Richard Butler

Richard Butler

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

Comments

Total: 3951, showing: 21 – 40
« First‹ Previous12345Next ›Last »
In reply to:

Leigh A. Wax: Read page 176 0f the 5D4 manual.

Manual from [Canon Asia](http://gdlp01.c-wss.com/gds/0/0300024970/01/eos5d-mk4-im-en.pdf).

Link | Posted on Aug 30, 2016 at 19:41 UTC
In reply to:

privatebydesign: Really? On close examination the difference between the 5D MkIV, which you call "significant" and the 1DX MkII which you say "exhibits pretty low levels" is minuscule.

The link below is to an overlay comparison.

http://www.canonrumors.com/forum/index.php?topic=30636

Both are 'accurate' in that both are representative of the performance the cameras give.

The most likely explanation is that the 1D X II shows less rolling shutter in 60p mode, which was used for the still frame grab focused update. The video shows the performance with both cameras in 30p mode.

Link | Posted on Aug 30, 2016 at 18:10 UTC
In reply to:

Treeshade: Why would DPR publish this test while they didn't do the same with a6300 and NX1?

Well, according to Canon users in this forum, Canon has the largest ILC market share. Most pros use 5Dmk2/3. They would never switch. 5Dmk4 will outsell a6300/NX1. Other brands doesn't matter.

So it is logical to put extra effort into testing this would-be popular camera of Canon, isn't it?

We'd have covered the a6300's rolling shutter if we'd noticed it to the same extent as the 5D IV.

However, being more aware of it now, I went out of my way to include the a6300 as a reference point so that we didn't appear to suggest that rolling shutter is peculiar to the 5D IV or fail to give some context to its performance.

Link | Posted on Aug 30, 2016 at 18:05 UTC
In reply to:

TigTillinghast: I know very little about video shooting, but isn't the difference in slope from those images almost completely a product of comparing the 5D4 shooting at 30fps and the 1DX shooting at 60fps? I imagine the answer is dependent on whether the sensor reads out faster with the higher fps. I think the 1DX2 shoots at 30 fps as well. It would be great to set the test with the same shooting mode to determine if the 5D4 is really doing anything differently, or if it just lacks the unique 1DX2 feature.

The video at the top shows both the 5D IV and 1D X II in their 30p modes, the stills lower down the page show the difference between 60p and 30p, since they're more concerned about the results from stills grabbing.

Link | Posted on Aug 30, 2016 at 17:55 UTC
In reply to:

mariuss: Why not using dual-pixel technology for (auto)calibrating AF from lens? Somehow like Nikon D500 does. But with dual-pixel techn. why not autocalibrating the lens automaticaly at more different focal lenghts?!
Now, that would be awesome/practical/handy!

That's a good question. Canon filed a patent for such a system a few years ago, yet we've not yet seen the fruits of that.

Link | Posted on Aug 30, 2016 at 17:50 UTC
In reply to:

Sonyshine: Jeeeez! This camera keeps getting damned with faint praise.... :(

Bear in mind we're only looking at the peripheral extra features, at this point: we're not criticizing the camera's core capability or image quality.

Link | Posted on Aug 30, 2016 at 17:47 UTC
On article An introduction to our studio test scene (110 comments in total)
In reply to:

samtheman2014: Is there some way that the test shoots could actually show equal brightness, due to the flexibility in the ISO naming conventions some camera results are darker than others suggesting that they would need more exposure to attain the same end brightness as others in essence you are not comparing like with like . If you are comparing two camera this difference becomes important

Only in a minority of modern cameras is there any significant difference between adjusting brightness through hardware amplification (which I'm assuming is what you mean by 'ISO' - even though this isn't always the case), and digital brightening.

And, in those cameras where these is a difference, it tends to stem from downstream read noise, which only plays a major role at base ISO. As soon as you start boosting the amplification, the role of downstream read noise becomes increasingly insignificant.

Digital correction allows more subtle adjustment (rather than 1/3EV steps), and avoids us conflating any effects of odd in-between ISO systems (where the 1/3 stops in between whole EVs can be more noisy than the whole stop steps).

Link | Posted on Aug 29, 2016 at 20:38 UTC
On article What is equivalence and why should I care? (2198 comments in total)
In reply to:

emxgarcia: Just to further understand and please correct me if I'm wrong.
An f/1.2 full frame lens mounted on an APS-C body it not a truly so, maybe an f/1.6.
But, an f/1.2 MFT lens on an MFT body is f/1.2

APS-C users are paying for F1.2 performance and are getting F1.2 performance.

**However**, what 'F1.2 performance' means changes depending on the format you're shooting. So, in this case, the APS-C's 'F1.2 performance' is an awful lot like full frame's 'F1.8 performance.'

They're getting exactly what they paid for. However, they're getting what they paid for both in terms of the lens and in terms of the camera that they paid for.

This has always been true: most people buying an APS-C camera are aware they're giving something up, relative to buying a full frame camera. Equivalence just helps you understand the magnitude of that difference (and the difference between *any* two formats - it's not specific to 'full frame.')

Link | Posted on Aug 29, 2016 at 19:22 UTC
In reply to:

Offside: From the home page: "The Canon EOS 5D Mark IV has significant rolling shutter, our tests show. We shot the 5D IV alongside both the EOS-1D Mark II and the Sony a6300 to show the effect."

That should read EOS 1DX Mark II.

Fixed. Thanks for that.

Link | Posted on Aug 27, 2016 at 15:55 UTC
In reply to:

DStudio: I don't think this 'defect' in Canon's new Wonder-camera is a huge publicity problem, nor do I believe this is scandalous behavior on DPR's part.

Nevertheless, I find it quite amusing that this is a "midnight release" - even on the West Coast! I wouldn't be surprised if this is about the slowest time of the entire week for the website!

What odd timing.

We're experimenting with this style of headline. We'd have just as readily used this headline about any other brand.

Link | Posted on Aug 27, 2016 at 15:48 UTC
In reply to:

wcan: I would like to know if when shooting 1080p, there are moire/aliasing/jaggies due to pixel binning or skipping, or has Canon taken the full sensor data and internally averaged it down to 1080p so as to avoid those issues?

We'll be posting that information very soon.

Link | Posted on Aug 27, 2016 at 15:43 UTC
In reply to:

rfsIII: Video sites have a rolling shutter measurement in terms of the time it takes the sensor to read out from top to bottom. It's in milliseconds IIRC. That would be a very helpful metric for DPR to add. (And then no one would complain about you whipping the camera around. )

I'm sure they'd complain if we told them any interesting results when we found them and would conclude we were doing it wrong. Or that it doesn't matter.

In all seriousness, though, yes we do need a test that allows us to measure the amount of rolling shutter do that we can test each camera without having to do a series of A/B comparisons. That's part of an ongoing project at the moment.

Link | Posted on Aug 27, 2016 at 15:42 UTC
On article What is equivalence and why should I care? (2198 comments in total)
In reply to:

emxgarcia: Just to further understand and please correct me if I'm wrong.
An f/1.2 full frame lens mounted on an APS-C body it not a truly so, maybe an f/1.6.
But, an f/1.2 MFT lens on an MFT body is f/1.2

Any lens, whatever you mount it on, *is* whatever it says it is.

So a 50mm F1.2 is a 50mm F1.2, since it knows nothing of the sensor size that's looking through it.

**However** if you mount a 50mm F1.2 on a four thirds sensor, it will give a result *equivalent to* a 100mm F2.4 lens on full frame, irrespective of what format it was designed for. Equally, a 50mm F1.2 on APS-C will behave in a manner *equivalent to* a 75mm F1.8 lens on full frame.

What the lens *is* doesn't change, regardless of what format it was designed for or what format you shoot in on. The results it gives *will* change as you mount it on different formats. Equivalence can help you understand what that change will be.

Link | Posted on Aug 27, 2016 at 09:04 UTC
In reply to:

Nobby2016: i would love when companys got rid of all video features and make the cameras a little less expensive. a two version system, one model with video features one without.

yeah yeah i know it will not happen.
and from a business perspective it makes no sense.

nikon was shunned for the lack of video in the DF, but i would not care at all.

still, all the development spend on video in DSLR/mirrorless cameras is wasted on so many of the customers. my friends have sony, nikon and canon cameras and the waste majority is using camcorders for filming.

i know it might be wrong to think so. but i can´t eliminate the feeling im paying for something i don´t need or want in a DSLR.

even when the implementation (hardware) of these video features don´t cost much, it sure has an influence on R&D cost.

ok a cameras with video features my attract more customers so the camera can be sold cheaper. who knows if that is actually true?

for video i have an extra camera i spend good money for.

Yes, and I think it's interesting that there's only one D810 (in terms of AA filter).

It's difficult to see how the D800E or 5DS R should cost more than the near-identical model.

I much prefer Ricoh's approach with their AA filter simulator, or Sony's in the RX1R II - an AA filtering effect only when you need it.

Link | Posted on Aug 27, 2016 at 08:44 UTC
In reply to:

DStudio: I don't think this 'defect' in Canon's new Wonder-camera is a huge publicity problem, nor do I believe this is scandalous behavior on DPR's part.

Nevertheless, I find it quite amusing that this is a "midnight release" - even on the West Coast! I wouldn't be surprised if this is about the slowest time of the entire week for the website!

What odd timing.

*This* is the article about the a6300 and I'll be adding a link to the Sony review in the morning.

The problem is that we don't (yet) have a standard test for rolling shutter and my 'style' (which is probably too grand a term) of shooting video didn't highlight the issue. However, having seen comments about the a6300's rolling shutter, I wanted to make sure that I gave the full context.

We do not, nor have ever, written exactly the same articles for each camera. We write about the things we notice. For instance, we wrote about how much better the 80D and 1D X II's DR was, compared to previous Canons, whereas we didn't highlight any aspect of the a6300's.

Impressions can be wrong and resilient to all manner of evidence to the contrary.

Link | Posted on Aug 27, 2016 at 08:39 UTC
In reply to:

Nobby2016: i would love when companys got rid of all video features and make the cameras a little less expensive. a two version system, one model with video features one without.

yeah yeah i know it will not happen.
and from a business perspective it makes no sense.

nikon was shunned for the lack of video in the DF, but i would not care at all.

still, all the development spend on video in DSLR/mirrorless cameras is wasted on so many of the customers. my friends have sony, nikon and canon cameras and the waste majority is using camcorders for filming.

i know it might be wrong to think so. but i can´t eliminate the feeling im paying for something i don´t need or want in a DSLR.

even when the implementation (hardware) of these video features don´t cost much, it sure has an influence on R&D cost.

ok a cameras with video features my attract more customers so the camera can be sold cheaper. who knows if that is actually true?

for video i have an extra camera i spend good money for.

I can only speak for our review, which didn't factor the Df's lack of video in, to any great extent.

Personally, I don't think adding video adds much to the cost of most cameras, but trying to differentiate between two models (one with video, one without), plus all the stocking errors that would occur as a result would *guarantee* extra expense.

Link | Posted on Aug 27, 2016 at 08:27 UTC
In reply to:

DStudio: I don't think this 'defect' in Canon's new Wonder-camera is a huge publicity problem, nor do I believe this is scandalous behavior on DPR's part.

Nevertheless, I find it quite amusing that this is a "midnight release" - even on the West Coast! I wouldn't be surprised if this is about the slowest time of the entire week for the website!

What odd timing.

Or, look at it the other way 'round.

I finished shooting and editing the video together last thing on a Friday night and stayed up until Europe woke up, so that it got seen. Believe me, I wouldn't have spent so much time annotating and preparing this if I *didn't* want people to see it!

For the record, part of the reason it took longer than planned is because I wanted to compare it to the a6300 - specifically because I didn't want anyone to think we were saying the Canon was bad and hiding that the Sony is similarly bad.

And, as I acknowledge in the text, nobody is likely to throw their camera around this much: it's a way of highlighting something that *might* subtly undermine your footage. Whether you shoot 5D IV or a6300 at 24p

Link | Posted on Aug 27, 2016 at 08:20 UTC
In reply to:

Nobby2016: i would love when companys got rid of all video features and make the cameras a little less expensive. a two version system, one model with video features one without.

yeah yeah i know it will not happen.
and from a business perspective it makes no sense.

nikon was shunned for the lack of video in the DF, but i would not care at all.

still, all the development spend on video in DSLR/mirrorless cameras is wasted on so many of the customers. my friends have sony, nikon and canon cameras and the waste majority is using camcorders for filming.

i know it might be wrong to think so. but i can´t eliminate the feeling im paying for something i don´t need or want in a DSLR.

even when the implementation (hardware) of these video features don´t cost much, it sure has an influence on R&D cost.

ok a cameras with video features my attract more customers so the camera can be sold cheaper. who knows if that is actually true?

for video i have an extra camera i spend good money for.

I'm pretty sure it wasn't the lack of video that put some people off the Df.

THe question is: how much money is actually being spent by companies on video development? Canon, Panasonic and Sony already make pro and broadcast cameras, for instance, so they're already done the R&D. So how much additional cost is really being passed on to the consumer?

There are enough cameras on the market that feel like they've had a [REC] button added because the sensor supplier pointed out that the chip could do it, with little further effort being made.

However, while adding video (in a perfunctory way, at least) costs next-to-nothing, trying to market two distinct models *would* invoke all sorts of additional costs.

Link | Posted on Aug 27, 2016 at 08:16 UTC
In reply to:

Dr_Jon: It would also be interesting to see how much less it is in 1080p if you ever get a sec? But many thanks, it's just what I asked for (although may just be coincidence of course). Pity you're not still in London, you could try my patented (err,-ish) moire/aliasing test...

I'll be passing through London in September...

Link | Posted on Aug 27, 2016 at 08:07 UTC
On article Canon EOS 5D Mark IV First Impressions Review (1245 comments in total)
In reply to:

left eye: So this dual-pixel sensor has 60MP photosites but a 30MP filter array.

In terms of noise, averaging 2 pixels will just about half visible noise.

In terms of dynamic-range the photosites are those of a 60MP sensor, so DR will not be great. I don't think averaging the result from two tiny photosites doubles DR, a blown highlight is a blown highlight.

On-chip ADC helps, but one wonders how down-scaling the 5DSR 50MP to 30MP would compare?

In many ways the 'dual-pixel' sensor of the 5D IV could be a 'single-pixel' 60MP. I'd buy the 'single-pixel' 60MP - with on-chip ADC, in an instant.

However, correct exposure on any given sensor tries to make full use of that highlight information. If you have two sensors that are identical except for a difference of full well capacity then the one with greater full well capacity would be considered to have a lower base ISO (because it can tolerate more light before overexposing).

To try to make maximal use of dynamic range, you (or the camera maker) will want to push the camera as close to clipping the first channel as possible because this maximises signal and hence minimizes noise.

Any highlight recovery comes from:
1) The two unclipped color channels (with color accuracy limited to certain colors)
2) Clever guesswork by the processing software
3) Poor exposure

All the differences in DR are in the shadows. Any highlight recoverability is a unreliable combination of luck and error.

Ultimately, digital is not film. There's a hard cut-off at one end of the sensor's response, not a linear, recoverable roll-off.

Link | Posted on Aug 26, 2016 at 18:47 UTC
Total: 3951, showing: 21 – 40
« First‹ Previous12345Next ›Last »