fPrime

fPrime

Lives in United States CA, United States
Joined on Jul 17, 2012

Comments

Total: 76, showing: 1 – 20
« First‹ Previous1234Next ›Last »

The LX3 was my second digital camera and I recall shooting everything at 16:9 on it. It was a marvel of miniaturization for its time but the desire to have a wide, fast lens on it clearly forced Panasonic to make some design compromises.

For example, soon after launch it was discovered that the LX3 lens produced images with significant optical distortion which required default correction post-capture. But even though the geometric distortion could be fixed, lots of micro-color fringes from the complex lens design appeared whenever the camera was shot wide open.

I also remember Panasonic's default JPEG colors were also a source of frustration for many back in the day. Fortunately the camera had RAW to get around that problem. I eventually found that C1 Pro had the most accurate color profile for the LX3 after first comparing LR, DxO, and Silkypix. C1 also had better correction for the LX3’s notorious color fringing and CA. Obviously there had been an LX3 enthusiast at Phase One!

Link | Posted on Mar 23, 2017 at 16:13 UTC as 52nd comment
In reply to:

Infared: WOW!....GREAT article!..
I had already come to a lot of those conclusions...thanks DPR for confirming them.

I am really looking forward to this:
"But remember you can emulate much of that color science in Raw converters with proper profiles (we're looking into a separate article on this)"
That is an article that I want to read!!!

@Infared - Don't hold your breath, DPR have no secret sauce here. They've fully bought into the theory that profiling small, high-density sensors can compensate for the poor color from their weakened CFA's.

What the experts won't tell you is that color profiles are wholly illuminant dependent... a profile built in sunlight breaks down under artificial light or even shade. Secondly, calibration targets are limited (24 samples typically) and the harder the profile has to pull the reference colors into line, the more likely that surrounding hues are shifted unpredictably. Not that you'll know until you shoot.

The only scenario worth profiling for is studio work where the lighting/lens/camera are completely static. For all other purposes the default profiles will suffice. Even better, use a camera built before 2008 to take advantage of it's stronger CFA. Today's sensors generally have weak CFA's. A strong CFA, however, is able to render color more accurately under artificial illuminants.

Link | Posted on Mar 21, 2017 at 15:59 UTC
In reply to:

helltormentor: This is what I am looking forward to :
"But remember you can emulate much of that color science in Raw converters with proper profiles (we're looking into a separate article on this)."
I hope you have found an easy way to accurately profile digital cameras. I have tried all methods proposed by experts but never succeeded in getting accurate color and had to switch back to the default profiles existing in RAW converters.

Cheers

@helltormentor - Welcome to the club! Don't hold your breath, DPR have no secret sauce here. I can tell from their text that they've fully bought into the theory that profiling high density cameras can compensate for their weakened CFA's.

What they don't tell you is that color profiles are wholly illuminant dependent... a profile built in sunlight breaks down under artificial light or even shade. Secondly, calibration targets are limited (24 samples typically) and the harder the profile has to pull the reference colors into line, the more likely that surrounding hues are shifted unpredictably.

The only scenario worth profiling for is studio work where the lighting/lens/camera are completely static. For all other purposes the default profiles will suffice. Even better, use a camera built before 2008 to take advantage of it's stronger CFA. Today's sensors generally have weak CFA's. A strong CFA, however, is able to render color more accurately under artificial illuminants.

Link | Posted on Mar 21, 2017 at 15:42 UTC
In reply to:

beatboxa: He's got a great attitude & accomplishments, but there were a few areas that I think were missing in this interview: specifically, reliability and image quality.

He seems to equate things like 'corner sharpness' with 'better', but there are plenty of IQ aspects where the Arts fall short for certain applications. Portraits is a great example: when contrast, bokeh, tones, etc. matter. And the Sigma Arts generally don't do well in these areas.

Ironically, the Sigma "Arts" are some of the most technical and least character lenses out there. They should rename them to the Sigma "Sciences". :)

And if you want to see some of flattest 135mm images in your life... look at these flat Sigma 135 Art samples just posted to Flickr:

https://www.flickr.com/photos/apmagazine/33110433366/in/album-72157678965478591#

The Buddha looks like a cardboard cutout.

Link | Posted on Feb 27, 2017 at 22:15 UTC
In reply to:

beatboxa: He's got a great attitude & accomplishments, but there were a few areas that I think were missing in this interview: specifically, reliability and image quality.

He seems to equate things like 'corner sharpness' with 'better', but there are plenty of IQ aspects where the Arts fall short for certain applications. Portraits is a great example: when contrast, bokeh, tones, etc. matter. And the Sigma Arts generally don't do well in these areas.

Ironically, the Sigma "Arts" are some of the most technical and least character lenses out there. They should rename them to the Sigma "Sciences". :)

@Rishi - Sigma ART lenses disappoint in microcontrast and tonality. They are flat lenses with clinically sharp rendering. Plenty of evidence of that was presented last year by Yannick Khong. http://yannickkhong.com/blog/2016/2/23/the-problem-with-modern-optics

@notpc - DPR doesn't test for 3D Pop in their lens reviews. Don't expect to see differences in their 2D comparator. If they actually tested, things would be different.

Link | Posted on Feb 27, 2017 at 19:17 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

@ HumanTarget

It's hardly overblown. Look at the member comments saying color declined in the transition to CMOS whenever Throwback Thursday involves CCD cameras. And why do you think there are several requests in these comments alone asking for DPR to add some old school CCD sensors to the new comparator? People on both sides of this debate are genuinely interested to know what color differences they might be able to see. If we get this done we'll have both RAW and JPEG examples to rule out camera JPEG engine improvements.

What's curious to me is why the editors have been so tone deaf to the comparison interest. At least Carey acknowledged the love among many here for both the SOOC JPEG's and RAW colors of older CCD cameras in this D80 Throwback. Progress, I hope.

Link | Posted on Feb 19, 2017 at 01:27 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

@HumanTarget

Although overall the same amount of light falls on sensors of equal size, the higher the megapixel count, the more that light is subdivided. Literally fewer photons arrive at each sensing photo-site the higher the density.

Now, the reason why sensor designers decided to give up some some color quality in modern sensors is simple... they prioritized high resolution and high ISO quality. So it was a scenario of two steps forward, one step back. You've fallen into the classic trap of believing that sensor development has no design tradeoffs, haha.

Link | Posted on Feb 17, 2017 at 23:04 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

zodiacfml: I have seen better photos from throwback compact cameras with CCD sensors compared to sample photos in the D80 review. Good thing is, I found it and it is there and evident on one sample photo. DPR should do a definitive image quality guide between the two sensors.
MF CCD sensors takes that IQ even higher. The MF CMOS sensors underwhelms.

@zodiacfml

The rationale is explained below:

Doug Peterson writes: "...the selection of a CFA, the color pattern put in front of the sensor, is a choice between quality of color, and ISO performance. If the CFA allows each pixel to see a broader spectrum of color (e.g. for the green pixels to see a bit further into yellow) a camera’s ISO range can be modestly increased. The resulting loss in color quality is subtle – subtle variations in color are missed and a handful of specific colors become difficult to photograph. In a market where a ISO 25,600 camera has a leg up on a ISO12,800 camera, the engineers are under enormous pressure to pick the modestly increased ISO over subtle color quality."

https://luminous-landscape.com/the-phase-one-iq250-cmos-fully-realized/

Link | Posted on Feb 17, 2017 at 20:48 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

@Marc - I'm with you. Also looking to buy more MF lenses and a better body to shoot them with. I might be getting a Sony unless Nikon come out with a mirrorless Df2 that is low density, has IBIS, EVF w/histogram, focus peaking, and strong CFA.

Despite the A7S offering a low density full frame option, my worry is that Sony may still have put a weak CFA on it simply for better high ISO performance. This insane push to make cameras into night vision devices is the ruination of base ISO color, I swear.

Link | Posted on Feb 17, 2017 at 19:10 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

@HumanTarget - Yes, simply because the pixels in old, low-density CMOS cameras had more surface area to gather light. Today's high-density CMOS pixels are much smaller and as a result gather less light in each pixel well.

Link | Posted on Feb 17, 2017 at 18:07 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

Josh Leavitt: Some of the 10+ year old cameras aren't half bad, even by today's standards. Sure they probably won't hold up well in low light, or produce decent files above 1600 ISO, or burst beyond 5 fps, but there's still a few gems out there. I just picked up a Canon 5D "Classic" and have been amazed by the color rendition.

The Canon 5D Classic had a low density sensor with big, fat pixels. As such it had a light sensitivity similar to crop CCD cameras and could use the old-school strong color filter arrays that they did.

Link | Posted on Feb 17, 2017 at 16:54 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

Will B Milner: Sorry to be THAT guy. "So, while the D80 could reach ISO 3200, it was so noisy that it up being a setting that I reserved for dire emergencies only." doesn't read very well.

Written over lunch while eating a sandwich as most of these Throwbacks seem to be. :D

Link | Posted on Feb 17, 2017 at 16:46 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

BlueBomberTurbo: You guys ever think about adding these cameras to the newest comparison tool? Maybr even the exposure latitude tool. It'll give everyone a good idea of how far digital cameras have come, and let people see the famed CCD colors.

DPR seemed to have gotten this message when they did the throwback on the Canon 5D Classic. It was a joy to see it added to the new comparator. But since then, nothing. Popular cameras like the D1X, D200, S5 Pro, and D700 warrant this extra effort as it allows many users to connect the performance of today's cameras with yesterday's gear.

Link | Posted on Feb 17, 2017 at 16:36 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

@Marc - The Canon 5Dc, D700, and D3/D3s were all low density CMOS sensors whose big fat pixels gave them similar light sensitivity to the original crop CCD sensors. As such they were still designed to use the stronger old-school color filters. These are the only CMOS sensors that ever came close to CCD's rich color rendering. Everything released since has color that's been digitally faked to some extent.

Link | Posted on Feb 17, 2017 at 16:20 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

MrBrightSide: Colors are SO much nicer than CMOS cameras, but can anyone explain why?

CCD sensors have greater light sensitivity which allows much stronger color filters to be used. Strong filters yield exquisite color fidelity.

After 2008 camera manufacturer's began to weaken their color filters in order to accommodate less sensitive CMOS sensors, boost high ISO performance, and enable higher resolution. Weak filters requite stronger digital gains to be applied to the RAW transforms yielding a less natural color rendition.

Link | Posted on Feb 17, 2017 at 16:12 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

zodiacfml: I have seen better photos from throwback compact cameras with CCD sensors compared to sample photos in the D80 review. Good thing is, I found it and it is there and evident on one sample photo. DPR should do a definitive image quality guide between the two sensors.
MF CCD sensors takes that IQ even higher. The MF CMOS sensors underwhelms.

CCD sensors are stunning for base ISO work. Their greater light sensitivity allowed much stronger color filters to be used yielding exquisite color fidelity. Only low density CMOS sensors with old school color filters ever came close to CCD's rich color rendering... this includes the Canon 5Dc, D700, and D3/D3s.

After 2008 camera manufacturer's began to compromise color fidelity in order to boost high ISO performance and enable high resolution sensors.

Link | Posted on Feb 17, 2017 at 15:55 UTC
On article Throwback Thursday: the Nikon D80 (243 comments in total)
In reply to:

martindpr: When are we gonna see the test scene shot by older cameras to see how they fare and the evolution of the pro line? I'm thinking of Nikon's pro bodies - first and second gen d1x, d2h, d2x, as well as canon's 1d, 1ds etc...

I agree completely, martindpr, but that would take real work to put together.

Sadly it's much easier to write up a few paragraphs of personal musings about an old camera over lunch and call it a day.

Link | Posted on Feb 17, 2017 at 15:39 UTC
On article Leica announces M10 with new sensor, slimmer design (113 comments in total)
In reply to:

fPrime: So sad... Leica users had signed up en masse calling for Leica to release an updated full frame CCD sensor with strict CFA:

https://prosophos.com/2013/11/16/an-open-letter-to-leica/

Instead they got a "me too" 24MP CMOS sensor with weak CFA that Nikon, Canon, and Sony have all either released or surpassed years ago.

What bold technical innovation, Leica! If you think that will carry a $4,000 premium... may God help you. I might have considered buying an M10 had you stuck with a uniquely positioned CCD, but now you're just another average CMOS sensor camera company.

@ Don Sata - CCD doesn't need further development per se as the entire sensor surface is already available for recording light from the lens. It is CMOS that still needs more development to reach equivalent light sensitivity as its amplification circuitry still occludes part of the light gathering surface in most full frame sensors.

That's why people keep harping on the industry to move every camera to backside illuminated CMOS sensors as the amplification circuitry is tucked entirely behind the pixel wells. If Leica had used a full frame BSI-CMOS sensor in combination with the same strict CFA used in the M9, you'd be surprised to hear that I'd be just as happy as I would be with a CCD sensor proper.

Link | Posted on Jan 20, 2017 at 20:20 UTC

Absolutely terrible colors in those M10 photos in the gallery. They either look artificially digital, thin, off color, or cartoonish. If this is what the M10 produces it will send Leica CCD users to Fuji who at least know how to build cameras in this this form factor with good color.

Link | Posted on Jan 18, 2017 at 22:33 UTC as 172nd comment | 1 reply
On article Leica announces M10 with new sensor, slimmer design (113 comments in total)

So sad... Leica users had signed up en masse calling for Leica to release an updated full frame CCD sensor with strict CFA:

https://prosophos.com/2013/11/16/an-open-letter-to-leica/

Instead they got a "me too" 24MP CMOS sensor with weak CFA that Nikon, Canon, and Sony have all either released or surpassed years ago.

What bold technical innovation, Leica! If you think that will carry a $4,000 premium... may God help you. I might have considered buying an M10 had you stuck with a uniquely positioned CCD, but now you're just another average CMOS sensor camera company.

Link | Posted on Jan 18, 2017 at 22:18 UTC as 19th comment | 4 replies
Total: 76, showing: 1 – 20
« First‹ Previous1234Next ›Last »