BobORama

BobORama

Lives in United States Allentown USA, Earth, United States
Works as a kernel application developer
Has a website at blog.trafficshaper.com
Joined on Jun 20, 2006
About me:

Find someone with a better plan, hit them over the head, steal their plan.

Comments

Total: 52, showing: 1 – 20
« First‹ Previous123Next ›Last »
In reply to:

BobORama: The "examples" very much show read noise and quantization errors. Also 100% crops are better than fake hand drawn diagrams. They conceal the underlying fallacy that is the motivation of this article. These sorts of errors do occur, and can be a significant source of IQ issues. But its noot proven by any of the examples.

What DOES prove something is statistical analysis of a studio shot of a uniform shadow region. Natural subjects are notoriously not uniform. You take that uniform shaded region and look at the statistical distribution of the RAW values. Some of the effect will be the Poisson noise, which si the type of noise being mentioned in the article, and some will be read noise, some will be quantization errors.

So the article is a lot of pretty words and handmade diagrams - with none of the mathematical analysis that proves the noise examples provided are actually the type of noise being talked about.

-- Bob

All you need is the curve. Shot noise has a vary particular profile. Also shot noise does not vary with the temperature of the camera, as does read noise. I made ( https://vimeo.com/817772 ) this years ago for a CCD. Shot noise is the least of your worries. The graph shown is the increase in dark current and readout noise over time as the camera heats up.

Direct link | Posted on May 4, 2015 at 17:31 UTC

If you look CAREFULLY at the R, G, and B ray path images provided, you see the problem immediately. While the lens is achromatic for in focus subjects, the OOF image areas are not. And they cannot be. "Show me the bokeh." I expect it to be an incredible melange of badness.

-- Bob

Direct link | Posted on Mar 13, 2015 at 00:23 UTC as 13th comment | 4 replies
On Opinion: The myth of the upgrade path article (1454 comments in total)

As a Pentax user, we have been ready for 50 years, given that the announced FF Pentax offering will use any lens that worked on any Pentax body going back 50 years.

The in-body share reduction technology means every lens I already have is image stabilized, and will provide a 2-3 stop advantage, and handles rotation around the optical axis, like shutter press blur - which OIS cannot.

There are already many decent older ( but not necessarily great ) full frame lenses for me to use. I have a couple of them. And some of the cheap lenses, by Samyang for example, which have very high image quality but lack OIS ( for your platform ) are magically image stabilized on the Pentax.

So yeah, I guess some platforms do not have an upgrade path. Weird, huh?

-- Bob

Direct link | Posted on Feb 22, 2015 at 19:47 UTC as 35th comment
On Under Water Selfie in the Underwater Photography challenge (1 comment in total)

Mission accomplished!

Direct link | Posted on Feb 14, 2015 at 03:43 UTC as 1st comment
On CP+ 2015: Canon shows 11-24mm in cross-section article (63 comments in total)

Who has the incredibly fun job of cutting lenses in half? Because that looks like a lot of fun.

Direct link | Posted on Feb 13, 2015 at 22:36 UTC as 11th comment
On Nikon D750 Review preview (1906 comments in total)

To summarize: "That can't be a problem, its a Nikon."

Direct link | Posted on Jan 23, 2015 at 01:26 UTC as 118th comment

"Raping the Internet has never been so easy!"

How is this any different from downloading the CC images and printing them? Not very different at all.

As for these images competing with "fine art" photographers - how? Not everything that is CC is something you'd want a poster print of. e.g.

http://upload.wikimedia.org/wikipedia/commons/4/45/Dog_feces.jpg

And to be honest, the difference between "artist" and the rest of us is time and money.

Direct link | Posted on Nov 23, 2014 at 03:06 UTC as 10th comment | 5 replies
On Light Field Cameras - Focusing on the Future article (136 comments in total)

The silly physical design of the camera betrays the struggle to differentiate it from "normal" cameras.

Consumers do not care HOW the camera arrives at "focus as a afterthought." The way the Lytro does it is unimportant to consumers. Most people use phone cameras, so they will not be a Lytro customers.

The market for such a camera is therefor "true believers" - which does not bode well.

Direct link | Posted on May 18, 2014 at 14:10 UTC as 20th comment | 3 replies
On Did Amazon just patent the seamless background setup? article (133 comments in total)

Congratulations, Amazon, on re-inventing the usage of the cyclorama.

Direct link | Posted on May 8, 2014 at 01:17 UTC as 24th comment
On Is it true? New service detects processed photos article (88 comments in total)
In reply to:

Mark Banas: It seems to accept, and trust, falsified EXIF info (via EXIFtool) with no problem, so my Pentax K-5 must actually be a 16MP Nikon D800? Not terribly sophisticated service...

If I stand where Ansel Adams stood, take the exact same shot, and do the same PP in the dark room, can I sell it as an Ansel Adams? Because that's the argument... Provenance matters, so by extension the EXIF matters.

Direct link | Posted on May 7, 2014 at 21:11 UTC
On Is it true? New service detects processed photos article (88 comments in total)
In reply to:

Marky boy: who cares.... really....we like what we like, regardless of how, what and where. We can tie ourselves up in knots about post processing, however, it is no different to darkroom skills of old......

The value to this goes beyond the trivial use of determining if the supermodels eyebrows were digitally waxed. It could easily establish forgeries, and settle ownership issues. Furthermore, a system could be constructed to create a certificate chain like verification mechanism that would establish derived works. So photo B is an edited version of photo A, but B's crypto gobbledygook is chained to A's.

PKI + EXIF

-- Bob

Direct link | Posted on May 7, 2014 at 20:57 UTC
On Is it true? New service detects processed photos article (88 comments in total)
In reply to:

Marky boy: who cares.... really....we like what we like, regardless of how, what and where. We can tie ourselves up in knots about post processing, however, it is no different to darkroom skills of old......

Yes and no. Film and the camera it was shot in could be examined by an expert and one could conclusively determine if it was faked. But the negative would provide a level of repudiation against a modified print.

The digital negative has no non-repudiation mechanism. So its inherently NOT the same. Security logs on truely secure systems will have a non-repudiation mechanism, usually a cryptographic checksum.

In the context of a camera, the camera would use a private key to sign / encrypt the cryptographic checksum of the image data. ( The signatory agent could also be a wifi attached service, so some other mechanism. ) Then should someone need to validate the digital negative they need only decrypt the checksum value with the cameras public key. The checksum can be verified against the image data, and you have verified that this particular camera took that particular negative.

Direct link | Posted on May 7, 2014 at 20:47 UTC
On Is it true? New service detects processed photos article (88 comments in total)

I'd love to have access to the original test images. Its super easy to tell if an image has been edited - even expertly - as each layer of edits has its own signature.

Perhaps DPR should have a "Is It Fake?" challenge with 10 pictures, and see if we, as human beings, are better than the soul-less cloudy-cloud. We fill out a true / false form for each and whoever gets it right is put in a pool to win something.

EXIF data does not contain a non-repudiation mechanism and can be forged easily. Its a bad litmus test for both real and fake photos.

-- Bob

Direct link | Posted on May 7, 2014 at 19:36 UTC as 13th comment
On Camera from NASA's moon missions sold at auction article (104 comments in total)

It wouldn't be the first time someone ( vastly ) overpaid for a Hasselblad....

Direct link | Posted on Mar 26, 2014 at 23:38 UTC as 5th comment | 2 replies
On CP+ 2014: Things we found that had been cut in half article (107 comments in total)

I want to know HOW they cut them in half. I've done this on a small scale for metallurgical specimen preparation. Same thing just bigger?

Direct link | Posted on Feb 16, 2014 at 13:44 UTC as 33rd comment | 7 replies

I, for one, cannot wait till the Hasselblad Maserati's come out: Carbon fiber composite draped over a Kia Soul.

Direct link | Posted on Feb 4, 2014 at 03:17 UTC as 176th comment | 2 replies

Luminance noise is the sort of noise we experience in nature ( haze, fog, smoke, dust ) and in film grain, and in our own eyes when its dark. Chroma noise is abrasive, unnatural looking, and distracting. There is a difference between us being physiologically more sensitive in our ability to detect something and it being horribly distracting.

Would you rather listen to a loud pounding rain -OR- a baby's desperate, spasmodic crying at half the volume?

Direct link | Posted on Jan 26, 2014 at 03:36 UTC as 10th comment | 1 reply
In reply to:

jonny1976: The beauty of medium format is ccd. Thats what create the magic. Whats the use of a camera that creates the same d800 IMAGES with more pixel? Which professional will use a blad at iso 1600or 3200? The blad, not like pentax645d, is not Made for street.

I think your prejudices for or against CCD vs CMOS are based on old info, or no info. Both can very competently capture an image. The highest DR sensors currently in production are all CMOS sensors. Because of newer fab methods CMOS is less expensive to produce in larger die sizes than was one possible. And the fab techniques scale up and down in size. In a sense your were using CCD because that's what they could make, not because it was inherently better. Even your $40K hassy rig is "value managed."

Direct link | Posted on Jan 22, 2014 at 02:00 UTC
On In photos: The beauty of bees article (49 comments in total)
In reply to:

Biowizard: Sorry, but as a biologist who loves and does everything he can to support the dwindling natural environment in which we all have a deep and profound share, I find an album of photos of collected-and-killed bees about as appetising as those dreadful cases of pinned, dead (some now extinct) butterflies, moths and beetles collected by Victorian "explorers".

Not impressed.

Brian

I'm right there with you, BioWizard! Appreciating these photos is much like "appreciating" a lush studio shot of Marilyn Monroe's corpse.

When I saw the DPR teaser shot, I said "Wow! How did they get shots of bees in flight like that?" I was prepared to learn a cool new technique or something. Then the anti-climactic answer: Dead bee + $10K of publicly funded camera gear and time.

The researcher is free to follow their bliss. I'm not critiquing the work. But rather DPR's highlighting these mortuary shots. How about an article on someone dealign with live insects, using equipment most of us have?

I do a lot of ( non publicly funded ) macro work, including super-high resolution / zoomable images of ( non-deceased ) bugs, e.g. ( http://maps.muhlenberg.edu/iip2b/moth.html )

-- Bob

Direct link | Posted on Dec 29, 2013 at 18:27 UTC
On Backstory of Phantom Flex4K video shot at 1,000 FPS article (91 comments in total)
In reply to:

BobORama: So its a $140K camera system that makes even the most exciting jobs look boring? OK, the slo-mo flames are nice, beautiful, and the water emerging from the nozzles was interesting.

I guess the REAL advantage is that you can shoot a scene at 1000FPS and then in post production get whatever frame rate you desire to achieve. Smooth adjustments in the time scale give you very cool effects. And if you were shooting Inception 2 - it might be useful.

The scene does underscore how much water is entirely wasted. 80% of it forms huge globules that never actually participate in putting the fire out. One of the profs I had years ago was involved in alternative systems and things like a pressure washer, or even steam put out a fire more effectively, and with so little water its truly amazing.

If you can put out a fire with 80% less water, you downsize every aspect of operation, from the tankers, trucks, ... the need for hydrants. Also less water means less water damage.

-- Bob

thx2238,

By your standards I've never had a single glass of "American Beer" But there are a lot of fantastic beers made in the USA. And if not beer, Laprhoaig?

And as to By Gediminas 8 concern... I would only take issue if you drink your beer through a straw. And after a few, I'd probably overlook that as well.

Direct link | Posted on Dec 18, 2013 at 04:49 UTC
Total: 52, showing: 1 – 20
« First‹ Previous123Next ›Last »