HumanTarget

Lives in United States United States
Joined on Jan 11, 2011

Comments

Total: 84, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

fPrime: Although CMOS sensors enabled better high ISO performance for cameras, their occluding circuitry also initially made them less light sensitive than CCD sensors. As such, they ushered in an era of weaker color filters which led to duller, more-muted colors in images. This ultimately became known as the infamous “CMOS-look” which to a large extent is still with us today.

This is not to say CMOS sensors can’t be made with strong CFA’s. It’s just that typically they are not. The pressure to increase pixel density and reach ever higher ISO’s keeps CMOS CFA’s weak and color performance middling. But because of this historical color gap there continues to be a size-able fan base for the older CCD technology. Sure, CCD falls apart at elevated ISO, but at base ISO many CCD cameras still produce colors that most CMOS cameras struggle to match.

Jim Kasson worked as a color scientist in the 1990s. Eric Fossum is not a color scientist, but he founded and ran a sensor company. I would expect him to know a thing or two about CFA selection. Meanwhile, you use promotional material to support your claims. Advertising is never the best source.

Your analysis of your 5D vs 5D2 post is pure fantasy.

I did read what The_Suede said. Where does he say anything that supports that the CFA was made "weaker" (whatever that means) for better ISO performance? As I said, he states the 5D2's CFA can cause more noise. How does that benefit higher ISOs?

And any data that opposes your view you immediately discount as bogus:
https://www.dpreview.com/forums/post/60668771

Link | Posted on Mar 23, 2020 at 00:27 UTC
In reply to:

fPrime: Although CMOS sensors enabled better high ISO performance for cameras, their occluding circuitry also initially made them less light sensitive than CCD sensors. As such, they ushered in an era of weaker color filters which led to duller, more-muted colors in images. This ultimately became known as the infamous “CMOS-look” which to a large extent is still with us today.

This is not to say CMOS sensors can’t be made with strong CFA’s. It’s just that typically they are not. The pressure to increase pixel density and reach ever higher ISO’s keeps CMOS CFA’s weak and color performance middling. But because of this historical color gap there continues to be a size-able fan base for the older CCD technology. Sure, CCD falls apart at elevated ISO, but at base ISO many CCD cameras still produce colors that most CMOS cameras struggle to match.

@fPrime:
1. I did not say Dr. Fossum was a retired color scientist. I was referring to Jim Kassen (who, granted, was an electrical engineer, though he did work in color science for IBM).

2: You did not answer all of his questions. You picked a graphic comparing the 5D and 5D2 spectral responses. Hardly proves a whole trend. You also shared a quote from The_Suede but seemed to not get it, again ignoring all he said that contradicted what you claim (for instance, the "weaker" CFA in the 5D2, as you put it, can actually lead to MORE noise, and so isn't ideal for low light). In fact, he never mentions improved ISO performance, suggesting the difference in CFAs is mainly for pleasing skin tones under various lighting conditions.

Third, I don't think anyone other than you would claim Eric folded. I think he felt it was useless to discuss the matter further with you, as you do not listen to reason and ignore any evidence that does not agree with you.

Link | Posted on Mar 22, 2020 at 22:49 UTC
In reply to:

fPrime: Although CMOS sensors enabled better high ISO performance for cameras, their occluding circuitry also initially made them less light sensitive than CCD sensors. As such, they ushered in an era of weaker color filters which led to duller, more-muted colors in images. This ultimately became known as the infamous “CMOS-look” which to a large extent is still with us today.

This is not to say CMOS sensors can’t be made with strong CFA’s. It’s just that typically they are not. The pressure to increase pixel density and reach ever higher ISO’s keeps CMOS CFA’s weak and color performance middling. But because of this historical color gap there continues to be a size-able fan base for the older CCD technology. Sure, CCD falls apart at elevated ISO, but at base ISO many CCD cameras still produce colors that most CMOS cameras struggle to match.

I think fPrime's still got a little animosity towards Dr. Fossum from when he called his "weak CFA" theory dumb (in a thread where a retired color scientist was unsuccessfully trying to get fPrime to clearly state his position):
https://www.dpreview.com/forums/post/60421216

Link | Posted on Mar 22, 2020 at 12:02 UTC
In reply to:

fPrime: Although CMOS sensors enabled better high ISO performance for cameras, their occluding circuitry also initially made them less light sensitive than CCD sensors. As such, they ushered in an era of weaker color filters which led to duller, more-muted colors in images. This ultimately became known as the infamous “CMOS-look” which to a large extent is still with us today.

This is not to say CMOS sensors can’t be made with strong CFA’s. It’s just that typically they are not. The pressure to increase pixel density and reach ever higher ISO’s keeps CMOS CFA’s weak and color performance middling. But because of this historical color gap there continues to be a size-able fan base for the older CCD technology. Sure, CCD falls apart at elevated ISO, but at base ISO many CCD cameras still produce colors that most CMOS cameras struggle to match.

As usual, fPrime, you're cherry-picking anything you think supports your view and ignoring all other evidence.

The author states in his final thoughts:
"For me, and I imagine for many others who couldn’t tell a definitive difference between either the head-to-head match-ups in Part 1 or the individual shots in Part 2, the results of the experiment are fairly clear. To restate: the hypothesis being tested was to see if the CCD look is real, unmistakable and couldn’t be emulated in post processing. To this end, I think I have at least demonstrated that with just a small amount of global adjustments in Lightroom, M240 files could make for some convincing M9 shots."

Furthermore, people only correctly identified the CCD image 36% of the time, which is worse than just a complete 50/50 guess. If there were any truth to your "CCD colors" theory, that would not be the case.

Link | Posted on Mar 21, 2020 at 20:11 UTC
On article Why Leica's M10 Monochrom is more than just a gimmick (666 comments in total)
In reply to:

(unknown member): Nice. Expensive.

But I really don't want to use color filters in front of the lens (again) to adjust the conversion of the different colors in grayed tones.

entoman, how is it any better than using a camera with a Bayer filter sensor at ISO 100? In other words, how does reducing the amount of light reaching the pixels not negate the advantage of having more light on the pixels?

Link | Posted on Jan 18, 2020 at 03:25 UTC
On article Why Leica's M10 Monochrom is more than just a gimmick (666 comments in total)
In reply to:

(unknown member): Nice. Expensive.

But I really don't want to use color filters in front of the lens (again) to adjust the conversion of the different colors in grayed tones.

Not to mention that using color filters reduces the additional light a monochrome sensor receives, negating one of its main advantages.

Link | Posted on Jan 17, 2020 at 15:55 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

Drogers23: In DPReview's professional opinion, does Canon have any aps-c native glass that can resolve 32 MP? If only a select few lenses can resolve that detail, this seems like it should be a huge topic of discussion.

Every lens should benefit from the 32MP.

Link | Posted on Oct 16, 2019 at 18:19 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

(unknown member): Increase in resolution is always a detail advantage, no matter what (canon should have removed the AA filter to really see the advantage though).

Is the 5DR supposed to look better in the studio shot? Because it doesn't.

Link | Posted on Oct 15, 2019 at 00:18 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

(unknown member): Increase in resolution is always a detail advantage, no matter what (canon should have removed the AA filter to really see the advantage though).

Yes, I'd rather have to apply sharpening to an image than try to remove moiré from it!

Link | Posted on Oct 14, 2019 at 18:48 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

(unknown member): Increase in resolution is always a detail advantage, no matter what (canon should have removed the AA filter to really see the advantage though).

Even with the filter, there is aliasing. It'd only be worse without.

Link | Posted on Oct 14, 2019 at 18:29 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

marc petzold: Funny thing, so...only because Canon, a being so-called "Marketing-Leader", and excuse my language - who gives a damn about this, does have the 32.5 MP Sensor into the M6 II, EOS 90D, and perhaps the next 20-25 Cameras following with this Sensor, the next couple years, like their old 18 MP Sensor debuted into the EOS 550D (2009), and it is still being used today, 10 years later - 2019 inside the EOS 4000D, the Timeline of 24 MP Sensors is ended?

I'd say Nonsense - Sony does sell 20 MP 1 inch Sensors since 2012 - the RX100, and RX10 Series since then.

Show me the APS-C/EF-S STM/NanoUSM Lenses, which *can* resolve full 32.5 MP - the answer is - bar none! None of the non-L Lenses can resolve that much MP. The new 18-135 USM is so-so.

Yikes, these folks - always lusting for more MP. Who is printing bigger than DiN-A3, A2 nowadays? I do order rarely 60x40cm prints, and being completely fine with 12 to 24 MP, mostly using my 12 to 16 MP Gear, which is from Megapixel Count quite enough.

Apparently you did not really look at the samples. 32MP is clearly not enough for the lens.

Look again:
https://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=canon_eos90d&attr13_1=canon_eos80d&attr13_2=canon_eos7dii&attr13_3=canon_eos800d&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=100&attr16_1=100&attr16_2=100&attr16_3=100&normalization=full&widget=1&x=-0.2036343014941565&y=-0.3037904050665845

Link | Posted on Oct 14, 2019 at 18:21 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

marc petzold: Funny thing, so...only because Canon, a being so-called "Marketing-Leader", and excuse my language - who gives a damn about this, does have the 32.5 MP Sensor into the M6 II, EOS 90D, and perhaps the next 20-25 Cameras following with this Sensor, the next couple years, like their old 18 MP Sensor debuted into the EOS 550D (2009), and it is still being used today, 10 years later - 2019 inside the EOS 4000D, the Timeline of 24 MP Sensors is ended?

I'd say Nonsense - Sony does sell 20 MP 1 inch Sensors since 2012 - the RX100, and RX10 Series since then.

Show me the APS-C/EF-S STM/NanoUSM Lenses, which *can* resolve full 32.5 MP - the answer is - bar none! None of the non-L Lenses can resolve that much MP. The new 18-135 USM is so-so.

Yikes, these folks - always lusting for more MP. Who is printing bigger than DiN-A3, A2 nowadays? I do order rarely 60x40cm prints, and being completely fine with 12 to 24 MP, mostly using my 12 to 16 MP Gear, which is from Megapixel Count quite enough.

Check out the 90D studio sample. Aliasing is present, indicating that the lens is indeed out-resolving the 32MP.

Link | Posted on Oct 14, 2019 at 18:07 UTC
On article Canon's 32MP chip marks the end of the 24MP APS-C era (493 comments in total)
In reply to:

(unknown member): Increase in resolution is always a detail advantage, no matter what (canon should have removed the AA filter to really see the advantage though).

No, they shouldn't have; aliasing is still an issue at 32MP.

Link | Posted on Oct 14, 2019 at 18:02 UTC
On article Canon EOS 90D Review (973 comments in total)
In reply to:

IamToddNorris: Why's everyone so bitter this camera didn't get an award? Should it get an award for the last aps-c dslr from canon? You think canon's going to start popping out new ef-s lenses to resolve that 32 mp sensor? The only decent, fast and wide ef-s lens is the 17-55 f2.8 and that was released in 2006 with the 30d. An 8.2 mp sensor camera. If canon releases a new aps-c lens it'll be for their "m" line not "ef-s". The lucrative "rf" mount will get resource priority.

(cont..)

Those RF lenses are also designed for full frame bodies, not APS-C.

Link | Posted on Oct 4, 2019 at 02:27 UTC
On article Canon EOS 90D Review (973 comments in total)
In reply to:

IamToddNorris: Why's everyone so bitter this camera didn't get an award? Should it get an award for the last aps-c dslr from canon? You think canon's going to start popping out new ef-s lenses to resolve that 32 mp sensor? The only decent, fast and wide ef-s lens is the 17-55 f2.8 and that was released in 2006 with the 30d. An 8.2 mp sensor camera. If canon releases a new aps-c lens it'll be for their "m" line not "ef-s". The lucrative "rf" mount will get resource priority.

(cont..)

There's no WYSIWYG if you're planning on processing from raws yourself or using flash, so an EVF doesn't have much to offer me (nor do histograms that are based on processed JPEGs).

Link | Posted on Oct 3, 2019 at 23:54 UTC

I don't find the "faked bokeh" to realistic or smooth. Looks like the same awful attempts they've been putting out for years(?) now; sharp patches that should be blurry, blurry patches that should be sharp, and ugly transitions between the two. Even at smaller sizes, it looks like a bad Photoshop job to me. Definitely more like a work in progress than a finished product.

But I suppose a lot of people must feel otherwise, since it's so ubiquitous these days.

Link | Posted on Oct 3, 2019 at 22:51 UTC as 22nd comment | 1 reply
On article Canon EOS 90D Review (973 comments in total)
In reply to:

Casch: I don't understand why Canon keeps upping the megapixels on an aps-c camera.
With a pixel size under 3.2 ums the lens becomes diffraction limited at f8 and unusable at f11 for print sizes of 20" or over. Sometimes, when you need a greater DOF and want to use f16 or greater the camera produces pictures that look out of focus and soft because of the lens diffraction or being diffraction limited. This isn't a problem with the lens (no matter what quality) but the limit of the sensor. That is why larger sensors are able to produce MUCH sharper images at smaller apertures (higher f-stop numbers). I would much rather have a sharp 20mp image than a soft 32mp image. Check out this site for a better explanation https://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
The megapixel wars are over. Just make a cleaner sensor.

Casch, that's not how diffraction works (Cambridge in Colour is a bad source for technical information).

Link | Posted on Sep 30, 2019 at 15:02 UTC
In reply to:

Lan: Thanks; that's just what I've always wanted, far heavier camera kit ;)

Better to make something like this water filled, that way unless you're shooting in a desert there's usually some weight to be added near your destination.

I would suggest a nice (lightweight) plastic bottle with velcro straps; yours for the price of an empty 2ltr bottle and some cheap velcro straps ;) Cheaper, a whole load lighter and probably about as useful.

I would think that if you were in the desert you'd be using sand bags.

Link | Posted on Sep 23, 2019 at 19:58 UTC
On article Canon EOS 90D sample gallery (DPReview TV) (69 comments in total)
In reply to:

whakapu: I really want to want this camera but with image quality like this I just can't. I know a lot of people are saying it's good but I just can't see that. Nothing is crisp at the pixel level. The guy at camerajabbers appears to see what I see. To my eye jpegs above 400 ISO are smeared like phone images at pixel level, and even base ISO isn't that great. I'm sure if the images were downsampled to, say, 24 MP they'd look better but then why not buy a camera that gives that sooc?

"With a good lens a 100% crop can look decent."

Which means that the good lens is being undersampled. The more resolution you get, the worse 100% crops are going to look, because you're looking at a smaller and smaller sampling of the lens.

Judging things by 100% crops is a meaningless endeavor.

Link | Posted on Sep 22, 2019 at 10:54 UTC
On article Canon EOS 90D sample gallery (DPReview TV) (69 comments in total)
In reply to:

whakapu: I really want to want this camera but with image quality like this I just can't. I know a lot of people are saying it's good but I just can't see that. Nothing is crisp at the pixel level. The guy at camerajabbers appears to see what I see. To my eye jpegs above 400 ISO are smeared like phone images at pixel level, and even base ISO isn't that great. I'm sure if the images were downsampled to, say, 24 MP they'd look better but then why not buy a camera that gives that sooc?

Things should not look sharp at the pixel level. "Sharp" pixels means the lens is being undersampled.

Link | Posted on Sep 22, 2019 at 03:17 UTC
Total: 84, showing: 1 – 20
« First‹ Previous12345Next ›Last »