D850 noise

[JoPoV]

Since I have my D850 I would like to know if my it is optimum or if it has problems. Maybe focus problem ?
I seem to not been able to get what other get with it. Even if I've got great shoots with it.
What would be the way to make it tested ?
Should I ask some test from Nikon itself ? I am a french user, so maybe I can't have as easy way than usa users to know this ?
Any help would be welcome.

 

[NatureLover]

Thank you everyone for the discussion.
To answer some of the previous questions:

1. I don't use Auto WB
2. If I was shooting in low light conditions I would understand more noise being present
3. I do understand that cropping can make noise more visible and that noise tends to show up more in the shadows than highlights
4. I have never used the in-camera NR; I thought it would slow down your frame rate.

Since my original post I have obtained very stunning images with the D850; occasionally I am surprised at the perceived noise level when I am post-processing but things are improving.

I use DeNoise -- do users of DeNoise first remove noise in the RAW file and then post-process or do you post-process first, then remove noise? I do the latter.

 

[AlexY]

I use DeNoise -- do users of DeNoise first remove noise in the RAW file and then post-process or do you post-process first, then remove noise? I do the latter.

When I used DeNoise, I employed it early on in my post processing workflow - right after I adjusted whites and blacks and WB and it worked very well. It could be problematic if it was applied after sharpening and contrast/texture/clarity adjustments.
I now use DxO's Deep Prime, which I like better. I'm not sure when Deep Prime is applied to the Raw file in the Photolab processing, but the resulting image is very good.
Cheers, A.

 

[Turberticus]

I would not choose f/7.1 1/2500" for a still bird unless I forget to check settings while set for birds in flight.
ISO6400 is at the ragged end of what I would want to accept on my D850

Ouch. Yeah that will kill quality quickly. For a still bird, OP could have taken it at 1/250 and the ISO would have been ~640, even stopped down at F/7.1. Even 1/500 would have given an acceptably clean image at ISO 1250.

 

[NatureLover]

I use the Sigma 150-600mm C which is a heavy lens. I find it difficult to hand hold without introducing camera shake and hence why I typically shoot at 1/3x to 1/5x the focal length.

 

[Turberticus]

I use the Sigma 150-600mm C which is a heavy lens. I find it difficult to hand hold without introducing camera shake and hence why I typically shoot at 1/3x to 1/5x the focal length.

In my opinion, your technique is the biggest reason for your photos being so noisy. A lot of us are focusing on the gear side of things, specifically the D850's noise performance and noise reduction software, but if you want to keep your ISO down you'll need to use a slower SS. There's no reason to shoot still subjects at such high shutter speeds unless you're introducing a significant amount of camera shake. I use the 200-500 and am able to handhold it at 1/250 or slower with good results but I think I'm lucky to have good a handholding technique. Have you tried taking photos at slower shutter speeds with stabilization turned on while taking short bursts? The key here is to use short bursts. I find that when I shoot around 1/200 or even slower in low light, using short 4-5 frame bursts help to ensure that at least one of the photos is sharp. I'd lose out on a lot of photo opportunities if I just took one frame at a time. You also may want to invest in a lightweight monopod preferably with the Wimberly monogimbal head. That will add another stop or two of stabilization and will allow you to cut down on the noise in your final images. Anyway, those are just my thoughts and I hope this is helpful.

 

[NatureLover]

The key here is to use short bursts. I find that when I shoot around 1/200 or even slower in low light, using short 4-5 frame bursts help to ensure that at least one of the photos is sharp. I'd lose out on a lot of photo opportunities if I just took one frame at a time. You also may want to invest in a lightweight monopod preferably with the Wimberly monogimbal head. That will add another stop or two of stabilization and will allow you to cut down on the noise in your final images. Anyway, those are just my thoughts and I hope this is helpful.

I do use short bursts and have IS on. I guess I just need more practice with the Big Sig whilst hand-holding, lol.... or as you said, consider a monopod. I have a tripod with gimbal but find that unless I am going to remain stationary in a spot for sometime, it is cumbersome to carry.
Cheers

 

[Turberticus]

I do use short bursts and have IS on. I guess I just need more practice with the Big Sig whilst hand-holding, lol.... or as you said, consider a monopod. I have a tripod with gimbal but find that unless I am going to remain stationary in a spot for sometime, it is cumbersome to carry.
Cheers

That's a good start. I think practice will help a lot, as well as a good monopod AND noise reduction software.

 

[JoPoV]

I now use DxO's Deep Prime, which I like better. I'm not sure when Deep Prime is applied to the Raw file in the Photolab processing, but the resulting image is very good.

Denoising in photolab is aplied at demosaicing step. Before any step which could add noise. I find it very good too.
About all corrections related to optic modules are too aplied at that step, I think.

 

[nautiboy]

I use DeNoise -- do users of DeNoise first remove noise in the RAW file and then post-process or do you post-process first, then remove noise? I do the latter.

I was pretty excited when they added RAW mode to DeNoise. However what I’ve found is that while the denoise seems to be better in RAW mode, it doesn’t have a very good RAW interpreter and options are a lot more limited in Lightroom when working with Denoise’s resulting DNG files.

So in my workflow, I tend not to do the denoise on the RAW except in some cases. I typically do a few initial adjustments (mostly exposure and tone), send to denoise, then do the rest of my adjustments. Just my $0.02.

 

[Woodpecker]

To me the only fair real world comparison is comparing the same output resolution at the same size and viewing distance, so I like the 'print' dxo comparison. It's similar to diffraction in a way. You set a diffraction calculator to pixel for pixel and the larger pixels of the low megapixel cameras will win, but set it to print, which adjusts for viewing size at a given resolution and then the diffraction is not so bad on the higher megapixel cameras even with smaller pixels.

May be I didn't choose quite the right way to describe what I mean. I am by no means against the comparison the way you describe it, but which one to choose depends pretty much on what you are looking for.

If the focus is to produce prints of a certain size and quality to equip an exhition or you use images downsampled to a normalized resolutions for a slideshow with a beamer, the "print mode" comparison is definitely the correct way of doing it. This is because you compare the qualities of normalized images:
If, on the other hand, you deliberately want to compre on pixel level, e.g. because finding the limits up to where a heavy crop of a high res body can compete with a full frame image of a lower res body at the same resolution, it is a different story. Here we are talking about a step prior to compare the quality of images, i.e. pure sensor performance.

There is no black and white, right or wrong, ..., the truth is always somewhere in between and the "somewhere" depends on which question you ask .

 

[bleirer]

May be I didn't choose quite the right way to describe what I mean. I am by no means against the comparison the way you describe it, but which one to choose depends pretty much on what you are looking for.

If the focus is to produce prints of a certain size and quality to equip an exhition or you use images downsampled to a normalized resolutions for a slideshow with a beamer, the "print mode" comparison is definitely the correct way of doing it. This is because you compare the qualities of normalized images:
If, on the other hand, you deliberately want to compre on pixel level, e.g. because finding the limits up to where a heavy crop of a high res body can compete with a full frame image of a lower res body at the same resolution, it is a different story. Here we are talking about a step prior to compare the quality of images, i.e. pure sensor performance.

There is no black and white, right or wrong, ..., the truth is always somewhere in between and the "somewhere" depends on which question you ask .

I find myself pixel peeping a lot too. I was just trying to relate it to the original question about why the OP's 45 megapixel camera looked noisier on screen at 100% than the 24 mp camera. I thought it might be more that the OP might be zooming way in in Lightroom to pixel peep each image so they ended up comparing a portion of a giant image from the 45 mp camera to a portion of a relatively smaller 24 mp image. To give an extreme example it would be like looking at a picture of an eyelash compared to a picture of the whole eye. Of course the eye would look better.

 

[DRwyoming]

I find myself pixel peeping a lot too. I was just trying to relate it to the original question about why the OP's 45 megapixel camera looked noisier on screen at 100% than the 24 mp camera. I thought it might be more that the OP might be zooming way in in Lightroom to pixel peep each image so they ended up comparing a portion of a giant image from the 45 mp camera to a portion of a relatively smaller 24 mp image. To give an extreme example it would be like looking at a picture of an eyelash compared to a picture of the whole eye. Of course the eye would look better.

Agreed, it's a pretty common thing when folks first jump to a much higher resolution camera. Both images are zoomed to 100% in a tool like LR but the effective zoom level and how tightly they're zooming into the image frame is substantially higher with the higher resolution camera and noise appears to be much worse when for equal subject level zoom the noise looks quite different.

 

[Replytoken]

Agreed, it's a pretty common thing when folks first jump to a much higher resolution camera. Both images are zoomed to 100% in a tool like LR but the effective zoom level and how tightly they're zooming into the image frame is substantially higher with the higher resolution camera and noise appears to be much worse when for equal subject level zoom the noise looks quite different.

Correct, as the OP acknowledged in post #14 on Jan. 15. I am not sure why this thread was revived and the OP questioned again, but the discussion has been interesting.

--Ken

 

[NatureLover]

I find myself pixel peeping a lot too. I was just trying to relate it to the original question about why the OP's 45 megapixel camera looked noisier on screen at 100% than the 24 mp camera. I thought it might be more that the OP might be zooming way in in Lightroom to pixel peep each image so they ended up comparing a portion of a giant image from the 45 mp camera to a portion of a relatively smaller 24 mp image. To give an extreme example it would be like looking at a picture of an eyelash compared to a picture of the whole eye. Of course the eye would look better.

No zooming in. No pixel peeping. Just normal viewing while editing.

 

[zemlin]

I don't hesitate to push ISO on my D850. The noise is pretty easy to deal with.
Here's a processed image I shot at ISO 25,600.

You can only see EXIF info for this image if you are logged in.

 

[JoPoV]

I don't hesitate to push ISO on my D850. The noise is pretty easy to deal with.
Here's a processed image I shot at ISO 25,600.
View attachment 32785

25600 ISO ! WOUCH ! Still lot of details un fur ! Don't know how to do that with mine.
Which lens, and which software(s) do you use for demosaicing, denoising and postprocess ? At what distance were you ? uncropped ?
And very nice expression. Very nice shoot.

 

[zemlin]

Lens was a Nikon 400mm f/2.8 VR with a TC-20e III 2x teleconverter. I use Topaz Denoise AI for noise reduction and I also like the Topaz Sharpen AI. This also had some photoshop work to smooth out noise left in the background and desaturate color noise in the fur on the chest. The image is cropped, but not heavily. I might have been around 30'/10m from the fox.

 

[JoPoV]

Lens was a Nikon 400mm f/2.8 VR with a TC-20e III 2x teleconverter. I use Topaz Denoise AI for noise reduction and I also like the Topaz Sharpen AI. This also had some photoshop work to smooth out noise left in the background and desaturate color noise in the fur on the chest. The image is cropped, but not heavily. I might have been around 30'/10m from the fox.

Thanks for the infos.
Nice work. And good use of topaz. I often see it used with abuse and being noticeable.
Happy to see D850 at 25600 iso with such result.
And nice approch of the fox. It seems curious and slightly surprised to see you lying in the grass. Nice expression.
Good job.