder if the light field approach Lytro has taken is the right one to achieve the goals they had and I had when I bought the camera.
Getting images with ~1.5 MP-2MP resolution and artifacts out of a 40MP sensor makes me feel there must be some better way. Panasonics "Post focus" function sounds like a better trade off for most photographers:
- You get 8MP resolution (enough for me)
- There are no limitations to how to place your focus points. (Yes, you have to focus with the Lytro Illum, and imho it's harder than with a conventional camera)
- You can get conventional JPG images out of it and are not tied to a proprietary, somewhat buggy software.
- Your images are not 120MB each (!)
On the other side:
- the Lytro Illum can photograph moving objects such as animals, although it is imho useless for fast motion, e.g. sports.
- the Lytro Desktop software offers useful features such as videos, perspective shift, virtual aperture.

Direct link | Posted on Nov 25, 2015 at 18:15 UTC
In reply to:

luisflorit: Why do I always see lots of artifacts in the blended (stacked) final results??
I've never seen a satisfactory HDR or panorama automatic implementation, always poorly blended. And the results here indicate to me that this focus stacking has the same problems (man face around the nose, and just everywhere in the church image). Manual blending or using dedicate software like Hugin always gives me MUCH better results, although with extra work of course.

I looked hard but could not detect any artifacts or problems on the men's portrait, nor do I find any in the church images.

But I have worked with the Lytro Illum and compared to the soft images (with ~1.5MP effective horizontal resolution) and the artifacts often seen the images of that camera these images here look flawless.

Maybe you can explain more precisely what problems you have seen in the sample images. Also note that artifacts in stacked images are likely an issue of the stacking software, not the camera creating the original images.

(With respect to in-camera HDR: I haven't seen a useful implementation so far either. My Canon camera doesn't even allow to set brightness or white balance when in HDR mode!)

Direct link | Posted on Nov 25, 2015 at 17:50 UTC

The specs above read:
* Microphone ... None,
* Microphone port ... None.
I guess the former is a mistake (or not yet known), otherwise the camera could only take Full-HD-Videos without sound, and I believe no manufacturer wouldn't do something that stupid.

Direct link | Posted on Oct 13, 2015 at 07:59 UTC as 46th comment | 1 reply

A whole bunch of ideas sprout from this concept. How about:
* cars which don't let you drive anywhere near traffic jams or during rush hour
* ovens and microwaves which won't work with fast food
* music and DVD players that allow a certain disk to be listened to resp. watched at maximum once per week
* TV sets which block blockbusters and important sports live events
* web browsers which disallow access to high-traffic pages such as Google, Youtube, ebay, amazon or maybe even ... dpreview.com

The possibilities are endless. Insanity too.

Direct link | Posted on Sep 11, 2015 at 16:11 UTC as 84th comment | 1 reply
On article Holga Digital camera project launched on Kickstarter (150 comments in total)
In reply to:

Elandreth: "The Holga Digital camera features original optics that are known for their dark corners..."

Did they just try to advertise heavy vignetting as a positive feature?

Yup.
Why pay thousands of dollars/euros for Photoshop if you can have the same feature built in. Now that I think of it ... I believe my camera can do this as well. Its hidden in those filters I never use. And best of all: I can turn the filter off.
;-)

Direct link | Posted on Aug 31, 2015 at 20:24 UTC

I don't believe it is primarily the complexity of dedicated cameras that is responsible for the drop in sales. I see it as a combination of following factors:

1. Smartphone cameras meanwhile can create good-looking photos.
2. Having a smartphone saves you taking a dedicated point and shoot camera with you, and if you don't care about DOF, minimizing noise or the like you don't loose much. Oh, and you don't have to care about shutter shock at all.
3. Smartphones have cool tricks like filters, blurred backgrounds, sweep panoramas etc. built in, so no need to go through a complicated software workflow in post. -> Time saver.
4. People who still like a dedicated camera often already have a capable one, and few see the need to spend money every two years to get some new features.

Maybe dedicated cameras will become again what they had been before the digital era: A tool for real photographers, not a gadget everyone has. And there's imho nothing wrong with that.

Direct link | Posted on Aug 28, 2015 at 17:26 UTC as 175th comment | 5 replies
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)

Concerning the AF improvements:

After one day of using the new camera firmware (2.0) I find that the AF is faster, but still slow compared to conventional cameras using phase detection.

Without having done scientific A-B tests I believe the big advantage of the new firmware is that the camera does not need to scan the full focus range any more in order to find the perfect distance, but starts off in the right direction immediately. However it subjectively still needs between 0.5s (wideangle) and 2s (telephoto) to get to the right point. In comparison, on my EOS 650D with the 18-135mm IS STM lens focusing from close to far (or back) is done after ~0.3 seconds or so at 135mm.

Direct link | Posted on Jul 12, 2015 at 08:33 UTC as 7th comment | 1 reply
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)
In reply to:

DStudio: When Lytro released their first camera it was written off by many as a gimmick. But I think the company's shown they are serious.

Rosenthal clearly believes in this vision, and it's slowly coming into focus for consumers.

I own an Illum and, honestly, it too is a gimmick, albeit with the form factor, look and build quality of a professional camera.

For me as someone doing experimental photography this is fine, but I can hardly imagine anyone (except Lytro) earning money with a camera whose images have an effective spatial resolution of 2MP and, depending on the type of subjects, often show artifacts. If I were to engage a portrait or wedding photographer, for example, I'd prefer one who gives me a sharp, clean 8MP 2D image instead.

I am not saying that light field cameras cannot be used seriously, but imho we are not yet there. It requires an even higher resolution sensor (60MP instead of 40), thus even more camera computing power, even faster desktop PCs to process the image. It also requires technical and usability improvements so that getting the focus right is not harder and slower than on a conventional camera. (Currently it is rather a science ...)

I hope we will get there some day.

Direct link | Posted on Jul 12, 2015 at 08:20 UTC
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)
In reply to:

random78: "And in case you've forgotten some of our earlier coverage of the Illum, F1.0 gets you 30-250mm F2.7 full-frame equivalent lens performance"

But the previous coverage said that the lens is constant f2.0 which would be f5.4 FF equivalent in terms of DOF. The camera picture also says f2.0. Where does this f1.0 number come from?

Um, before you continue arguing about the exact crop factor in order to get an "equivalent DOF" value keep in mind that the Lytro Illum can get you almost every depth of field you want, so the crop factor is imho largely irrelevant in practice.

P.S. @Rishi Sanyal
The fact that you can now process your image to F1 even in the camera firmware already doesn't change the real (fixed) aperture of the lens, which to my knowledge (and backed by Wikipedia) is F2.

Direct link | Posted on Jul 12, 2015 at 07:52 UTC
In reply to:

kty: Now the big question for video folks is:

SHALL I WAIT FOR THE "A7s II" OR NOT ?!

@ Enginel
If I prefer cameras with lower pixel count then there are reasons beyond the (potentially) higher sensitivity. They include:
- smaller RAW files and faster processing of them at my PC
- camera needs less computing power, thus can potentially save some energy (and cooling is easier)
- at lower resolutions the camera can offer higher burst rates or can use its processing power for other useful things, like CA correction (which on my EOS 650D extremely reduces the burst rate).
- cameras with very high resolutions commonly do not offer small file sizes at all, and I don't want to shoot timelapse sequences in 21MP, just as an example.

Direct link | Posted on Jul 12, 2015 at 07:40 UTC
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)
In reply to:

random78: "And in case you've forgotten some of our earlier coverage of the Illum, F1.0 gets you 30-250mm F2.7 full-frame equivalent lens performance"

But the previous coverage said that the lens is constant f2.0 which would be f5.4 FF equivalent in terms of DOF. The camera picture also says f2.0. Where does this f1.0 number come from?

> Where does this f1.0 number come from?

I believe it is an error. As far as I know the camera has an aperture of F2.0 in respect to exposure times and at the same time covers a depth range corresponding to F16.

The F1 maybe made its way into the article from the fact that the Lytro Desktop software allows to set the "virtual aperture" in post processing to anything from F1 to F16, meaning that you can actually compute the out-of-focus areas to be even more unsharp than the default value.

Concerning the sensor size: The specs at dpreview.com state 1/2", which seems to be wrong as well because the official website (https://www.lytro.com/illum/specs/) claims 1/1.2" (10.82 x 7.52 mm active area). A typo maybe.

Direct link | Posted on Jul 11, 2015 at 18:45 UTC
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)

My favourite software improvements in Lytro Desktop 4.3:
- The depth map editor: No more exporting/importing with 3rd party software (e.g. Gimp) in order to correct for depth map artifacts. This is a big time saver.
- Ability to just import the images from the camera without processing them immediately.

@Everlast66
Yes, this camera is expensive for a high-tech toy (which is what I consider it). But hey, how many other camera manufacturers give you so significant software/firmware improvements for free? (My old Nikon D5000 did not get one single firmware update since I bought it ... and it wasn't free of quirks either.)

Direct link | Posted on Jul 10, 2015 at 23:01 UTC as 9th comment
On article Lytro ILLUM and Desktop software get major updates (36 comments in total)
In reply to:

creaDVty: It's great that they keep updating the firmware and desktop software to add more features. However, I don't understand the addition of phase detection AF. I thought the whole point of Lytro is that you don't have to focus in the first place because the focus could be chosen afterward. So I don't understand what difference it makes to add phase detection.

It is a common misconception that because it is a light field camera that can shoot 3D scenes then it needs no focusing. This camera neither has unlimited focus nor does it have a distinct range with everything within being equally sharp.

In reality this camera has two distinct sharpness maxima (at "-4" and "+4" on the relative range scale of the camera) with a slight dip inbetween and sharpness continually decreasing outwards. (Just imagine two overlapping gaussian curves.)

This is why this camera gives the sharpest images if you have a foreground subject and a background subject and place the focus range in a way that the foreground subject is at -4 and the background one at +4. Sounds complicated? It is, even though you can tell the camera to place the autofocused subject at -4 (or +4 at your liking).

Direct link | Posted on Jul 10, 2015 at 22:53 UTC

First, if a law is not enforced in practice, then this law is useless and thus obsolete (or should not be ratified in the first place).

Second, it is my personal opinion that if an artist or architect designs a sculpture or building to be placed in public space and gets paid for this then this payment should include the right (licence) for the public to take photos of the subject for *any* possible usage, including commercial.

If this is not possible then I think that the city councils should enforce a policy of only buying art for public spaces which includes the licence for public photography of any type.

Matthias

Direct link | Posted on Jul 4, 2015 at 18:08 UTC as 151st comment | 1 reply
In reply to:

matthiasbasler: A question to the community (or dpreviewers), since the topic of battery life is often discussed here:
One of my use cases is timelapse photography. Many cameras do not allow to change battery while on a tripod so battery life *is* important. Leaving accessory grips out of discussion for now, is there anybody with practical experience just how much the battery life of a 300-CIPA-shots rated mirrorless camera compares to a 450 or 600-CIPY-shots rated DSLR in this specific use case where the camera is on/standby for hours continuously and shooting, lets say, one photo every 15 or 20 seconds?

I have an EOS650D, rated 440 shots, and during a night timelapse session at ~20°C temperature it worked 5.5hrs with one full (original Canon) battery. It did roughly 1000 shots every 20s with 4s exposure each.
Do you have any comparable data for a mirrorless ILC cameras?

Thank you all for your answers and suggestions.

@bwana4swahili
> The new A7R II allows one to power the camera via its USB port while shooting.
This sounds very good. My EOS 650D unfortunately doesn't take power from the PC although it is constantly connected to one. :-(

I am not aware that it is really possible to "power off" the display. When connected to the PC the EOS650D doesn't show anything on the display during my timelapse recordings, but it might still use battery on it. I don't know how to tell.

Direct link | Posted on Jun 21, 2015 at 08:29 UTC

A question to the community (or dpreviewers), since the topic of battery life is often discussed here:
One of my use cases is timelapse photography. Many cameras do not allow to change battery while on a tripod so battery life *is* important. Leaving accessory grips out of discussion for now, is there anybody with practical experience just how much the battery life of a 300-CIPA-shots rated mirrorless camera compares to a 450 or 600-CIPY-shots rated DSLR in this specific use case where the camera is on/standby for hours continuously and shooting, lets say, one photo every 15 or 20 seconds?

I have an EOS650D, rated 440 shots, and during a night timelapse session at ~20°C temperature it worked 5.5hrs with one full (original Canon) battery. It did roughly 1000 shots every 20s with 4s exposure each.
Do you have any comparable data for a mirrorless ILC cameras?

Direct link | Posted on Jun 20, 2015 at 10:46 UTC as 69th comment | 10 replies
Total: 36, showing: 1 – 20
« First‹ Previous12Next ›Last »