Is it just me or the lowish light files are too noisy?

[petergabriel]

No, we're talking specific conditions. In perfect light conditions, low-mid ISO they are excellent. It's when light becomes poor and you crank up the ISO is where it falls apart imo. now playing around with shadow, highlight, noise reduction etc settings can have a big impact on your SOOC pics so YMMV but for me it's been very hit&miss whereas i can get way more consistent SOOC pics out of my 5D3 with minimum fuss. I shoot everything manual btw.

 

What do you consider mid-iso?

[algrove]

What do you consider mid-iso?

 

Specific ISO numbers or range of ISO settings would be most helpful to all of us.

[graflex]

People keep posting images at iso 6400 in well lit situations. We are talking about the high iso performance in LOW lit situations. You get a completely different result then. Losing details, increased smudging, less saturation.

 

This is nonsensical. See my earlier post #39 this thread.

 

That bookcase is not well lit. It's in a dimly lit room with a single overhead ceiling light. The exposure was 1/8 sec at f/6.4. That's very LOW light. At ISO 200 that would have required an exposure of 4 seconds at f/6.4.

 

"Just because you've raised the ISO doesn't mean you don't get a good exposure at that ISO."

[kimcarsons]

I'm just going to toss the terms high-key and low-key out there for anyone struggling to find the right word.

 

A low-key, normal, and high-key exposure at ISO 6400 will look very different in terms of subjective noise---because we find noise most objectionable in dark tones and shadow areas (if you think about it, the shadows are under-exposed and the highlights are over-exposed, locally speaking.)

Edited January 7, 2017 by kimcarsons

[Adam Woodhouse]

@

Hermelin

 

That pic is too cute.

[Adam Woodhouse]

I recall when I got the Nikon D300s and read its higher ISO results were suppose to be better (best at the time for Nikon APS-C cameras).  I tested high ISO outside in fair lighting (resulted in high shutter speed) and sure enough I thought 'wow, really good'.  Then I used it at my next wedding job and discovered that the results of the same higher ISO in a dark environment showed drastic different results.  

 

It taught me that it is important to get exposure correct at the time of shooting in low light, darker scenes because any post edit to the high ISO file showed serious degradation.  Also showed me that dark scenes versus bright scenes (at the same high ISO ... which at that time was 3200) produced noticeably different results in the pixel quality of the image.

Edited January 17, 2017 by Adam Woodhouse

[Hermelin]

People keep posting images at iso 6400 in well lit situations. We are talking about the high iso performance in LOW lit situations. You get a completely different result then. Losing details, increased smudging, less saturation.

Here's ISO 10 000 in LOW lit situation. Ofcourse it's not as detailed as ISO 6400 @ a better lit situation but I still find it impressive and terrific for aps-c.

-3 in NR, no post editing. JPEG Straight out of cam. 

 

 

DSCF9672 by Filip Hermelin, on Flickr

[Torturro]

I know its a bitt off - but its related in my opinion: 

 

Here is 100% crop of my picture: river at night - somo palms against sky.

800 seconds, f8 ISO 200

 

Can someone tell me if one is getting such a horrible hot pixels or whatever it is? Those red blue dots - amount of those on my 6 months 9000 accutations XT2 is so horrible it is visible on my small monitor with picture filled in - and for sure will be visible at small size prints!

 

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!

[kimcarsons]

I know its a bitt off - but its related in my opinion: 

 

Here is 100% crop of my picture: river at night - somo palms against sky.

800 seconds, f8 ISO 200

 

Can someone tell me if one is getting such a horrible hot pixels or whatever it is? Those red blue dots - amount of those on my 6 months 9000 accutations XT2 is so horrible it is visible on my small monitor with picture filled in - and for sure will be visible at small size prints!

 

You're going to have this problem with any digital camera. You either need to use the built-in long exposure noise reduction function (assuming you're using the JPEGs) or shoot/apply your own dark frame images if you're processing the RAW files. If you really want the best results, shooting multiple shorter exposures and averaging them will produce better results than one long exposure of the same total duration.

Edited February 13, 2017 by kimcarsons

[jlmphotos]

I too found the X-T2 noise unacceptable and refused to use it in low light.  I kept using my X-T1.

 

I think you would see more noise in a 24mp camera than you would in a 16mp.  That's the magic of physics.  I hated shooting my 36 megapixel D800e at anything over 640ISO even though I KNEW that the extra noise, was just the massive amount of resolutions and pixels.  I believe this is the same situation we face with the difference in noise between the X-T1 and X-T2 (in my case)  If you downsample the T2 files, down to T1 size, it should clear up some of the noise, but then why bother getting the X-T2? Right?  

 

So, yes upon initial assessment in LR it does look noisier, but once you fix it a bit, and PRINT IT, the increase in resolution, and clarity make up for the noise and it pretty much disappears in print.  

[Torturro]

Im not talking about noise. Im talking about some randomly scattered high signal red/blue big dots.

I'd call it hot pixelss - except - I cant believe there is such amount of them  - so these are probably not hot pixels.

 

The noise is uniform pattern, regular.

Edited February 14, 2017 by Torturro

[jakku]

You can try capture black image ISO 200-6400 1/8000th, then import to LR and increase clarity or exp. You will find more of them in white color... For me it's strange :/ but I think it's sensor problem in all new Fujis...

[Torturro]

very very bad, again.

[kimcarsons]

Im not talking about noise. Im talking about some randomly scattered high signal red/blue big dots.

I'd call it hot pixelss - except - I cant believe there is such amount of them  - so these are probably not hot pixels.

 

The noise is uniform pattern, regular.

 

This is an 800 second exposure though, right? There's a reason why digital cameras make you jump through hoops to take exposures longer than 30 seconds... Thermal noise becomes a huge problem and even slightly 'warm' pixels will saturate at this duration. Any CMOS sensor without an active cooling solution will suffer in the same way (although hot pixels do look particularly ugly with the X-Trans pattern). You can totally eliminate the problem by changing your practices. Shoot a dark frame, limit your exposures to 30 seconds (use the intervalometer function to take 26 shots for a total of 800 seconds and change), and then subtract the dark frame and average the results. Then you will have a very clean image. Again, the problem and solution are common to all brands of camera.

Edited February 14, 2017 by kimcarsons

[Torturro]

Ok - so how to make a picture with running river being flat thanks to long time exposure - at dark?

[kimcarsons]

Ok - so how to make a picture with running river being flat thanks to long time exposure - at dark?

 

Exactly like I described. If you average the images (and don't have lengthy gaps in time between taking each consecutive shot), the motion blur will be the same, but the noise level (effective ISO) will be much, much lower.

[Torturro]

Would it be possible to put those images togeather in PShop? is there a tool for this? I tried some time ago with night sky bot stars didnt match ..

[kimcarsons]

Would it be possible to put those images togeather in PShop? is there a tool for this? I tried some time ago with night sky bot stars didnt match ..

 

Yeah you can do it in photoshop and with many other tools too. I think Tony Northrup has some tutorials on YouTube for image averaging in PS. Stars are their own problem---you may want a camera that can do astro-tracking for that (i.e. a Pentax) Pretty much all good looking star/milky way shots you see are composites, not single exposures. There are many more techniques involved than just managing noise.