Q2 ISO 50 broken ?

[Jared]

Advertisement (gone after registration)

OK, now here is a 100% crop from the rear leg area of the right-hand piggy.  I have added back 0.7EV of exposure (as well as the 100% shadows adjustment already mentioned).  This will, of course, exaggerate the effect somewhat, but again it gives you a good idea of the limitations as you near the camera's noise floor.  This is the ISO 100 shot.  We don't generally think of ISO 100 shots as having any noise or grain, but if you push the files very hard they definitely do.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Jared]

And here is the ISO 50 shot.  Not that there is noticeably less shadow noise, but it's really only "noticeable" because I pushed the shadows pretty hard.  In practice, for most images and most print sizes you wouldn't notice the difference.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Jared]

And finally, here is what I did for the third shot.  I took twelve separate exposures on high speed drive mode, all handheld, and brought them into Lightroom.  I made the same adjustments as I did before except I reduced sharpening to zero, reduced color "de-noise" to zero, and I added back a full stop of exposure instead of 2/3 of a stop (since I took this one a full EV below the ISO 100 shot--meaning f/4 and 1/250s).  I then exported all twelve images to Photoshop as layers.  I had Photoshop auto-align the twelve layers.  I then deleted two of the images that were a little bit less sharp than the rest.  They weren't very different, but enough that I could tell at 200%.  

I then converted the twelve layers into a smart object and changed the combine method to "Mean".  This averaged all the layers together.  I brought the result back into Lightroom and added the same sharpening as I had used on the other two images (except without the color noise reduction since it wasn't required when averaging stacks of images.

Here is the result...

Note that the noise level is much lower than with the single exposures.  By stacking images, I effectively lowered the noise floor pretty significantly.  

I'm certainly not saying one would want to do this with every image.  Heck, this type of photography isn't really what the Q2 is for.  But it's still kind of cool, and the experiment does show you how much/how little you are losing or gaining by not using ISO 50.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[nicci78]

Thanks for you test shots.

However Leica should fix live view histogram at ISO 50. It is not normal. Seems like a software issue. Engineers got lazy and do not take into account the different dynamic characteristics at 50 ISO

Not sure to stack my images one day. 

Edited August 27, 2019 by nicci78

[Leica Guy]

13 hours ago, Jared said:

And finally, here is what I did for the third shot.  I took twelve separate exposures on high speed drive mode, all handheld, and brought them into Lightroom.  I made the same adjustments as I did before except I reduced sharpening to zero, reduced color "de-noise" to zero, and I added back a full stop of exposure instead of 2/3 of a stop (since I took this one a full EV below the ISO 100 shot--meaning f/4 and 1/250s).  I then exported all twelve images to Photoshop as layers.  I had Photoshop auto-align the twelve layers.  I then deleted two of the images that were a little bit less sharp than the rest.  They weren't very different, but enough that I could tell at 200%.  

I then converted the twelve layers into a smart object and changed the combine method to "Mean".  This averaged all the layers together.  I brought the result back into Lightroom and added the same sharpening as I had used on the other two images (except without the color noise reduction since it wasn't required when averaging stacks of images.

Here is the result...

Note that the noise level is much lower than with the single exposures.  By stacking images, I effectively lowered the noise floor pretty significantly.  

I'm certainly not saying one would want to do this with every image.  Heck, this type of photography isn't really what the Q2 is for.  But it's still kind of cool, and the experiment does show you how much/how little you are losing or gaining by not using ISO 50.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Jared,

Its hard with the resolution on these images to see much difference. I can see lower noise on your stacked image. 

Is the bottom line that ISO 50 is worth using for landscapes? That’s my take away. I would bracket if there was any concern about blowing highlights or wide dynamic range in the scene. I’m beginning to believe that under exposing maybe 1/2 to 2/3 stop might be a good idea to preserve highlights with the Q2  

I had an extreme situation recently where I was taking a portrait with sun behind the subject highlighting her hair. Even though I bracketed, I still ended up with highlights I could not recover.  ISO 100. I used a fill flash and that worked well, but the blown highlights ruined the shots  

Thanks for your efforts to discover the value of ISO 50. 

Dan

[Jared]

Oh, I agree it would be nice if the histogram at ISO 50 were as accurate as at ISO 100 and above. In the meanwhile, it’s not hard to work around.  

As far as stacking... I’m not proposing it per se, but it’s good to know the option is out there for those images that just plane require it.

Normally I wouldn’t bother for a traditional landscape.  Images with 13+ stops of dynamic range still generally like too heavy handed to my eye.  Maybe that’s as much my lack of skill as anything else.  But for situations where there isn’t enough light it can be very helpful.  Nightscapes virtually require the technique.  Also, low light shots where you don’t have a tripod allow you to get rid of high ISO noise.  It’s useful.

[Jared]

Advertisement (gone after registration)

8 minutes ago, Leica Guy said:

Jared,

Its hard with the resolution on these images to see much difference. I can see lower noise on your stacked image. 

Is the bottom line that ISO 50 is worth using for landscapes? That’s my take away. I would bracket if there was any concern about blowing highlights or wide dynamic range in the scene. I’m beginning to believe that under exposing maybe 1/2 to 2/3 stop might be a good idea to preserve highlights with the Q2  

I had an extreme situation recently where I was taking a portrait with sun behind the subject highlighting her hair. Even though I bracketed, I still ended up with highlights I could not recover.  ISO 100. I used a fill flash and that worked well, but the blown highlights ruined the shots  

Thanks for your efforts to discover the value of ISO 50. 

Dan

Bottom line is that ISO 50 has demonstrably more dynamic range as it does drop the noise floor.  It’s not simply a “pull”.  If you have the time and inclination in any given exposure, you can recover more shadows with ISO 50 even after you lower the exposure to recapture lost highlights.  The lost highlights are real, though, so you can’t simply use the camera the same way at ISO 50.  I’d probably rate it a “true” ISO 64 or thereabouts.

If you really want the most dynamic range possible out of the Q2 (or any camera) for static shots, stack.  That will get you more than the extra half stop or so you will get from ISO 50.  It’s especially useful for higher ISO images (where you have much more limited dynamic range) and for nightscapes.

Bracketing remains a good safety net.  There is only so much you can tell about highlights from the JPG histogram.  It’s more than adequate for an image with limited dynamic range, but for one that exceeds the camera’s capabilities it can save an image.

[bags27]

4 minutes ago, Jared said:

Bracketing remains a good safety net.  There is only so much you can tell about highlights from the JPG histogram.  It’s more than adequate for an image with limited dynamic range, but for one that exceeds the camera’s capabilities it can save an image.

Jared,  I'm unclear about your last sentence. The "It's" refers to bracketing or using the histogram? Ditto, the "it" 5 words from the end of the sentence. Thanks, Ken

[Jared]

10 hours ago, bags27 said:

Jared,  I'm unclear about your last sentence. The "It's" refers to bracketing or using the histogram? Ditto, the "it" 5 words from the end of the sentence. Thanks, Ken

Yes, vague pronoun references, sorry.  

The histogram is not very accurate for ISO 50 shots.  You may still have clipped highlights even when the histogram indicates you don’t.  One solution is to just be a bit more conservative—underexpose by an extra ⅓ stop or even ⅔ stop.  The other solution is to bracket.  You can then pick whichever exposure is farthest “to the right” without losing whatever highlight details you care about .  You want to preserve the blues in the brightest parts of the sky?  Yet keep as much shadow detail as possible?  Bracketing is a good approach.

[like glass]

What this chart taught me is to avoid shooting at 400.

 

source https://www.photonstophotos.net/Charts/RN_ADU.htm

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited December 3, 2022 by like glass