Low Light Shooters

[Gavjenks]

Apparently, Gavjenks thinks that signal to noise is the ONLY valid criteria. The hell with dynamic range. The hell with color. Right?

Also, no, not at all.

Of course those are important. But like any multi-factorial data, it is always better to be able to view each factor independently, if possible. You want to be able to pull up a dynamic range graph, and then SEPARATELY a color IQ graph, and SEPARATELY a noise graph.

Slap dashing them all together in multiple different (and redundant) arbitrary mixtures is stupid. That makes it vastly more difficult to meaningfully interpret, and it's never ever how you properly report experimental data like this if you can at all help it (i.e. it's okay if you CAN'T separate them out, and are forced to show a combination. Unacceptable if you have separated them, which they have). And anybody who does so would normally be seriously calling into question their analytical competency.

In the case of DxO, I think it's a little more excusable that they have such summary data / I am not writing them off as slackwits, because of the obvious marketing aspect involved, which is probably why they do it. I.e. newbie shooters are overwhelmed with stats and need a single number, or they will go elsewhere. So maybe they sort of need to do that to stay in business, okay fine. But it is absolutely done IN SPITE of scientific appropriateness, not in service of it. And so anybody who actually know what they're doing and what their shooting style is and how to read a graph should look up the graph.





Saying that you actively should look at the summary data and not the raw graphs is like telling people they should judge nutritional value based purely on the big comic sans font "now with 70% less fat!" on the front of packages, instead of the nutrition labels...

Both "70% less fat!" and "Summary sports/low light score!" are designed for distractable people or people who don't really know what they are looking for just yet. Which is fine, I was once one of those people too when I started out. But once you know better, you can do better.

 

[Derrel]

I did a search for "the low light king", and lo and behold, an article written by Mason Resnick, the Managing Editor of Popular Photography magazine popped up near the top of the list.

20 Best Cameras For Digital Low-Light Photography: Tested and Compared! from Adorama Learning Center

The article was published January 20, 2014, so it's current. It is entitled 20 Best Cameras For Digital Low-Light Photography: Tested and Compared!


Winners, by sensor size: FULL FRAME: Nikon Df, Nikon D800/Nikon D800e, Nikon D610. Runner up: Sony A7r.

APS-C sensor size winners: Nikon D5300, Nikon D7100, Pentax K3.

m4/3 sensor winners: Olympus Pen E-P5, Olympus Pen E-PL5, Panasonicx Lumix GH3.

I have omitted the small-sensor digicam winners, and some of the runners-up. There is not a single Canon-sensored camera in the whole bunch.

 

[Gavjenks]

Derrel, that author simply looked up various scores on DxO and found pictures of the results (after filtering it through some additional unspecified formula of the author's own design to make it less valid and more confusing) and "wrote" an "article" about it...

Okay, fine, DxO is nice and all, yay. But this doesn't provide any additional evidence or information beyond the same thing you already posted earlier, since it comes directly from the exact same source.




If people want actual convergent evidence and agreement or disagreement from more than one source, they need to look somewhere that isn't simply based on DxO. For example, one of the various sites that shows you actual photos to compare yourself, like this one:

http://www.dpreview.com/reviews/studiocompare.asp

 

[Derrel]

Now, kind of going back to what runnah suggested...how about a camera and a flash unit? Here's a fellow who made his whole career, shooting MOSTLY after the sun had gone down...wayyyyy back in the day before autofocus. Of course, his camera probably had a Kalart rangefinder, and heck I dunno, maybe he even used a flashlight. Suffice it to say, his generation was no lucky enough to have a new-fangled d-slr with an in-body or in-flash AF assist beam to be able to NAIL focus in pitch-black conditions.

This is low-light shooting, OLD-school style. picture of Weegee with camera and flash - Google Search

and a second link to "Weegee"

http://lens.blogs.nytimes.com/2011/12/16/weegees-other-naked-city/?_php=true&_type=blogs&_r=0

 

[sk66]

The best "low light camera" is going to be the one of current technology with the largest pixels/photoreceptors.
Almost all of the penalties of low light are due to not saturating the sensor with light; and it is caused by using higher ISO's. ISO is *not* actually part of the captured exposure...it's the equivalent of "pushing/pulling" film during processing.
The "low ISO noise" is a matter of current technology/processing and is to make up for the inability to saturate the sensor... the processing often includes a certain amount of noise reduction even for a raw file.
So in this context a faster lens is better than high ISO performance, but it also has it's compromises (i.e. DOF/sharpness).

 

[runnah]

Thanks for the nod Derrel.

Truth is that even the best rated camera still sucks at really low light. I love how camera maker s brag about super high ISO but in reality most suck after 6000-8000.

There comes a very clear point where you need to add light or use a tripod.

 

[Gavjenks]

Better Light SuperModels with Enhanced Resolution

This digital scanning back gives 48 megapixels, each of which is about 4x larger as the sensors in the largest mass produced large format digital cameras in Nikon/Canon today.
It only costs a measly $15,000!!