Henrik Herranen

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

Henrik Herranen: Dear Damien Demolder,
is there a good reason why you don't revise this article and remove the following incorrect statement: "for the specification it is a good deal more compact than a similar lens for a full-frame or even APS-C system".

It is not. In the Full Frame world there are lenses like the EF24mm f/1.4 which both perform significantly better, and have much shallower DoF + much higher total light transmission than this lens.

(Before the I-don't-understand-equivalence-so-you-must-be-an-idiot-brigade trolls in, let me just remind you that because FF has 4 times the sensor area of m43, then the noise over image area, given similar sensor technologies, are similar when FF uses ISO 400 and m43 ISO 100. ISO is just a number without any real, physical dependence. When equivalent aperture (e.g. f/1.9 vs f/0.95) and ISO (e.g. ISO 1600 vs ISO 400) is used, then noise over image area, DoF, exposure time (or in a word: everything) is equivalent.)

PS.
I wouldn't be writing this if the author of the article wouldn't have made an incorrect comparison between m43 and FF lenses. If you want to avoid this kind of discussion, write of m43 lenses on their own merit, and don't INCORRECTLY compare them to lenses of different sensor systems.

Actually, don't even CORRECTLY compare them, because even that will cause lots of miscorrections and unnecessary arguments.

Direct link | Posted on Jul 14, 2015 at 19:03 UTC

Dear Damien Demolder,
is there a good reason why you don't revise this article and remove the following incorrect statement: "for the specification it is a good deal more compact than a similar lens for a full-frame or even APS-C system".

It is not. In the Full Frame world there are lenses like the EF24mm f/1.4 which both perform significantly better, and have much shallower DoF + much higher total light transmission than this lens.

(Before the I-don't-understand-equivalence-so-you-must-be-an-idiot-brigade trolls in, let me just remind you that because FF has 4 times the sensor area of m43, then the noise over image area, given similar sensor technologies, are similar when FF uses ISO 400 and m43 ISO 100. ISO is just a number without any real, physical dependence. When equivalent aperture (e.g. f/1.9 vs f/0.95) and ISO (e.g. ISO 1600 vs ISO 400) is used, then noise over image area, DoF, exposure time (or in a word: everything) is equivalent.)

Direct link | Posted on Jul 14, 2015 at 19:00 UTC as 15th comment | 11 replies

"The lens is neither especially small nor light, though for the specification it is a good deal more compact than a similar lens for a full-frame or even APS-C system."

Hmmh... 10.5mm f/0.95 lets in as much total light and has the same DoF as a FF 21mm f/1.9 lens. Canon's EF 24/1.4L II is both lighter and shorter, and lets in more than double the amount of total light. So what are the "similar lenses" that is won by this lens being "a good deal more compact"?

Don't get me wrong. I'm not saying the Voigtlander isn't a useful lens. But I do believe that this article begins with an incorrect statement.

Direct link | Posted on Jul 14, 2015 at 12:29 UTC as 31st comment | 32 replies
On Lytro ILLUM and Desktop software get major updates article (33 comments in total)
In reply to:

Everlast66: The price checking thingy top right reads:
"Buy on Amazon.com From $1,111.11"

Ha ha, yeah right, as if someone would pay $1.1k for this thing ;p

random78: It's not. The numbers come from thin air, or rather software blurring.

Direct link | Posted on Jul 11, 2015 at 08:34 UTC
On Phase One 645DF+ with IQ250 field test article (139 comments in total)
In reply to:

Henrik Herranen: Once again:
Why is sensor size not in the "Key specifications" list? You cannot expect your readers to remember these things by heart for every camera - I certainly don't.
Plus, honestly, if sensor size isn't a key specification, then what is?

Otherwise, an interesting read. Still, I don't think the "shallow DoF" argument works for MF anymore, with the sensors only slightly larger than FF, and with so many FF ultra-large aperture lenses (f/1.2 to f/1.4) available.

Yes, slightly larger.

At least the lenses used for this test don't have particularly impressive shallow DoF properties (crop factor of 0.78 used for calculations):

- 45mm f/2.8, equivalent to 35mm f/2.2 FF lens
- 80mm f/2.8, equivalent to 62mm f/2.2 FF lens
- 120mm f/4 macro, equivalent to 94mm f/3.1 FF macro lens

With us having 35mm f/1.4, 50 and 85 mm f/1.2, and 100mm f/2.8 lenses, the DoF advantage seems to quite definitively be in the FF camp. There certainly are advantages to MF (nicely presented in the article), but I don't think DoF is one of them.

Direct link | Posted on Jul 11, 2015 at 08:29 UTC
On Phase One 645DF+ with IQ250 field test article (139 comments in total)

Once again:
Why is sensor size not in the "Key specifications" list? You cannot expect your readers to remember these things by heart for every camera - I certainly don't.
Plus, honestly, if sensor size isn't a key specification, then what is?

Otherwise, an interesting read. Still, I don't think the "shallow DoF" argument works for MF anymore, with the sensors only slightly larger than FF, and with so many FF ultra-large aperture lenses (f/1.2 to f/1.4) available.

Direct link | Posted on Jul 11, 2015 at 07:40 UTC as 19th comment | 4 replies
In reply to:

acidic: My 24mm 1.4 can also do 35mm 1.4, and everything in between. It's called Crop Tool.

But your image is still not similar to an image taken with a 35/1.4. It will have similar DoF as a picture taken with a 35/2, and of course you lose over half of your pixels when cropping (24/35)^2 = 0.47 = 47% of pixels left.

Direct link | Posted on Jun 23, 2015 at 09:08 UTC
On 2015 Superzoom Camera Roundup article (166 comments in total)

Dear DPreview,

why isn't sensor size at (or close to) the top of the "Specs Compared" table on Page 1? What is it that makes you unwilling of putting sensor size as a key parameter in your tables, be it a camera or mobile phone comparison?

Please?

Direct link | Posted on Jun 17, 2015 at 22:00 UTC as 50th comment | 4 replies
In reply to:

Zeisschen: APS-C upgrade for M43 cameras for 650 bucks?

Androole, I'm afraid you are confused.

1) The Canon 1Ds series is FF, not APS-H.
2) The Canon 1D series without the 's' is APS-H.
3) The crop factor with the 0.64 Speedbooster is 1.28, while the crop factor of Canon's APS-H cameras has been 1.26-1.29, so they are the same.
4) "difference by area is 1.23 stops for FF -> M4/3"; what? FF sensor area (864 mm²) is around 3.8 times that of M43 (225 mm²). If you want to convert that to "stops", that would be log(864/225)/log(2) = 1.94.

Direct link | Posted on Jun 9, 2015 at 08:00 UTC
On Nikon D7200 Review preview (555 comments in total)
In reply to:

Papi61: Nowadays lots of videographers buy DSLR's and mirrorless cameras for video. Canon and Nikon are deliberately ignoring this and their camera sales are going south accordingly.

I'm a Nikon shooter and I would have loved to buy a D7200 with 4K video. Instead I bought a Samsung NX1. To my surprise, it bested my D5300 and even D750 (except for high ISO on the latter) even as a still camera. If Samsung keeps releasing bodies like this and adds a sufficient number of lenses to become competitive with Canon/Nikon, I seriously doubt I will stay in the Nikon camp in the future.

Papi61:
Canon ignoring videographers? Don't they have the best video AF system ever?

Direct link | Posted on Jun 4, 2015 at 17:44 UTC
On Nikon D7200 Review preview (555 comments in total)
In reply to:

naththo: Much better than Canon 7D MKII for the dynamic range and less shadow noise. But AF seems a concern though is a let down due to critical problem in review. Still that beats Canon again and again without stopping.

Even though I respect and quite agree with the notion that cameras with OVF's have and will always have AF issues, and even though EVF's have become better and faster over time, in my opinion it would still be worth mentioning that OVF's are really, REALLY fast. After all, they literally work at the speed of light!

Summary: if you are going to gripe a lot about inherent negative characteristics of OVF cameras, it would be only fair to give some credit for the equally inherent benefits. OVFs are not there _just_ because traditional camera companies are visionless and greedy.

Direct link | Posted on Jun 4, 2015 at 17:40 UTC
On Nikon D7200 Review preview (555 comments in total)
In reply to:

Edgar_in_Indy: Why the heck is 1080p mentioned prominently as a "feature" a couple times early in the review? At this point, when even cell phones are starting to record 4K, 1080p is no longer a "feature" in a high-end camera. It's a liability.

And for a lot of people, that 1080p "feature" will remove a camera like this from their consideration.

Q: "So what are the compelling arguments *against* giving a 2015 state-of-the-art, semi-pro, $1200 DSLR 4K capability?"

A: How about Canon 70d's unparalleled focusing capability with its dual-pixel CMOS AF? While 70d's video quality is not the best, your shots are very much in-focus. And, in my opinion, In-focus video trumps out-of-focus video any time (you said you make home videos, so I assume you don't have a professional focus puller alongside you).

(I know, this has little to do with the Nikon D7200. But it _is_ an answer to the question.)

Direct link | Posted on Jun 4, 2015 at 17:25 UTC
On Nikon D7200 Review preview (555 comments in total)
In reply to:

Henrik Herranen: Page 13: "The ISO 100 shot pushed by 5EV is essentially indistinguishable from the native ISO 3200 shot, even down to the darkest tones in the image."

What?
They are not even remotely indistinguishable. ISO100+5EV is purple and ISO3200 is gray. It's not even close!

Don't get me wrong. I really respect the concept of ISO invariance and that DPreview has chosen to make it a Big Thing. In time, it will force my camera maker of choice, Canon, to improve. Still, if you give zero attention to obviously different colour, then what we are eventually going to get is that no-one will care about colour anymore. We'll get even weaker colour filter arrays, missing IR filters, etc.

Suggestion:
How about making sensor spectral response a part of the standard things to test? That way we'd, with time, end up with better colour accuracy. While I respect high-ISO capability, I think it has been "good enough" since Canon's 5D2. All this while colour fidelity, which used to be pretty good, has suffered.

Richard: Thank you for your answer. Obviously such a colour test should be done at least in the RAW image domain, and perhaps additionally with a low-key JPEG engine mode, too.

Dr_Jon: I've also been noticed those rumours, and would very much like to know the truth.

As said, I think all modern cameras have "good enough" high-ISO performance. However, we have very little scientifically tested and presented information about colour fidelity. Spectral response curves of some common cameras would go a long way in mitigating that.

Direct link | Posted on Jun 4, 2015 at 10:42 UTC
On Nikon D7200 Review preview (555 comments in total)

Page 13: "The ISO 100 shot pushed by 5EV is essentially indistinguishable from the native ISO 3200 shot, even down to the darkest tones in the image."

What?
They are not even remotely indistinguishable. ISO100+5EV is purple and ISO3200 is gray. It's not even close!

Don't get me wrong. I really respect the concept of ISO invariance and that DPreview has chosen to make it a Big Thing. In time, it will force my camera maker of choice, Canon, to improve. Still, if you give zero attention to obviously different colour, then what we are eventually going to get is that no-one will care about colour anymore. We'll get even weaker colour filter arrays, missing IR filters, etc.

Suggestion:
How about making sensor spectral response a part of the standard things to test? That way we'd, with time, end up with better colour accuracy. While I respect high-ISO capability, I think it has been "good enough" since Canon's 5D2. All this while colour fidelity, which used to be pretty good, has suffered.

Direct link | Posted on Jun 3, 2015 at 19:04 UTC as 124th comment | 3 replies
In reply to:

rrccad: this article is only partly right, foveon sensors can pretty easily shoot black and white as well.

Actually, badi, because I don't know you, I didn't know that you knew. But, you know, now I know you know, it's all good, no?

But, in all seriousness, many people don't know these things about sensors and their spectral response, so I felt the need to elaborate. Plus, it's not like Leica always bothered to put the necessary filters in front of their sensors. We all still remember what happened with the M8 and IR, don't we?

Direct link | Posted on Jun 3, 2015 at 18:11 UTC
In reply to:

rrccad: this article is only partly right, foveon sensors can pretty easily shoot black and white as well.

badi wrote:
"With the "no fliter" (leica monochrome) approach means you have maximum lightness fidelity."

If there really is no filter in front of the sensor, that would not be true at all.

A CMOS sensor's frequency response is way different from human vision, it being most sensitive from red to infrared, while human vision is most sensitive around green to blue (greenish in good light, more bluish near dark).

In other words: without any kind of filter, the lightness perception of a CMOS sensor is not at all what a human sees.

(And no, I don't know how Leica has corrected for that. My guess would be that some kind of IR-blocking filter is pretty much a must for any fidelity. It would be interesting if someone were to measure and report the frequency response of the Monochrom Typ 246 camera with some typical, nice lens.)

Direct link | Posted on Jun 3, 2015 at 10:33 UTC
In reply to:

Photofix101: It's the 21st century. Any camera can shoot black & white, yes even my phone. I'm impressed that Leica's brand is so strong that it can demand such s premium. Well played.

Dr_Jon:
What you write would be true if BOTH of the following conditions hold:
1) You photograph something that is pure red or blue, and
2) Your camera's colour filters are ideal.

Actually, point (2) isn't nowadays true at all. To make cameras more sensitive, the colour filters of many cameras have been weakaned during the last five years, making red filters let through also some green and blue, and so on. So, you can - if signal processing is done properly - get surprisignly good B&W renditions from many modern cameras even if of subjects with strong colours.

Direct link | Posted on Jun 3, 2015 at 10:23 UTC
In reply to:

p1726: I kind of wonder is this a color LCD panel at the back or they managed to source BW ones from somewhere?

It's a colour panel.

Reference:
http://www.dpreview.com/articles/2035743820/leica-m-monochrom-typ-246-hands-on?slide=14
"Peaking comes in red, green and blue flavors, though each of the colors stands out particularly well against the black and white images on screen"

Direct link | Posted on Jun 3, 2015 at 09:56 UTC
On Samsung working on slimmer RBW camera sensors post (82 comments in total)
In reply to:

ProfHankD: This is an interesting variation on the old CMYG and RGBW CFAs. The biggest issue I see is that RB will get far less light than W pixels, meaning they'll probably be noisier, and hence G derived from W-(R+B) will be noisy too. Then again, you could almost get away with just using the W channel as G... or W is roughly Y in JPEG's YUV. W pixels also will saturate fast, possibly limiting dynamic range. In sum, there's a lot to empirically test here....

Prof:
Using a W channel would help reduce luminance noise. Chroma noise wouldn't get much better (W would have less noise than G, but not by much, and R+B would has as much noise as before), but it can be filtered much more harshly than the luminance information without serious visible effects.

Barty:
Putting an ND filter in front of the W elements would negate the whole point of using a W channel in the first place: better quantum efficiency.

Direct link | Posted on May 30, 2015 at 10:05 UTC
On Samsung working on slimmer RBW camera sensors post (82 comments in total)
In reply to:

GrahamSeventy1: What's the obsession with thinner & thinner phones ?, I'd rather have a phone were the battery last's a few days !

viking: I think that's the point: why not make the whole phone one millimeter thicker? Then the camera module wouldn't protrude, and battery life could be increased.

I am still using an ancient Nokia not-so-smartphone. Although its battery is nothing like it was when new, a full charge still lasts for almost a week in light use. I don't find it personally appealing to replace it with a phone that ties me to a charging station every few hours.

Direct link | Posted on May 30, 2015 at 10:02 UTC
Total: 159, showing: 1 – 20
« First‹ Previous12345Next ›Last »