Richard Butler

Richard Butler

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

Comments

Total: 6008, showing: 41 – 60
« First‹ Previous12345Next ›Last »
On article Fujifilm X-H1 Review (1566 comments in total)
In reply to:

Ebrahim Saadawi: Great, great camera, yet the price, considering the release of the Full-frame Sony A7 III, is a bit high.

That example was for a uniformly lit scene but it works similarly well for any real scene. If given a lens with the same angle of view (so both 'see' the same scene), then every element in the scene is recorded over a larger area (larger in proportion to the increased sensor size). Since F-number dictates light per unit area, this means each element is made up from more light (since it takes up a larger area of the sensor).

This explains the better low light performance (at the same F-number, ISO and shutter speed, a larger sensor gets more light. And hence is less noisy). It also helps explain the great dynamic range: if both sensors can tolerate the same number of photons per sq mm before clipping, then the large sensor still records more light about each object, before clipping. This means every tone will be cleaner, so your darkest tone that's distinguishable before being swamped with noise (so long as your sensor isn't contributing too much noise), will be a darker tone. More DR.

Link | Posted on May 10, 2018 at 18:23 UTC
On article Fujifilm X-H1 Review (1566 comments in total)
In reply to:

Ebrahim Saadawi: Great, great camera, yet the price, considering the release of the Full-frame Sony A7 III, is a bit high.

At its most simple, it's because the same F-number gives the same number of photons per sq mm, so if you can project a larger version of the image onto a larger sensor, while maintaining the same number of photons per sq mm, then you get more photons describing every element in the scene. More light = potential for cleaner images (and, unless your sensors is *terrible*, then actually cleaner images).

Just to make the maths easier, let's assume your APS-C chip is exactly a 1,5x crop. You set your lens to 18mm on the full frame camera and 12mm on the APS-C camera. Both will see the same view of the scene, if shot from the same location. Set them to the same F-number (let's say F4).

The full frame sensor sees the world through a 4.5mm diameter aperture. Let's say it receives *x* photons per sq mm (eg you're pointing at a featureless grey sky). It receives *x* photons \* 864mm. The APS-C sensor sees through a 3mm dia aperture. It receives *x* photons \* 384mm (the sensor area).

Link | Posted on May 10, 2018 at 18:14 UTC
In reply to:

Daft Punk: I recently switched from Sony to Fuji.

The FF factor is exaggerated. I can't see any massive difference in print quality Fuji compared to Sony. I do prefer Fuji files SOOC.

I am absolutely loving my Fuji move. They have got me interested in photography again. The XT2 and XH-1 are giving me fun like I used to have with my Nikon FM2 when I learned the basics all those years ago.

Agreed.

At matched equivalent aperture, there's rarely a large difference (which is why I never understand the accusation that equivalence favours a particular format: if anything it also highlights how wide the range of settings is where they're no significant difference).

Link | Posted on May 9, 2018 at 23:16 UTC
In reply to:

RedFox88: What's a flog?

It's a tone curve with a logarithmic response across much of its range. ie: it devotes the same number of available levels to each stop of captured light (except in the shadows). Unlike most still image tone curves that devote most values to the tones around middle grey.

The advantage is that it tends to increase the ability to capture dynamic range, while maintaining maximum processing flexibility. The downside is that, especially in 8-bit capture, you're spreading your 256 available values pretty thin, so you risk posterization, especially if you mis-expose the footage.

Link | Posted on May 9, 2018 at 21:13 UTC
In reply to:

Daft Punk: I recently switched from Sony to Fuji.

The FF factor is exaggerated. I can't see any massive difference in print quality Fuji compared to Sony. I do prefer Fuji files SOOC.

I am absolutely loving my Fuji move. They have got me interested in photography again. The XT2 and XH-1 are giving me fun like I used to have with my Nikon FM2 when I learned the basics all those years ago.

However, larger sensors will often offer more dynamic range in bright light conditions and, if you can mount lenses that are brighter (in equivalent terms - eg by being the same F-number) on the larger sensor, it will usually offer shallower depth of field, better low light performance.

Link | Posted on May 9, 2018 at 21:08 UTC
On article Blackmagic Design announces Pocket Cinema Camera 4K (408 comments in total)
In reply to:

historianx: "Full size" 4/3rds sensor? That's redundant. All 4/3rds sensors are one "full" size.

Except the ones that are oversized (eg: GH2 and GH5S).

Link | Posted on May 9, 2018 at 19:55 UTC
In reply to:

LP0WELL: I'd suggest studying how feedback-controlled motion stabilization works its magic before leaping into the fray with your finest gear. A GH5s with a Sigma 18-35mm T2 weighs almost 5 lbs. The DJI Ronin M may be rated at up to 8 lbs payload, but that is a lot more weight than you'd want for optimizing its real-life stabilization performance. A lightweight, fixed-length native Lumix lens with video-friendly OIS would be a far more practical choice. The Lumix 30mm f2.8 macro and the Lumix 42.5mm f1.7 are good starting points for evaluation.

This is another thing I learned. I'd got hold of some nice gear to let me see what the camera was capable of, but it didn't necessarily play well with the gimbal, particularly not with a first-time operator.

Link | Posted on May 9, 2018 at 19:43 UTC
In reply to:

rationalist: I just made some experiments with my Fuji X–T2. Looking at the raw files in Capture One there is only a very slight difference in the highlights between DR100 and DR400, more noticeable in the histogram. Differences in the Jpegs are much more prominent. I therefore assume much more is done in the tone curve during jpeg conversion. At DR400, however, there is visible noise in the image which can be expected at ISO800. DR expansion done manually gives much better results. So whats the point of this function?

As soon as you use a shorter exposure, you'll get more noise everywhere. It'll be most visible in the shadows, but all tones in your image will have a worse signal to noise ratio because they're made from less light.

In bright situations, where you might move from DR100 to DR400, and hence see the minimum available ISO increase, this is the only real change: a two stop reduction in exposure (indicated as two EV increase in ISO). There's then a change in the JPEG tone curve that ensures that middle grey looks right but with additional highlights incorporated in the DR400 shot.

DR400 uses the same amplification at ISO 800 as DR100 does at ISO 200 (per the table in the article).

You can test this by shooting DR100/ISO 200 and DR400/ISO 800 with the same exposure settings. The two Raw files will be identical in an analysis tool, such as Raw Digger. ACR and C1 will initially render them differently, but if you brighten the DR100 shot by 2EV, it'll look an awful lot like the DR400 shot.

Link | Posted on May 9, 2018 at 19:01 UTC
On article Fujifilm X-H1 Review (1566 comments in total)
In reply to:

Ebrahim Saadawi: Great, great camera, yet the price, considering the release of the Full-frame Sony A7 III, is a bit high.

A 50mm F1.8 will project the same number of photons per sq mm, regardless of what sensor you put it in front of (after all, it has no knowledge of such things).

However, as soon as you relate it to photography (where you're presumably composing a specific framing), it becomes more complex.

Mount a FF 50mm F1.8 on a FF camera and an APS-C camera and, from the same position, you get all the edges cropped off on the APS-C camera, which is useless. Move further back with the APS-C camera (if you can) and you change the perspective.

Ultimately, the larger sensor 'sees' a larger proportion of the imaging circle. And, since you're getting the same number of photons per sq mm, the sensor with more sq mm of surface area gets more photons. This occurs regardless of pixel size and is the main reason larger sensors are able to take cleaner images.

Link | Posted on May 9, 2018 at 17:25 UTC
On article Video: Understanding the science of camera sensors (72 comments in total)
In reply to:

Gesture: What about Foveon?

FLruckas: I tried to make that point in the Quattro H review (which is much more direct than me trying to say it to them, since they then also get to see the public reaction to it).

The Quattro H is a fascinating, if rather niche camera. Given how specialised it is, many photographers will need a second camera (for low light or moving subjects), so the use of a more popular lens mount (or one that can be adapted to other mounts) would appear to broaden its appeal.

Link | Posted on May 8, 2018 at 21:32 UTC
On article Video: Understanding the science of camera sensors (72 comments in total)
In reply to:

Gesture: What about Foveon?

MrBrightSide: Sigma also provides the access and trips you assume are corrupting the entire press. (Despite there clearly being a clear business benefit to be the truth teller that cuts through any hype/lies).

There are fundamental engineering challenges that Foveon faces, in terms of noise and quantum efficiency. The Foveon concept is brilliant, but it's very hard to implement. Bayer, by contrast, will always suffer from an additional form of spatial aliasing (color moiré) and the cost of throwing away half of the received light at every pixel, but it does well *enough* in enough circumstances that it dominates the market. Part of the reason I know this is because of the time I've spent talking to Foveon executives and engineers.

It would be an ethical failing to solely concentrate on the benefits without proportionately addressing the (observable) limitations. We worked very hard to strike that balance in our review of the Quattro H.

Link | Posted on May 8, 2018 at 20:09 UTC
On article Video: Understanding the science of camera sensors (72 comments in total)
In reply to:

MrBrightSide: Everything he says directly contradicts the DPR model of photons and buckets. I'm calling baloney.

No, it doesn't. What he's describing around [six minutes into the video](https://www.youtube.com/watch?v=MytCfECfqWc&feature=youtu.be&t=6m8s) is exactly the same thing. Collecting a charge (electrons) produced by the number of photons that have hit the sensor (there's a maximum capacity for electrons: the 'Full Well Capacity' that defines the depth of the metaphorical bucket).

Link | Posted on May 8, 2018 at 19:57 UTC
In reply to:

rationalist: I just made some experiments with my Fuji X–T2. Looking at the raw files in Capture One there is only a very slight difference in the highlights between DR100 and DR400, more noticeable in the histogram. Differences in the Jpegs are much more prominent. I therefore assume much more is done in the tone curve during jpeg conversion. At DR400, however, there is visible noise in the image which can be expected at ISO800. DR expansion done manually gives much better results. So whats the point of this function?

I'd be interested to hear what you mean by 'DR expansion done manually.'

If there's litter difference in Capture One between DR100 and DR400 shots, then it sounds like it's doing what Adobe Camera Raw does, by default: applies the same tone curve, clipping the additional highlights that were captured (you need to use the highlight recovery tool to reintroduce them).

Note that the greater noise comes from using a shorter exposure, so it's a direct consequence of the action that captured those extra highlights.

Link | Posted on May 7, 2018 at 18:43 UTC
On article Sony a7R Mark III review (1255 comments in total)
In reply to:

HenryDJP: Hilarious! On the list of CONS, it says "No Playmemories app Support". As far as I'm concerned that's a huge plus. The desktop version as well as the app in my Sony A6300 is absolutely atrocious. Sure it "WORKS", but it's obvious to the most computer/camera illiterate that it's a poorly written app. Sluggish performance, not 100% reliable, and the setup could be a lot simpler. Until Sony gets it's act together with Playmemories, I have removed it from my MacBook and I don't even use the Smartphone control feature.

It's support for the in-camera apps that have been removed.

Link | Posted on Mar 30, 2018 at 17:31 UTC
In reply to:

captura: When will Sony filter this new sensor technology over to the Axxx series cameras?

BSI [significantly pre-dates the NX1](https://www.dpreview.com/articles/4281372431/sonycmos).

Even at the time Sony said that most of the efficiency benefits were in (very) small pixels in (very) small sensors.

Link | Posted on Mar 23, 2018 at 20:04 UTC
In reply to:

rmexpress22: This article convinced me to switch to Sony so that I can make high-quality charts.

/sarcasm

We tend to find the results from our charts correlate pretty well to real-world use, or we'd not bother shooting them.

But nothing about them is designed to convince anyone to take any specific action, they're intended only to provide information about camera performance.

Link | Posted on Mar 23, 2018 at 18:47 UTC
On article Fujifilm X-H1 Review (1566 comments in total)
In reply to:

Ebrahim Saadawi: Great, great camera, yet the price, considering the release of the Full-frame Sony A7 III, is a bit high.

The a7 sensor underperforms while the a6500 (and X-T20, since it's likely to be essentially the same sensor) performs very well. This will narrow the 1.3EV gap in terms of the difference in total light they get.

Light per unit area is the same at matched exposure but if two sensors have different areas then they get different amounts of total light. And total light available to the sensor, with some correction for how efficient the sensor is, is pretty much the best predictor of noise/IQ.

Though it's worth noting that small sensors tend to be more efficient, since they tend to have more advanced technology than large sensors (which inherit older technologies as the fabrication lines are replaced with finer-scale ones for the next-gen of smartphone sensors). Large sensors can get away with being a bit more wasteful of light.

Link | Posted on Mar 23, 2018 at 00:20 UTC
On article Sony a7 III Review (2167 comments in total)
In reply to:

Rick Knepper: Here's what I do not understand about the DPR studio comparison methodology and the consumers of said comparison. Unless this is a system to system comparison, using different lenses does not really provide accurate insight. Especially with Canon, Nikon and Sony, there are plenty of 3rd party lens makers producing the same lens on different mounts.

Can't Mr. Bezos peel off a few bills from his coffee money to bolster the DPR budget?

All the lenses we use are very sharp at F5.6 in the center. If you see significant difference in sharpness/detail at the edges that aren't present nearer the centre, then it's likely to be a lens difference, but the difference in good primes stopped-down are not huge.

Link | Posted on Mar 22, 2018 at 20:10 UTC
On article Sony a7 III Review (2167 comments in total)
In reply to:

Jonathan Brady: Do you guys think any work got done at Sony's headquarters when the image quality tests started coming out from the Canon 6D Mark II, or did the overwhelming chorus of laughter keep them distracted all day?

**Joed700**: Are you using an Apple computer with a 'Retina' screen, or its equivalent?

Our site shows 100% crops (ie: a 1:1 relationship between the pixels in the file and the displayed image), if you're seeing them magnified, that's something occurring at your end.

Link | Posted on Mar 22, 2018 at 17:56 UTC
On article Sony a7 III Review (2167 comments in total)
In reply to:

Rick Knepper: Here's what I do not understand about the DPR studio comparison methodology and the consumers of said comparison. Unless this is a system to system comparison, using different lenses does not really provide accurate insight. Especially with Canon, Nikon and Sony, there are plenty of 3rd party lens makers producing the same lens on different mounts.

Can't Mr. Bezos peel off a few bills from his coffee money to bolster the DPR budget?

On the previous two occasions we've tested the use of a single lens design across multiple systems, we've found more inconsistency than with our current system (there's also the issue of which focal length to use: a 50mm is longer than we want for Four Thirds sensors but shorter than ideal for full frame).

On the basis that most people use own-brand lenses on their cameras, using a good, sharp, consistent lens stopped down to F5.6 on all systems seems like the most real-world relevant answer (since all answers have their pros and cons).

Link | Posted on Mar 22, 2018 at 17:44 UTC
Total: 6008, showing: 41 – 60
« First‹ Previous12345Next ›Last »