Low light performance of various cameras

| | Comments (1)

Low light performance is important for a lot of street photography. Being able to meter down to EV 1 @ ISO 100 is often a necessity. A very dark bar or cafe might be EV 0. Some manual meters and the Leica M6/M7 are able to meter down to EV -2 which is just about the limit of "available darkness."

Bessa R2/3A: EV1 @ ISO 100 @ f/1.4
Bessa R2/R/T: EV1 @ ISO 100 @ f/1.4
Leica M6/M7: EV-2 @ ISO100 @ f/1
Hexar RF: EV1 @ ISO100 @ f/2
Leica CL: EV3 @ ISO 100
Minolta CLE: EV 3 @ ISO100 f/2

Canon EOS 3: EV 0 @ ISO 100 @ f/1.4
Canon EOS10D/20D: EV 1 @ ISO 100 @ f/1.4
Nikon F2 with DP-12AS: EV -2 @ ISO 100
Hasselblad 203FE: EV 0 @ ISO 100 @ f/2.8
Hasselblad 205FCC: EV 1 @ ISO 100 @ f/2.8

VC Meter 2: EV1 @ ISO 100
Gossen Digisix: EV0 @ ISO 100
Minolta Autometer IVF: EV -2 @ ISO 100 (incident)

Now EV 1 @ ISO 100 sounds pretty low (it's 1 second @ f/1), but using ISO3200 film (which is what you'd use in a really dark place), it's 1/30th second @ f/1.4. Many of us with f/1.4 lenses, rangefinders, steady hands, or monopods/tripods can shoot lower than this. Also, with TTL metering, the EV numbers are reduced by the aperture of the attached lens. So a camera meter is rated at EV 0 @ ISO 100 with an f/1.4 prime lens is only EV4 @ ISO 100 with a f/5.6 zoom. The Leica M7 with an EV-2 @ f/1 is only an EV-1 @ f/1.4, still excellent but not as stunning.


Note that this is metering and not auto-focusing. Many autofocus systems will fail at EV 2 or lower especially with dark lenses such as most consumer zoon lenses. Also, film suffers from reciprocity failure and thus the metering may not be accurate at low-light levels with long shutter speeds.


6 Comments transferred from Blogger.com:

Luis said...

That's a quite interesting question.Recently I have been taking some images of a baroque organ on the church's upper floor , light conditions were comparable to a dark jazz club. Took a Leica M6 with HP5+ pushed up to 1600, a Olympus C-5050 and an old selenium cell meter (sekonic auto-lumi 86).The meter, as it was expected, was unable to read anything except close to the small window, which isn't too useful. The C-5050 (it has a very accurate meter) in puntual mode cannot give a clue reading over the wood tones close to "that gray" at ISO 64 (dont ask, I dont like the rest of ISOs available there). The M6 at least was able to give some clues metering in lights and shadows shooting almost always with an f2 aperture and between 1/4 and 1/15 which is relatively safe handheld for me. That can be estimated (based on http://www.geocities.com/thombell/1600chart.html) about EV 2 or 3 metering on lights, around 0 in the dark places. A faster lens (1-1.4) and 3200 film should make possible shooting handheld to the film very limits, if the motive does not move, indeed :)

11:24 PM

nasukaren said...

It does add an interesting twist to the digital vs. film debate. The Canon 20D can only meter down to EV 1 @ ISO 100 @ f/1.4. In the dark church environment you're describing, the meter will start to fail before the camera sensor does. This is definitely an area where Canon and other digital camera manufacturers need to pay more attention as they bump up the available ISO speeds.

11:36 PM

cjs said...

Interesting. I do mostly low-light, non-flash photography (my most common film is Fuji Super Presto rated at 3200), and I hadn't thought about the metering because it's just never been an issue for me, even when using 0.25-1 second shutter speeds at f1.4. I guess that's just another joy of the Olympus OM-4 for low-light photography. (The brouchure claims that the light meter range is -6.5 to 19 EV, ISO 100 @ 50mm f1.2. I'm not sure I buy that, though, but given the metering it's certainly testable, and I'm going to try it out.) I'm also going to start playing around with rating Super Presto at 6400, since I've done some 8 x 10 prints of stuff taken at 3200, and the grain is not really objectionable at all--certainly much better than what you'd see in typical photojournalism of the 1940s and 1950s.

11:05 AM

Luis said...

Imho grainness isnt the major drawback that the digital wave is claiming. Most times that pretty high ISOs are used the final quality in a giant print isnt the main issue.Certainly less grain/noise is better as rule of thumb, but that does not invalidate the main rule : better one shot that nothing. BTW 35mm or digital equivalent are designed for different uses than maximum print quality at 16x20 or bigger (altought they are perfectly possible), if I want other things I will take another equipment more suited for the case, no matter what Canon claims :PPlease, let us know the results of your tests,

7:52 PM

nasukaren said...

Dr. Jean Louchet writes:Hi Karen,I have read your very interesting page on lightmeter performance in lowlight. However I think there is a slight inaccuracy there. Measuring the limits at ISO 100 does not represent the low light performance in a relevant way, as there may be a real and an apparent limit given by the display system.On the other hand it is unfair to compare TTL cell cameras with differentlenses, e.g. the M7 with a f:1 Nocti and a CLE with a f:2 rokkor, as bothcameras do accept the same lenses (as you point out). Here are my own measurements - keeping in mind that in very low light it makes (not always but frequently) more sense to use ISO 1600 than ISO 100:Olympus PenF # ISO 100: 1/4 s at full aperture, this is EV3 w/f:1.4 lens.Nikon F/FTn @ ISO 1600: 1/15 s at full aperture (whatever the full aperture is). This is EV5 w/ f:1.4 lens.Nikon F/FTn @ ISO 100: 1 s at full aperture. This is EV1 w/ F:1.4 lens.Nikon F3 @ ISO 1600: 8 s at full aperture. This is EV -2 w/ f:1.4 lens. @ ISO 100: 8 s at full aperture. This is EV -2 w/ f:1.4 lens. (note that here the limit is that of the speed knob, and that auto exposure being able to give much longer exp times. Setting the ISO at 1600 the auto shutter gives 15 sec exposures say EV -3; at 100 ASA it gets up to 2 mins i.e. EV -6 with f:1.4 lens.M6classic @ ISO 1600: 1/8 s at any aperture. This is EV4 w/ f:1.4 lens.(I hope my M6 is not defective, at slower speeds the right diode lits butthe left one never does).M6classic @ ISO 100: 1 s at any aperture. This is EV1 w/ 1.4 lens. (note that at ISO 100 the cell is switched off by the speed knob getting to B, rather than by the intrinsic cell limit).Minolta CLE @ ISO 1600: 1/2 s at any aperture, this is EV2 w/ f:1.4 lens. CLE @ ISO 100: 1/2 s at any aperture, this is EV2 w/ f:1.4 lens. (note that with the CLE the limit is given by the diodes that stop at1/2 s, but the auto shutter's actual limit is around 30 secondsi.e. EV -4: thus the CLE might well be the best low-light sensitiveM-camera! - would have to compare with the Bessas anyway)vcmeter1 @ ISO 1600 : 1 s @ f: 1.4, this is EV1 @ ISO 100: 1 s @ f: 1.4 this is EV1 but again this is unfair asthe limit is due to the dial markings! Counting the clicks after the 1 secmarking allows up to 15 secs i.e. EV -3.Lunasix3 @ ISO 1600: 2 s @ f:1.4 (very slow reaction!), this is EV0 @ ISO 100: 30 seconds @ f:1.4, this is EV -4. In practicethe Luna3 is really usable from 1 stop over these values otherwise it has too much memory.To summarise: I believe that the EV limits at ISO 100 do not give adequateinformation on the actual low light metering limits. EV at high ISOsettings give a more accurate and useful idea. Using your criteria, the M6would be more low-light sensitive than the CLE, in reality it is theother way round.This table summarises the EV values I have measured @ ISO 100 and 1600,with a f:1.4 lens if applicable: ISO 100 ISO 1600Olympus PenF 3 NAM6classic 1 4CLE display 2 2CLE shutter -4 ? (have to test)NikonF 1 5NikonF3 display -2 -2NikonF3 shutter -6 -3Luna3 slow -4 0Luna3 practical -3 1VCmeter1 dial 1 1 VCm1 off-scale -3 1 > Leica M6/M7: EV-2 @ ISO100 @ f/1so would be EV -1 @ ISO 100 @ f/1.4, as metering is TTLwhich would mean the M6TTL is so much better than the classic?This is possible with the M7, but - I am not familiar with the M6TTL,does the M6TTL really have a 4 secs position on the speed dial?> Minolta CLE: EV 3 @ ISO100 f/2Yes but this is the number of diodes, not the actual sensitivity.>>> Lightmeters> VC Meter 2: EV1 @ ISO 100Yes. Again, congratulations for your outstanding web pages which are among my favourite readings on the web and so informative.Best,Jean------------------------------------------------------------ Dr Jean Louchet COMPLEX Project INRIA Rocquencourt http://fractales.inria.fr/~louchet

1:04 AM

nasukaren said...

I posted the above with the permission of Dr. Louchet. Note that in my defense, I took the specs from the manufacturers. Measurements of EV meter performance are done @ ISO 100 in order to standardize the specifications. Karen

1:05 AM

<< Home

Monthly Archives

Sponsored Links

Powered by Movable Type 5.11

Sponsored by



Sponsored Links

About this Entry

This page contains a single entry by Karen Nakamura published on December 30, 2004 5:40 PM.

New Year's events in Japan was the previous entry in this blog.

Nikon F6: a film camera in a digital age is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

September 2013

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30