noise level acceptable?

[philip518]

Hi:

Is it possible that I can get some opinions on the noise level of a picture I took recently?

I took a group of shots with manually bracketed exposures. But when zoom in in 100%, many of them show pretty high noise level. the one I attached here is one of the better ones.

It is only a very small, but representative portion of the original picture, converted to jpeg (with highest quality) to meet the size and format requirements for attachments. the histogram is from the original picture.

there is no post processing on this particular picture yet. I tried to use noise reduction, but couldn't decide how much was good enough without losing too much details of rock face.

So what do you think about the noise shown in this picture? would you use noise reduction on it?

Thanks.

Phil

[TonyW]

To my eye, this is not too noisy. It looks better with a bit of sharpening though. Even then, the noise does not seem to me to be high.

[Stagecoach]

Phil, I do not see any noise at this size.

Zooming in at 100% may very well show some noise but how you treat this should also be determined by the end use of the image. Any attempt at noise removal is also going to reduce sharpness which may be more detrimental than leaving any noise as it is.

With a well exposed image as yours is and at ISO320 at screen viewing sizes there's unlikely to be any noticeable noise.

Grahame

[revi]

No visible noise for me (but is that a 100% crop you showed, or reduced?).

What I do notice is that the histogram starts far to the right, and that the rock face lacks contrast.
You could try increasing the black point a bit, to see what that does (as a side effect, that often makes shadow noise less visible as well).

[Shadowman]

Would have been better to post the other bracketed shot which showed the noise, it's not visible here. The posted exif data, is this from the same image posted?

[xpatUSA]

I took a group of shots with manually bracketed exposures. But when zoom in in 100%, many of them show pretty high noise level. the one I attached here is one of the better ones.

It is only a very small, but representative portion of the original picture, converted to jpeg (with highest quality) to meet the size and format requirements for attachments. the histogram is from the original picture.

there is no post processing on this particular picture yet. I tried to use noise reduction, but couldn't decide how much was good enough without losing too much details of rock face.
Thanks.

Phil

I see next to no noise in the picture but it looks like plenty of NR has been applied whether you wanted it or not. I base this assertion because a selected portion of the sky shows a very narrow histogram, nowhere near what it should be according to the laws of quantum physics (Poisson distribution, aka 'shot noise', one std deviation = the square root of the signal).

The rock edges are showing some slight halos indicating that some sharpening, or possibly micro-contrast adjustment, has been applied. Any more sharpening could make the halos more obvious.

My feeling is that the same scene shot RAW (NEF) on a tripod would give a much better starting point.

P.S. I see that the ICC profile in your image is Adobe RGB (1998) - some might say that sRGB is better for the Web.

[rpcrowe]

I cannot detect the noise level from your small image I can view. IMO, ISO 320 is not too high of an ISO level and I sometimes shoot at that level when capturing moving subjects or when shooting hand held with a long lens. You really should not experience a lot of noise at that ISO but, I will normally shoot at the lowest ISO that will allow the f/stop and shutter speed I desire.

As an example, for the below image shot with my Canon 7D and a 300mm f/4L IS lens; ISO 200 allowed me to shoot at 1/2000 second wide open at f/4. I wanted a very high shutter speed to shop the action and wanted to shoot wide open at f/4 to blur the background)...



I shot the below image at ISO 100 to give me the most noise free image. This was shot with a Canon 40D and a 70-200mm f/4L IS lens at 200mm with a CPL filter added. The ISO 100 allowed me to shoot at 1/250 second @ f/4.5, even with the exposure loss due to the CPL...



However, I had been shooting some birds when I came upon this idyllic scene in the lake below. I had been using ISO 400 to get a high shutter speed for the birds and did not change my ISO for this shot, but, I did change my f/stop to f/9 and accepted a slower shutter speed. The reason I am showing this is that I don't detect a significant amount of noise in this ISO 400 image...



OTOH, if I had been shooting your image of the rock formation, I would likely have used a lower ISO. I have been torn between shooting my Canon cameras at ISO 160 or ISO 100. (I believe that ISO 200 is the lowest many Nikon cameras go). I have read that ISO 160 is a native ISO for Canon cameras but, I really don't see any difference between the quality of my images at ISO 160 or ISO 100... ISO 160 would have allowed you to capture this image at 1/400 @ f/8 while the ISO 200 of a Nikon would have allowed you to shoot at a slightly faster shutter speed. Your 1/800 speed was overkill and not necessary for a still subject with an 85mm lens.

I asked a Canon representative at a recent Canon seminar and the rep said that he has "heard" about the Canon native ISO being 160 but, has heard nothing from the "Canon Gods" to confirm or disclaim this....

[Peeshan]

>Richard: Not to dismiss all those details and informations in your post, but ISO numbers don't really means anything by themselves when it come to noise, it depends of the camera; ie. ISO 500 will probably produce lot of visible noise if max ISO is 800 but you won't notice it the camera max ISO is 6400.

As a precision, nikon minimum ISO is 100 on more recent model

[IzzieK]

I do not like looking at small images, so I will just welcome you to this forum...and to remind you to please fill up the details in your profile.

[jcuknz]

Another person worrying about the irrelevant ....I rarely see noise, none in the example for sure, and gave up using noise reduction programmes years ago ... but then I am an ex-film user who was used to grain as opposed to somebody presumably starting in the foolishly hyper-critical digital age where any 'defect' is 'bad'.
Whatever .. welcome to CiC
One of my favourite experiments using max USM on a file twice .... 1000 USM "Off to Battle"



correction, I se from the filename it is only 500USM from back in the days when I used USM

edit . I remember my amusement at a judge's horrified reaction to it in a club competition

[jcuknz]

The 1K USM image ... repeating the process didn't bring much change I remember ... back in the days when I was using a 3.3Mp P&S and a cheap tele-adaptor from a pawnshop. Gave me a 160mm telephoto view
Quite something for somebody usually using a 50mm lens on full frame.

[mrinmoyvk]

Noise is acceptable for the attached image for my eyes.

From the EXIF data it seems that the exposure compensation was -2
I have noticed from my personal experience (might be wrong) changing the exposure compensation reduces the image quality a bit. Degradation is very less but it is.
Also spot metering was used, as this is the small part of image, there might be some subject where in it was required, however if only landscape was to be shot, I would prefer matrix metering.

Rather than changing the exposure compensation you may try the below things.

1. Find out the aperture range between which your lens performs best and stick to that range
2. Also stick to lowest ISO available on your camera
3. I would suggest to always play with shutter speed to adjust the amount of light

[HaseebM]

To me it looks like it was not focused properly or maybe its to do with the small image size?