Jump to content
synthesaur

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

Recommended Posts

Since i bought the XT2 something has been bugging me. I find that the files even at iso 400 are pretty noisy. My old Xt1 even at higher iso produced very clean images.
Also the JPEGs are just too "plastic" looking, too smoothed or something. I used to love classic chrome on xt1 but I cant stand it on xt2 now.
Sometimes the shadows are too dark....image looks cartoonish.  Can't figure it out. 

Share this post


Link to post
Share on other sites

The film simulations on the X-Pro2 and X-T2 are a bit different than previous models. The auto exposure algorithm is different too. Noise in low light is slightly worse than the X-Trans I/II cameras. However, the noise reduction in the JPEGs does appear to be better. If you think the shadows are too dark, then why not set SHADOWS to -1? Or try the Pro Neg S film simulation (I find Provia to be a bit contrasty/garish myself.)

 

Unfortunately, for now anyway, all the new tech is bundled with the 24MP sensor. Personally, I'd love it if Fuji made a 16MP version of the X-Trans III sensor for better low light performance (actually, if I could really get what I want it'd be Bayer), but I don't think that's going to happen.

 

As for the plastic look, I keep the NR set to -2 or -4 (it's too bad there's no way to set this based on ISO like on other camera brands). Or you may try NR 0 to 2 with the NOISE EFFECT enabled... This produces similar levels of detail but with a less objectionable appearance to the noise. (again, it sucks to have to fiddle with this every time you change ISO...)

 

Also, don't forget the relationship between ISO and resolution. You may just need to accept that at 6400 ISO your X-T2 is an 6-8MP camera and print/display your images at the appropriate size.

Edited by kimcarsons

Share this post


Link to post
Share on other sites

Thank you. I do seem to really prefer the results from the 16mp sensor myself. It has some "film-like" magic that I am in vain trying to find in this new camera.
It is bewildering to me that so many people praise the new sensor so much. I am trying to find reasons to keep it but more and more get disappointed with the results. 
Check out this image shot at iso 5000 on xt1 with not noise reduction applied, I can't get this on the new xt2:
https://www.flickr.com/photos/35210521@N03/30636578161/in/dateposted-public/
At 100% in lightroom it is very clean, flickr isn't showing it.

Edited by synthesaur

Share this post


Link to post
Share on other sites

I find my x-pro 1 with x-trans 1 produces better images than my x100t with x-trans 2. The x-trans 2 has the terrible waxy skin at high iso and I hate it. Now you state that x-trans 3 has more noise and look plasticy at high iso as well and so did dpreview. I am beginning to have second thoughts about buying the xt2. Damn!

Share this post


Link to post
Share on other sites

Thank you. I do seem to really prefer the results from the 16mp sensor myself. It has some "film-like" magic that I am in vain trying to find in this new camera.

It is bewildering to me that so many people praise the new sensor so much. I am trying to find reasons to keep it but more and more get disappointed with the results. 

Check out this image shot at iso 5000 on xt1 with not noise reduction applied, I can't get this on the new xt2:

https://www.flickr.com/photos/35210521@N03/30636578161/in/dateposted-public/

At 100% in lightroom it is very clean, flickr isn't showing it.

 

Well, I don't know that there's anything 'film-like' about it. Certainly lower resolution digital images are more 'film-like', so are ones shot with softer/less constrasty lenses.

 

I do know what you mean though. I felt the same way when I made the move from 12MP to 16MP (why are my low light shots so noisy, what happened to those smooth tonal gradations?)

 

Sony seems to be the only one that gets it (with the A7S).

 

That being said, once you scale the images to the same output resolution, much of the difference disappears.

 

The real improvements with the new cameras are not in image quality per se, but in things like AF and sensor read-out speed/viewfinder blackout time, etc.

 

I still have my X-Pro1 and like it better than the X-T2 and X-Pro2 (the IQ is the same for all practical purposes). AF and write speed are slow enough to be a hindrance in many situations though, and that's what I was hoping to fix with the X-Pro2 and X-T2. Well, that and usable video.

 

I think with the X-Pro2 and X-T2 Fuji toned down the noise reduction to deal with the 'waxy skin tones' complaint (which your flickr image does exhibit). I would suggest you try NR +1 or +2 to better match the X-T1's output. Don't forget that if you shoot RAW or RAW+JPEG you can always re-render the JPEG at different NR settings in camera from the playback Q menu. 

Share this post


Link to post
Share on other sites

How about a shot of a person? Fujifilm high iso is great until you use it with human skin where it gets that horrible waxy look.

 

OK, a person then: Shot indoors with only a ceiling light fixture holding a couple CF bulbs -- ISO 12800. For an APS class camera I think this is pretty impressive. Excellent color and noise barely an issue.

 

 

joe_iso_12800.jpg

Share this post


Link to post
Share on other sites

The film simulations on the X-Pro2 and X-T2 are a bit different than previous models. The auto exposure algorithm is different too. Noise in low light is slightly worse than the X-Trans I/II cameras. However, the noise reduction in the JPEGs does appear to be better. If you think the shadows are too dark, then why not set SHADOWS to -1? Or try the Pro Neg S film simulation (I find Provia to be a bit contrasty/garish myself.)

 

Unfortunately, for now anyway, all the new tech is bundled with the 24MP sensor. Personally, I'd love it if Fuji made a 16MP version of the X-Trans III sensor for better low light performance (actually, if I could really get what I want it'd be Bayer), but I don't think that's going to happen.

 

As for the plastic look, I keep the NR set to -2 or -4 (it's too bad there's no way to set this based on ISO like on other camera brands). Or you may try NR 0 to 2 with the NOISE EFFECT enabled... This produces similar levels of detail but with a less objectionable appearance to the noise. (again, it sucks to have to fiddle with this every time you change ISO...)

 

Also, don't forget the relationship between ISO and resolution. You may just need to accept that at 6400 ISO your X-T2 is an 6-8MP camera and print/display your images at the appropriate size.

 

...."Also, don't forget the relationship between ISO and resolution. You may just need to accept that at 6400 ISO your X-T2 is an 6-8MP camera and print/display your images at the appropriate size.".

 

Huh???

 

Do you mean SOOC jpegs?

 

If you mean the sensor is a base ISO sensor of native ISO 200 that a ISO 6400 image is actually an image pushed 5 stops...and thus the inherent noise, OK, but I'm at a loss as to what the 6/8 megapixel reference is based upon?

 

Can you tell us what that means???

 

(My two cents on the topic of is the images from the X-T2 is that they are fantastic when the settings are properly applied and/or proper PP. A notch up from the X-T1.)

 

Peter

Share this post


Link to post
Share on other sites

...."Also, don't forget the relationship between ISO and resolution. You may just need to accept that at 6400 ISO your X-T2 is an 6-8MP camera and print/display your images at the appropriate size.".

 

Huh???

 

Do you mean SOOC jpegs?

 

If you mean the sensor is a base ISO sensor of native ISO 200 that a ISO 6400 image is actually an image pushed 5 stops...and thus the inherent noise, OK, but I'm at a loss as to what the 6/8 megapixel reference is based upon?

 

Can you tell us what that means???

 

(My two cents on the topic of is the images from the X-T2 is that they are fantastic when the settings are properly applied and/or proper PP. A notch up from the X-T1.)

 

Peter

 

Noise isn't information. If you take an 8MP image, and upscale it, then add fake noise, you'll find at some point you've matched the (noisy) high ISO image of the same scene shot at 24MP, because the noise in the 24MP image was just that... noise. Not actually information from the scene. Or conversely, if you take a noisy ISO 6400 24MP image and start scaling it down, you'll reach a point where the noise becomes imperceptible (it averages out). This is why you never see noisy photos on instagram... The images there are only displayed at like 3MP.

 

People fall into this trap every time a new, higher megapixel image sensor comes out. They look at a 16MP ISO 6400 image at 1:1 and compare it to the 24MP ISO 6400 image at 1:1. That's not a valid comparison. If you downscale the 24MP image to 16MP, then maybe you can compare. But chances are nobody but the photographer will ever see the image at 1:1 anyway, they'll see some 3-6MP downscaled version on the web.

 

The moral of the story is: if you want to crop a lot or make a giant print (i.e. benefit from that 24 megapixels) then you're going to have to shoot at ISO 200. If you accept that you will only make smaller prints or only display on the web, then you can crank the ISO as high as you want (the higher you go, the smaller the dispaly MP you can support without the image appearing noisy).

Share this post


Link to post
Share on other sites

Noise isn't information. If you take an 8MP image, and upscale it, then add fake noise, you'll find at some point you've matched the (noisy) high ISO image of the same scene shot at 24MP, because the noise in the 24MP image was just that... noise. Not actually information from the scene. Or conversely, if you take a noisy ISO 6400 24MP image and start scaling it down, you'll reach a point where the noise becomes imperceptible (it averages out). This is why you never see noisy photos on instagram... The images there are only displayed at like 3MP.

 

People fall into this trap every time a new, higher megapixel image sensor comes out. They look at a 16MP ISO 6400 image at 1:1 and compare it to the 24MP ISO 6400 image at 1:1. That's not a valid comparison. If you downscale the 24MP image to 16MP, then maybe you can compare. But chances are nobody but the photographer will ever see the image at 1:1 anyway, they'll see some 3-6MP downscaled version on the web.

 

The moral of the story is: if you want to crop a lot or make a giant print (i.e. benefit from that 24 megapixels) then you're going to have to shoot at ISO 200. If you accept that you will only make smaller prints or only display on the web, then you can crank the ISO as high as you want (the higher you go, the smaller the dispaly MP you can support without the image appearing noisy).

 

Good post. 

Share this post


Link to post
Share on other sites

Yeah, I'm not convinced that the X-T2 is cleaner (free from noise) at higher ISOs than the X-T1. I have to do some testing to generate some data and find out between the two cameras. I have a hunch, though, that the overall "noise floor" at a given ISO (above say, ISO1600) is lower on the X-T1. 

 

I will say that the noise performance of my X-Pro1 was very, very good up to ISO 1600. 

Share this post


Link to post
Share on other sites

The jpegs from the xt2 are just horrid as if taken with a cell phone. Flat and cartoonish.. Classic chrome is lifeless compared to xt1's.
Here is classic chrome @ 1600 SOOC XT1 shot:
_T011613_zpsbvj5izth.jpg


And here are some SOOC images from xt2 @ 400
DSCF1508-2_zps2fxmihgx.jpg
DSCF1508_zps6fjhsqca.jpg

@800

DSCF1494-2_zpsctql8sq3.jpg
DSCF1494_zpsg9xafa19.jpg

Edited by synthesaur

Share this post


Link to post
Share on other sites

@synthesaur I'm not sure what your pictures are meant to illustrate. It's a very subjective comparison without using the same lens, same scene, same subject, same lighting (not to mention the same camera settings.) Doesn't have to be a studio shot. Your X-T1 shots look either a little underexposed or with SHADOW TONE > 0. Don't forget that if you like that look, SHADOW TONE goes to +4 on the X-T2.

 

If you're upset that X-Trans III isn't better than X-Trans II, then I'm with you there, the IQ is only better at ISO 200, and only very slightly better (which is nice for cropping.) But of course you get a lot more than just a new sensor in the X-T2 (which is why I kept mine).

 

Perhaps a more direct comparison would put your mind at ease. Do you still have the X-T1?

Edited by kimcarsons

Share this post


Link to post
Share on other sites

Tested the xt-2 in a camera shop. Lighting was indoor lighting in a normal shop.
I find that noise tends to come in around 3,200 & 6,400 was very noisy & grainy to me.
This kinda holds me back from purchasing the xt-2 for a moment.

 

Here's a quick simple comparison. Please download the actual files here & tell me what u guys think:
https://drive.google.com/open?id=0B-Mudsk627xpS0t2V2dfSUY5RVE

- Raw files using Lightroom

- Shot using 35mm f2

- constant f2 for all shot

post-17133-0-60408600-1478652770_thumb.jpg

post-17133-0-14641200-1478652780_thumb.jpg

post-17133-0-36802500-1478652786_thumb.jpg

post-17133-0-03730700-1478652793_thumb.jpg

post-17133-0-82616300-1478652797_thumb.jpg

Share this post


Link to post
Share on other sites

I would suggest turning the sharpening down by a notch or two, and then sharpen in LR (JPEGs)...

 

In LR you can do selective sharpening with a mask, while the in-camera sharpening sharpens the whole image, including OOF areas which tends to amplify the perceived noise..

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...