Richard Butler

Richard Butler

DPReview Administrator
Lives in United Kingdom Seattle, United Kingdom
Joined on Nov 7, 2007

Comments

Total: 2932, showing: 1 – 20
« First‹ Previous12345Next ›Last »
On Canon EOS-1D X overview article (368 comments in total)
In reply to:

Pascal Parvex: Well, Amazon.com now has it, so where is the review? :)

leica f64: Why would DxO's supposed bias have any impact on our editorial decisions? (Answer: they wouldn't/don't).

Also, where does the fact that we weren't able to conduct a full Nikon D4s review sit with your conspiracy theory about the decision being brand-based (which it wasn't), rather than complexity/resource based?

Direct link | Posted on Apr 20, 2015 at 20:16 UTC
In reply to:

maxnimo: Why do they call it a 1" sensor when not even the diagonal is even close to 1" in length?

maxnimo: it's both. Four Thirds type sensors are both 4/3"-type, using the ridiculous standard terminology and are 4:3 aspect ratio.

Direct link | Posted on Apr 20, 2015 at 16:53 UTC
On Opinion: Why the Canon XC10 is a big deal article (802 comments in total)
In reply to:

regordyoll: Why a 1" sensor when a Sony A7 has a full frame sensor or a Sony A6000 has an APS-C sensor? I'd like to understand the reason. Thanks!

@ColdViking. It's not so much 'absolute nonsense' as 'demonstrably true.'

The 'with a 27-270mm equiv zoom' part of the statement is important.

Even with equivalent apertures, a GH4 with 13.5-135mm lens would be larger and heavier than this, while an a7S with a 27-270mm lens will be larger and heavier still.

The URSA Mini is also a cool product but it wasn't publicly known when this article was written. It's certainly more capable but it's also larger and twice the price (suggesting it's aimed at a slightly different audience).

Direct link | Posted on Apr 16, 2015 at 21:49 UTC
In reply to:

Nukunukoo: Just trying to do the math. Most anamorphic lenses would turn a 4:3 ratio into the more conventional 16:9-ish ratio. I may be wrong, that means that if I want cinema aspect, my SLR Magic 1.33x won't do. I need a 2x for that.

I think the SLR Magic 1.33x was specifically designed to squeeze a *roughly* 'Scope ratio down onto 16:9. (16*1.33)/9=2.36, so it's 1:2.36. Consequently it's not so useful on 4:3 (as you point out, that would only give you 15.96:9).

However, 2x on 4:3 would give you 8:3, which is 2.66:1

Direct link | Posted on Apr 13, 2015 at 17:40 UTC
On Samsung NX1 Review preview (1018 comments in total)
In reply to:

DGrinb: No in-body image stabilization, proprietary lenses....how come it's good??

@DGrinb - That's true, but not necessarily the only outcome. Canon's EF mount is, to the best of my knowledge, a closed, proprietary system.

Direct link | Posted on Apr 10, 2015 at 19:26 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

Lou P Dargent, our statements agree with one another.

I agree entirely that the elegance (and value) of the f-number system is that, with the same F-number, any two lenses will project the same amount of light for any given subject in the scene. Light per-unit-area on the sensor is the same.

My point is then that, with equal light-per-unit area, the sensor with the largest area receives more light. Nothing I've said disputes your most recent statement.

Direct link | Posted on Apr 10, 2015 at 19:06 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

Exposure as it's used in photography (and I'm not suggesting abandoning it) is based on F-numbers, which tells you about light-per-unit-area. Clearly, at this point, the amount of area available for light capture then plays an important role.

However, the ISO standard then requires whatever amplification and digital pushes are necessary to cancel-out the sensor size difference and provide a consistent light intensity-to-output brightness, which obscures these differences.

And yet most people recognise these differences: smaller sensors have a narrower field of view, noisier images and less shallow depth-of-field, when using the same lens.

The oddity (to me) is that most people are happy to accept the first of those (equivalent focal lengths) but are reluctant to examine the effects of the exposure system we all use, to recognise the reason for the other effects of sensor size that they see.

Direct link | Posted on Apr 10, 2015 at 18:53 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

And, if you could open the APS-C lens up to F2.67 (so that you had a 37.5mm entrance pupil diameter for both lenses), then you'd get more light on the APS-C sensor.

The crucial thing is that (with the same shutter speed), you'd get more light per-unit-area on the small sensor in exact proportion to how much smaller the sensor is, than the full frame camera.

At which point, you'd get a better pixel-level performance on the APS-C camera and the two images would look extremely similar (I won't claim identical) when scaled to the same output size.

Direct link | Posted on Apr 10, 2015 at 18:42 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

Like this, for instance: 16MP D7000 vs 36MP a7R. [Alternate between full size and 'Print' resolution to see thee effect of scaling to a common output size](http://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=sony_a7r&attr13_1=nikon_d7000&attr15_0=raw&attr15_1=raw&attr16_0=6400&attr16_1=6400&normalization=full&widget=170&x=0.034839129400584874&y=0.4957593209940143)

The full story is, of course a little more complex, but this illustrates the point that the broader point works in the real world, not just in theory.

Direct link | Posted on Apr 10, 2015 at 18:38 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

For photographic examples of this topic, [please read this article](http://bit.ly/equivap).

But let's imagine this situation:
36MP Full frame camera, 150mm F4 lens
16MP APS-C camera, 100 F2.8 lens

Both cameras have essentially the same pixel density. Shot from the same position relative to the subject, they have the same field of view. This should be fairly uncontroversial.

Set both lenses to F4 and they will both project the same intensity of light onto their sensors, so every pixel will receive the same amount of light. The pixel-level noise performance will be identical.

However, one of them captured 36million points of data, the other only 16million. Take the 36million pixels and down-scale them to 16MP and you'll generally end up with a cleaner image (The noise averages as you downscale, so you end up with cleaner images).

Direct link | Posted on Apr 10, 2015 at 18:36 UTC
On Samsung NX1 Review preview (1018 comments in total)
In reply to:

DGrinb: No in-body image stabilization, proprietary lenses....how come it's good??

I think we're using different definitions of the word 'proprietary.'

The Micro Four Thirds mount is owned by two companies, rather than one, but it's still owned and controlled by those companies.

Direct link | Posted on Apr 10, 2015 at 17:34 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

Because the primary source of noise in an image is often shot noise (the randomness inherent in light), the total amount of light captured is often *the most* relevant factor in noise performance.

And, because total light is *strongly* dependent on sensor size, it ends up being the most significant factor in image noise in most situations (assuming contemporary sensor design).

To stress the impact of pixel density (at large sensor sizes), the Sony 16MP and 24MP APS-C sensors produce essentially the same noise, when scaled to a common size, and Sony's a7R is less noisy than its a7II, despite a higher pixel count. It may play more of a role in high pixel count, tiny smartphone sensors, but that's not the topic in hand.

Direct link | Posted on Apr 10, 2015 at 17:30 UTC
On Samsung NX1 Review preview (1018 comments in total)
In reply to:

DGrinb: No in-body image stabilization, proprietary lenses....how come it's good??

Which camera brand doesn't use proprietary lenses?

Direct link | Posted on Apr 10, 2015 at 17:15 UTC
On Samsung NX1 Review preview (1018 comments in total)
In reply to:

Albert Silver: Just a few notes:

1) Judging by the noise seen in the Real World DR comparison, the NX1 appears to be about a full stop worse than the D5500. Whether this qualifies as almost as good is of course subjective, but it is still a distance away.

2) As of this comment, the ISO-invariance at the end is broken, and no image appears when trying to compare with other cameras.

3) The AF performance results are all over the place when scrutinizing reviews, which suggests that Samsung has considerable work to do.

It is fascinating to see Samsung enter the DSLR market precisely now, when everyone is biting their nails on what the future holds.

I'm not sure what you mean when you say the ISO-invariance section is broken. Could you please send me a private message and I'll try to get to the bottom of any problem.

It's also worth noting that AF performance has improved significantly with some of the more recent firmware updates, so any reviews based on v1.0 are likely to tell a very different stories to test being conducted today.

Direct link | Posted on Apr 10, 2015 at 17:07 UTC
On Canon XC10: What you need to know article (227 comments in total)
In reply to:

John C Tharp: Some goofy points:

-305Mb/s is still possible on the faster SD cards
-That baseplate looks like it will get in the way of gearing for zoom and focus pulling
-The type of AF motor used isn't mentioned- if it's a USM-type with direct manual override, cool, but if it's an STM-type, focus pullers will likely be frustrated!

@John C Tharp: as in the example I gave above, the SanDisk card you quote can write at *up to* 90MB/s. It's the minimum sustained write speed (which isn't always published), that's the issue.

Direct link | Posted on Apr 10, 2015 at 17:02 UTC
On Opinion: Why the Canon XC10 is a big deal article (802 comments in total)
In reply to:

The Squire: CORRECTION?

In the comparison table with the FZ1000 and RX10 it lists the Canon as featuring "2X oversampling for HD video", but none of the other cameras do this.

The RX10, at least over-samples - It does a full sensor scan to create the HD image - No pixel bining. No centre of sensor crop.

Or do you mean something else?

If you look at the Sony a7S's 1080 video, you'll see it is essentially moire-free because it captures 3840 x 2160 footage (1920\*2 x 1080\*2), and intelligently downsamples it.

This means it is shooting at a the precise resolution that allows 1080lph to be represented, without aliasing (1080 is the Nyquist frequency of 2160).

The RX10 and others may perform full sensor readout (at still higher frequencies), but what they don't manage to do is intelligently reduce this to moire-free 1080.

The '**2X**' aspect is relevant, here and the XC10 supposedly gains the same downsampling system Canon developed for its C500, so it's reasonable to expect it to be very good, in much the same was as the a7S is and the RX10 and FZ1000 aren't.

Direct link | Posted on Apr 9, 2015 at 18:47 UTC
On Canon XC10: What you need to know article (227 comments in total)
In reply to:

jtan163: @Barney Britton - any news on the video toolset - e.g. zebras, focus peaking (for those of us who don't trust the AF), vector scopes and the like?

I like it, I can't afford it, but I'm very tempted anyway.

It has zebras (70-100%), peaking (three colours), markers, time code and colour bars (EB/SMPTE on the PAL versions, at least).

Direct link | Posted on Apr 9, 2015 at 01:06 UTC
On Canon XC10: What you need to know article (227 comments in total)
In reply to:

John C Tharp: Some goofy points:

-305Mb/s is still possible on the faster SD cards
-That baseplate looks like it will get in the way of gearing for zoom and focus pulling
-The type of AF motor used isn't mentioned- if it's a USM-type with direct manual override, cool, but if it's an STM-type, focus pullers will likely be frustrated!

As I say, it's quite possible that some of these cards are capable of such speeds. The problem is that there's not yet a standard stringent enough to guarantee it.

Yes there are UHS-II cards that say '280MB/s' on the front. These say they can write at *up to* 250MB/s, but the U3 symbol on the front only *promises* that they won't dip below 30MB/s.

Direct link | Posted on Apr 9, 2015 at 01:01 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

The thing I find interesting is that people recognise the relationship between sensor size and focal length (and are quite happy to use equivalent focal lengths), they also recognise the relationship between aperture and sensor size (increased depth-of-field) and they recognise the relationship between ISO and sensor size (noisier images on smaller sensors), but some people are fiercely resistant to following the logic through and acknowledging the pattern.

(Not that I'm suggesting using equivalent F-numbers for anything other than identifying the different *potential* of different camera/lens systems). JPEG-based ISO is also limiting for Raw photographers, but that's another matter entirely.

Direct link | Posted on Apr 9, 2015 at 00:42 UTC
On Fujifilm X100T Review preview (643 comments in total)
In reply to:

FodgeandDurn: Apologies for putting this here instead of the 'report issues' button wherever that is - the intro says this has a "23mm f/2 lens". This actually got me excited, but unfortunately it is an error. If you've divided 35 x 1.5 to give some 35mm equivalence wouldn't it be 50mm equiv?

Hack is probably an exaggeration. My point is that, in order to provide comprehensibly consistent exposure between cameras with different sensor sizes, it ends up obscuring the fact that the effect of 'F2' changes depending on what format you're shooting on. On top of this, the current standard is really woolly and allows so much leeway that it's almost meaningless.

However, while everyone realises that a full frame camera, a Micro Four Thirds camera and a 1/2.3"-type sensor compact will all end up with different image quality (if shot with the same ISO, F-number and shutter speed), the way ISO works obscures the *reason* for this difference: each format is making its image from different amounts of total light.

Direct link | Posted on Apr 9, 2015 at 00:36 UTC
Total: 2932, showing: 1 – 20
« First‹ Previous12345Next ›Last »