The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

leica m9 iso 800 vs 640

gooomz

Member
i have been shooting my m9 a lot at iso 800 but after reading some of the threads here am i better off underexposing a bit and shoot at iso 640 and then just brighten in post?

if this is better why?


thank you.
 

Double Negative

Not Available
When shooting dark scenes or scenes with deep shadow areas, never (or try to avoid) general brightening. This is what brings out noise... You're always better off getting the exposure right in the first place - or if anything, darkening.

You can also hide noise sometimes by increasing contrast (or bumping down the dark areas anyway).

Obviously, lower ISOs are better for noise as well. I try to keep the ISO to 1250, max.

Nevertheless, proper exposure is perhaps the single most important thing.
 

douglasf13

New member
When shooting dark scenes or scenes with deep shadow areas, never (or try to avoid) general brightening. This is what brings out noise... You're always better off getting the exposure right in the first place - or if anything, darkening.

You can also hide noise sometimes by increasing contrast (or bumping down the dark areas anyway).

Obviously, lower ISOs are better for noise as well. I try to keep the ISO to 1250, max.

Nevertheless, proper exposure is perhaps the single most important thing.
This is not actually the case with the M9. The M9 is essentially an ISO-less camera. Pushing ISO 160 three stops in Lightroom 4 yields as good, if not better, results than using ISO 1250. Of course, the more light you get to the sensor, the better, but raising ISO in camera is not actually necessary, in terms of IQ. Granted, if you don't raise ISO in camera, you won't have a usable review image on the camera's LCD, so it may not always be practical.
 

douglasf13

New member
i have been shooting my m9 a lot at iso 800 but after reading some of the threads here am i better off underexposing a bit and shoot at iso 640 and then just brighten in post?

if this is better why?


thank you.
If you shoot, say, 1/60, f2 at ISO 800, and then shoot again at the same shutter and aperture, but use ISO 640 and then boost the exposure in Lightroom, the latter will be as good, if not better. The same is the case even if you keep the same aperture and shutter and use ISO 160 boosted in LR.

With the M9 raws, what matters is exposure, meaning shutter and aperture. ISO is just gain after the fact, and raw converter exposure gain works just as well as in-camera ISO with the M9, and it prevents you from blowing highlights.
 

Maggie O

Active member
This is not actually the case with the M9. The M9 is essentially an ISO-less camera. Pushing ISO 160 three stops in Lightroom 4 yields as good, if not better, results than using ISO 1250. Of course, the more light you get to the sensor, the better, but raising ISO in camera is not actually necessary, in terms of IQ. Granted, if you don't raise ISO in camera, you won't have a usable review image on the camera's LCD, so it may not always be practical.
Wow, I'm going to have to try this. How do you meter for this and keep speeds hand-held in low light?
 

douglasf13

New member
Wow, I'm going to have to try this. How do you meter for this and keep speeds hand-held in low light?
Hi, Maggie. It's so easy that it almost feels like cheating. You essentially meter like normal with ISO 160, and, whenever you're in light low enough that the meter shows you're underexposing, you can pick whichever aperture and shutter speed that you'd like, of course knowing that, the more light that hits the sensor, the less noisy your image will be.

The practical issues with this are that you often won't have a usable image review on the LCD screen, and you'll have to individually adjust the exposure of all of your shots in your raw converter.

I usually split the difference, and shoot ISO 160 whenever the light is good and ISO 640 whenever the light is low, and I push in the converter from there. Using 2 ISOs usually leaves me with a camera review image that is somewhat usable, even if is a little dark. It's essentially like shooting two different speeds of film and pushing in development.

It's important to remember that all digital sensors only have one true ISO. After that, we usually amplify the signal after exposure in one of a few ways, depending on the camera design:

1) raise the ISO in-camera, which adds analogue gain to the signal
2) raise the ISO in-camera, which adds digital gain to the signal (or some combination of analogue and digital gain)
3) raise the exposure in the raw converter, which adds digital gain to the signal.

Some cameras perform better by raising the in-camera gain (ISO,) like Canons, and others perform a little better by raising the raw converter gain, like the M9. It just depends on the sensor and A/D design. Some MFDBs don't actually raise in-camera gain at all when you raise ISO. Meta tags just tell the raw converter to boost the gain under the hood when you import the image. Tricky!
 

Double Negative

Not Available
Well, yes - of course. There's a sensor's native sensitivity - and then there's everything else. That is usually just an algorithm in the firmware saying "amplify the signal X times to represent ISO Y."

Unfortunately, that concept is wrapped up in a lot of things - the meter, for example. I'd just as soon select an ISO using the ISO button, have my LCD images, etc.

I get what you're saying though. Rather than let the camera apply its own algorithm - do it yourself via RAW/DNG conversion. The benefits are you can choose from multiple tools (ACR, Apple, Bibble, etc.) and exert more control (exposure, highlights, etc.).
 

douglasf13

New member
The beauty is that you rarely blow highlights with this technique, the noise is slightly lower, and you don't have to take the time to adjust the ISO. The big issue, like you mentioned, is that your LCD review may be dark. That's why I use a couple of ISOs, although I don't review shots in my LCD a lot.

Metering isn't really an issue. Once I get to the point where my largest aperture and slowest shutter speed (that I want to use) are under exposing, I know that every stop I underexposed is going to be adding a stop of noise in the converter. Metering is one of the easiest things about this technique. It's just like pushing film, but you're not stuck with pushing the whole roll the same amount. You can push each shot a different amount.
 

Amin

Active member
Well, yes - of course. There's a sensor's native sensitivity - and then there's everything else. That is usually just an algorithm in the firmware saying "amplify the signal X times to represent ISO Y."
Based on the DxOmark data, there are a variety of native ISO values for the M9. As represented at Sensorgen.info, sensor read noise seems to be very similar at all of them, so it makes little difference what ISO you use as long as you don't choose one which is high enough to clip highlights or low enough that your software can't brighten it enough.

Read noise is marginally lower at ISO 640 than at ISOs higher than that, so if you're looking to eek out a tiny bit more image quality for lowlight, handheld photography, I wouldn't go higher than ISO 640.
 

Godfrey

Well-known member
I'd been thinking I want to experiment with this ISO-less shooting style but it slipped my mind to give it a test since the last time we had this discussion. Thanks for bringing it up again.

I made a matched pair of exposures, set manually*to f/2.4 @ 1/45 second, at ISO 160 and ISO 1250. I brought the DNGs into LR, set the color temperature to 3400°K, tint to +1, and set the ISO 160 exposure to +3 EV exposure adjustment. I nipped them to 1:1 in the Compare view and did a screen grab, grabbed the whole frame from the Navigator view on the left so you could see the portion of the frame, and created this composite to show the result:


Far as I can tell, modulo my ability to hold still and focus properly, there's very little to no difference at all in the output image quality.

This implies two things to me:

- Lightroom's exposure control allows up to +/- 5 EV adjustment*range. Effectively, using an M9 this way nets up to ISO 5120 as possible rather than just ISO 2500.

- The key thing to be careful of is photosite saturation. Once saturated, there's no way to adjust the exposure downwards and pull back detail in the saturated area.

What it opens up is the ability to look to metering purely for that purpose: to prevent photosite saturation. Otherwise, set the aperture and exposure time that best suits your subject and put ALL of your exposure finesse into post processing, setting base and curve relationships.

Might as well turn off JPEG production entirely and operate the camera without any previews at all... Everything you're doing is going to be seen properly only after you bring the exposures into the image processing environment and bring the numbers up into the appropriate range.

Quite an intriguing way to work. I'll have to give it a go. Whether the processing in LR or other tools is always superior to how the camera elevates the effective ISO ... well, that remains to be seen, for me. It's worth some playing with.

What a curious and versatile camera the M9 is. ;-)
 

Amin

Active member
Nice demonstration, Godfrey. If the DxOmark data are correct, an ISO 640 image pushed two stops should look slightly better than an ISO 2560 image with matched exposure (shutter speed and aperture).
 

thrice

Active member
I did a lot of testing and pushing scenes with a lot of green/blue gets you plenty of banding. Otherwise neutrally lit scenes push very well.

The following are both ISO160 files pushed:



 

douglasf13

New member
Interesting, Daniel. I've not seen any banding disadvantages, yet. How far did you push these? Which raw converter? Also, have you tested your SD card to see if it causes banding, especially when shooting quickly in a series?
 

Amin

Active member
I did a lot of testing and pushing scenes with a lot of green/blue gets you plenty of banding. Otherwise neutrally lit scenes push very well.
As an aside, Nik Dfine does an amazing job of removing banding like that.
 

Double Negative

Not Available
As an aside, Nik Dfine does an amazing job of removing banding like that.
That's what I use on the rare occasion where the noise is fairly obvious and my other tricks haven't worked well. Though it's very simple. Noise Ninja has worked well for me in the past, but I haven't used it w/Leica images.
 

edwardkaraa

New member
I have been traveling the last couple of days so I missed this thread. I can confirm Douglas findings, since I have done many tests after he told me about it, and came to the same conclusions. It's curious that my M9 doesn't show the banding I see in Daniel's shot. I have pushed a few shots at iso 2500 by over a stop and couldn't notice any banding at all. I believe it is body specific. However, I think shooting compressed may exacerbate the problem. This said, even though I like to underexpose to preserve the highlights, I confess to be a "chimper" and therefore have to shoot at higher iso in order to see a proper image review.
 
Top