Henrik Herranen

Lives in Finland Tampere, Finland
Works as a Digital Signal Processing Software Engineer, MSc
Joined on Oct 6, 2005
About me:

Plan: To baldly shoot what everyone has shot before.

Comments

Total: 218, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

qwertyasdf: 2x TC on 100-400mm f4.5-5.6 gives 800mm @ f11.2, if this qualifies as bringing 1219 to a system, I might as well stack five 2x TCs on a Canon 1200mm and and bring 8,246,337,208,320mm to the EF system, on a crop body.

qwerty: for that you'd need thirty-two stacked 2x TCs (1200*2^32*1.6 ~= 8.24*10^12).

With five stacked 2x TCs you'd end up with 1200mm*2^5 = 38400mm f/180, which would be equivalent to 61440mm when used on a 1.6x crop body.

Warning: if really pixel-peeping, you might notice some slight diffraction softening.

Link | Posted on May 19, 2016 at 07:13 UTC
In reply to:

The Squire: Only 1080p? At least it's "spectacular 1080p"!

BartET:
I'll take your bet. As we live in different countries (I assume), let's choose a common subject: The Moon. The one who gets the better picture without any extra optics, wins.

So, how much was your monthly salary and how are you going to pay? Or are you just trolling?

Link | Posted on May 11, 2016 at 14:23 UTC
In reply to:

tkbslc: f6.6? I guess it's good for outdoors in the afternoon.

phototransformations: you are of course absolutely right. But take into account that what you are answering to must be a feeble attempt at trolling. Nobody could be so clueless.

Link | Posted on May 11, 2016 at 14:14 UTC
In reply to:

tkbslc: f6.6? I guess it's good for outdoors in the afternoon.

f/6.6 on a crop factor 5.6 system will have similar diffraction as a Full Frame lens stopped down to f/37, so regardless of whether the optics are good or not, image quality will by necessity be limited. Then again, what choice is there if you need a superzoom in such a small form factor?

Link | Posted on May 11, 2016 at 05:50 UTC

DPR writes:
"It also brings a modest increase in resolution because you're sampling luminance (green) information at every pixel position and not effectively blurring it by borrowing it from surrounding pixels."

Luminance is _not_ green. Green is green. Luminance is a weighted sum of red, green, and blue. Remove the incorrect and confusing "(green)" qualifier and you're golden.

Btw, if you want to compare high-resolution images, why is the Nikon D810 in the default resolution comparison instead of the higher resolution Canon 5DS R?

Link | Posted on May 5, 2016 at 22:04 UTC as 124th comment | 1 reply
In reply to:

Henrik Herranen: Assessing skintones by taking a photo of a photo is a bit like a broken pencil - pointless.

The spectral features of a 4-colour print (under artificial light) are vastly different from the spectral densities of real skin. Hence, the results may be quite different.

Real-world example:
- If I take a photo of hot coals with my Canon S90 with white balance set to Incandescent, IR leakage will make the coals appear purple, even close to blue.
- If I take the same photo with my Canon 5D Mark II, the coals are red as they should because this camera has a proper IR filter.
- If I print the 5D2 photo on paper, take a photo of this photo with my S90, then upload it to the web for all to see, it will give a completely incorrect impression of the S90's colour capabilities.

Fair enough, Rishi.

By the way, speaking of colour of light...

Wouldn't it be _really_ interesting if there was a way to take a photo of a continuous spectrum, starting from deep UV through visible light right into deep IR territory? That would show some significant colour production differences, particularly with small sensor cameras, including camera phones. My experience is that some of them have proper IR filters, but most of them don't, which affects e.g. skin tones under incandescent light.

So, how about getting a continuous-spectrum light source plus an appropriately big prism, then adding the output of such a system to the standard test image suite? That way you'd get the separate frequency responses of the R, B and B filter elements: how much do they bleed to each other, and what colours are difficult to represent accurately. Might be pretty interesting...

Link | Posted on May 4, 2016 at 21:37 UTC

Assessing skintones by taking a photo of a photo is a bit like a broken pencil - pointless.

The spectral features of a 4-colour print (under artificial light) are vastly different from the spectral densities of real skin. Hence, the results may be quite different.

Real-world example:
- If I take a photo of hot coals with my Canon S90 with white balance set to Incandescent, IR leakage will make the coals appear purple, even close to blue.
- If I take the same photo with my Canon 5D Mark II, the coals are red as they should because this camera has a proper IR filter.
- If I print the 5D2 photo on paper, take a photo of this photo with my S90, then upload it to the web for all to see, it will give a completely incorrect impression of the S90's colour capabilities.

Link | Posted on May 4, 2016 at 16:45 UTC as 105th comment | 4 replies
On article The Canon that can: Canon EOS 80D Review (674 comments in total)
In reply to:

Zerixos: Its nice to see that they stept-up there game. I think it once-again adds something to there line up since the 50D (the 60D and 70D where more like a 50Ds and 50DII in my opinion) and make it a great allround camera. I'm just a little sad not to support 4K video. Anyone saying you don't need 4K is right, for now. But knowing you want it in the (near) future is gonna make this a quick aging camera.

1080p that keeps consistently in focus without hunting or breathing looks better than 4k that is out of focus or constantly jumping and pumping.

So yeah, it's not 4k, but it's consistently good 1080p.

Link | Posted on Apr 28, 2016 at 07:35 UTC
In reply to:

emfor: So consequently the 80D review on DPR will have to compare its DR with a seven years old Sony or Nikon :))

Richard,

joking aside, although I understand and appreciate your point, I think the point of the complainers is that you compared an excellent DSLR from 2009 with the best mirrorless has to offer today, in 2016 (Nikon 1 doesn't count because sensor size). Yes, the frame rates match, but to many the comparison feels unfair. It can feel a little like saying the new iPhone has developed greatly by comparing it to a 7-year old Nokia Symbian phone just because their cameras happened to have the same number of megapixels.

Yes, mirrorless has progressed. But not comparing state-of-the-art mirrorless to state-of-the-art DSLR makes it feel like you _want_ to give mirrorless an advantage - even though I am quite certain that wasn't your intention.

Link | Posted on Apr 13, 2016 at 19:24 UTC
In reply to:

Clint Dunn: ....and this image is a perfect example of why I sold all my DSLR gear and switched to mirrorless. Ridiculously big setup.

EthanP99:
The Voigtländer is also slow (f/5.6) and not a zoom. At least in the "DLSR fool" world it's a bit of a non-news that you can make a slow prime smaller than a faster zoom. In the troll world these things can of course be conveniently forgotten.

Link | Posted on Mar 31, 2016 at 06:53 UTC
In reply to:

billslatteryjr: What a useful tool. Thank you dpreview for being on top of this.

bmw:
That wouldn't be fair to other manufacturers. As Roger Cicala has pointed out in his blog articles, adapted lenses never work as well as native ones. The adapters are never perfect (two lens mounts have more tolerances than one), and the thickness of the glass in front of the sensor (which is different with different brands) have an effect on how lenses should be designed for optimum performance.

All in all, using a XYZ lens adapted to all other brand cameras would favour XYZ cameras.

Link | Posted on Mar 28, 2016 at 16:29 UTC
In reply to:

mais51: Most focus problems are lens + camera body related that is why camera has fine tune feature built-in so how adjusting a lens independent of the specific body would fix that problem at all. The USB solution could only fix whichever problem exits with the lens. Could someone more knowledgeable in this field please explain

mais51
You have not really understood how this works, have you?
:
The lens is NOT adjusted indepent of the body. You tune it for YOUR camera body, using the docking station, and after tuning it works as well as it can with YOUR camera body. True, it cannot be optimized to work optimally with multiple camera bodies, but for most users this is not a problem in the slightest.

Link | Posted on Mar 16, 2016 at 08:43 UTC
In reply to:

mick232: the old firmware must have been pretty bad then and did by far not exploit the hardware capabilities

So, in your opinion they should just have left it as it is and not continued developing the firmware?

Sheesh, you guys are unbelievable. Sigma offers new firmware that may have required thousands of man-hours to develop, tune and test, to get even more out of their very good zoom, and all you can do is bitch and moan.

You must be the guy I read about at notalwaysright.com. In the real world, when we get a free update to an already good product, we get happy, NOT angry. But you, you somehow feel entitled to pour crap over everyone and everything, just for the sake of it.

For your information: developing software that takes the last 10% out of a given piece of hardware is hard, really hard. It's not like the 10% worse-performing version was bad, but you need to be _really_ good to take the final step. That's always true, but particularly with low-level programming that is required for such an embedded and complex electro-mechanical product as a lens is.

Link | Posted on Mar 16, 2016 at 08:39 UTC
In reply to:

Doctor Manhattan: >allowing users to optimize [...] vibration compensation

So I have to become an expert on vibration compensation too, now, and buy and learn how to use a dock? What is wrong with a switch on the side to change VR modes whenever I want?

Call me old fashioned, but I preferred it when the manufacturer didn't make me pay extra to do their job for them.

So you think it was better before when, if your camera-lens combination would require calibration, you sent them to factory calibration?

You know, nothing has been taken off. If you don't want to calibrate and adjust, then don't. Situation is as it always was. But, for those who are willing to take that extra step, this is nice to have.

Bu I guess complaining just for the sake of it is more fun than thinking.

Link | Posted on Feb 22, 2016 at 14:03 UTC
In reply to:

halfwaythere: Extremely limited range of compatible lenses. I wish they'd add some zooms since primes are much easier to adjust in-camera.

Of course the lens selection is small in the beginning. Expect that to change over the coming years.

Link | Posted on Feb 22, 2016 at 14:00 UTC
On article Newly enthused: hands on with the Canon EOS 80D (708 comments in total)
In reply to:

jpeghorror: Still no dedicated AF-assist lamp? Tell me I'm hallucinating.

Rishi:
A red beam would be ok, but an infrared beam would not help at all. At least not unless the camera was aware of the infrared focusing properties of every single lens that might be used with it and perform automatic compensation.

Link | Posted on Feb 18, 2016 at 10:32 UTC
In reply to:

Henrik Herranen: This is easy:
If you thought a Bayer matrix did bad things to your quantum efficiency by throwing out over 50% of the light, how about leaving 95-99.9% of the light in your pinhole mask? While this system may have its applications, it surely isn't going to replace even the most humble camera phone module. This is a hard constraint no amount of signal processing or other cleverness can overcome.

Another thing is that as much as I appreciate the videos, and even though their nominal resolution were 512x512 pixels, they contain 100 pixels of real resolution at most in either direction, and even that is marred with noise.

All in all, very interesting, but at the moment this looks like a solution looking for its problem.

hjulenissen: yes, they claim they have more holes than some previous pinhole cameras. However, they make no claims against lens-based systems in that sentence.

In other words: a hundred small holes would be two orders of magnitude better than one small hole, but still several orders of magnitude worse than a bigger-holed lens-based system.

Actually, on the top of the same page 2 of the document you quoted you can see an image that shows how little information the sensor receives. No wonder the reconstruction in the last column is lacking.

Let me stress that I don't claim this system is useless. It may well have some industrial application in the future, and I have the deepest repect for people trying to invent new things. But one thing is sure: this thing isn't going to put the photography world on fire.

Link | Posted on Feb 17, 2016 at 14:56 UTC

This is easy:
If you thought a Bayer matrix did bad things to your quantum efficiency by throwing out over 50% of the light, how about leaving 95-99.9% of the light in your pinhole mask? While this system may have its applications, it surely isn't going to replace even the most humble camera phone module. This is a hard constraint no amount of signal processing or other cleverness can overcome.

Another thing is that as much as I appreciate the videos, and even though their nominal resolution were 512x512 pixels, they contain 100 pixels of real resolution at most in either direction, and even that is marred with noise.

All in all, very interesting, but at the moment this looks like a solution looking for its problem.

Link | Posted on Feb 17, 2016 at 07:45 UTC as 18th comment | 3 replies
On article Key features explained: Canon EOS-1D X Mark II (97 comments in total)
In reply to:

RPJG: Does anyone know why the AF module only covers "only" a relatively small part of the frame? Would it be prohibitively expensive to implement a larger module? Presumably bird/sports/etc photographers would appreciate the ability to track objects outside that central part of the frame.

On a related note, if every pixel is split into two separate photodiodes, does anyone know why "only" 80% of the frame is available for focus using DP AF?

Rishi, yes, I think you made perfect sense.

Link | Posted on Feb 2, 2016 at 23:19 UTC
On article X-Factor: Canon's EOS-1D X Mark II examined in-depth (623 comments in total)
In reply to:

Joe Ogiba: 4K video with an OVF ? I would much prefer a high resolution 2.3-4.4 mp EVF. but this looks like the best DSLR for video so far.

KAllen: how you would shoot 4K video with a "real image in the viewfinder" using this camera is beyond me. Can't be done.

kangoo1707: The OVF can give you black in video mode, and not much else.

Link | Posted on Feb 2, 2016 at 14:49 UTC
Total: 218, showing: 1 – 20
« First‹ Previous12345Next ›Last »