Noise

[Stephenaweiss]

Can anyone point me to a thread or answer my question about noise


I have done some long exposures at dawn with interesting results, but significant noise in the sky.
I am shooting at ISO 400, aperture 2.8, for about 30-60 seconds

Is there a way in photoshop to reduce the noise, without losing variations in light..
thanks, sw

[marcmccalmont]

Can anyone point me to a thread or answer my question about noise
I have done some long exposures at dawn with interesting results, but significant noise in the sky.
I am shooting at ISO 400, aperture 2.8, for about 30-60 seconds

Is there a way in photoshop to reduce the noise, without losing variations in light..
thanks, sw
[a href=\"index.php?act=findpost&pid=88509\"][{POST_SNAPBACK}][/a]

Have you tried noise ninja? it has noise profiles for some cameras and works very well or switch to Lab mode and do your noise reduction there
Marc

[erick b]

or neatimage a plugin (or standalone for photoshop

[John Sheehy]

Can anyone point me to a thread or answer my question about noise
I have done some long exposures at dawn with interesting results, but significant noise in the sky.
I am shooting at ISO 400, aperture 2.8, for about 30-60 seconds[a href=\"index.php?act=findpost&pid=88509\"][{POST_SNAPBACK}][/a]

30-60 seconds brings up a lot of dark current noise.  You should either use the camera's long exposure noise reduction feature if it has one, or take a shot of the same length with the lens cap on immediately after, to subtract from the first to reduce the bright pixels.

With some cameras, it is better to expose well at ISO 1600 or 800 than to under-expose at ISO 400.  Under-exposure can result in a lot of noise.

[Jonathan Wienke]

It's always better to expose properly at a higher ISO than underexpose at a lower ISO, as long as the higher ISO isn't faked by math tricks in the camera, like 3200 on the 1D-MkII.

[bjanes]

It's always better to expose properly at a higher ISO than underexpose at a lower ISO, as long as the higher ISO isn't faked by math tricks in the camera, like 3200 on the 1D-MkII.
[{POST_SNAPBACK}][/a]

The above is not always true. It is always desirable to give as much exposure as possible to reduce photon noise, but when ISOs above the unity gain of the camera are used, one gains little additional information. Unity ISO varies from 1600 to 800 with many common DSRLs. The higher ISO will give somewhat better AD conversion.

[a href=\"http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/]http://www.clarkvision.com/imagedetail/dig...rmance.summary/[/url]

In the case of the Canon 1D MII, ISO 1600 will give slightly better read noise, but ISO 800 will give more DR. At long exposures (more than several seconds, thermal noise becomes a problem)

http://www.clarkvision.com/imagedetail/eva...-1d2/index.html

[John Sheehy]

In the case of the Canon 1D MII, ISO 1600 will give slightly better read noise, but ISO 800 will give more DR. At long exposures (more than several seconds, thermal noise becomes a problem)

[a href=\"index.php?act=findpost&pid=88590\"][{POST_SNAPBACK}][/a]

Dark current noise is like photon shot noise; it exists relative to absolute signal, and is independent of ISO (IOW, it exists beyond and before any readout or digitization).

[bjanes]

Dark current noise is like photon shot noise; it exists relative to absolute signal, and is independent of ISO (IOW, it exists beyond and before any readout or digitization).
[a href=\"index.php?act=findpost&pid=88727\"][{POST_SNAPBACK}][/a]

I did not say that dark current (thermal) noise was related to ISO. As I said, it is encountered with long exposure, such as in astronomical work. As implied by its name, it is related to the temperature of the sensor and the time of the exposure, and not to the absolute signal, which will depend upon the luminance of the main subject and the exposure (f/stop and shutter).

[John Camp]

Stephen,

Where you're talking about noise in the sky, are you talking about the grey wave-like noise where it looks like the program is trying to distinguish between two close shades -- almost like a brush had been dragged across it? Or are you talking about some kind of speckling effect?

JC

[Jonathan Wienke]

The above is not always true. It is always desirable to give as much exposure as possible to reduce photon noise, but when ISOs above the unity gain of the camera are used, one gains little additional information. Unity ISO varies from 1600 to 800 with many common DSRLs. The higher ISO will give somewhat better AD conversion.

And that's why the results are still better overall. I've done a lot of low-light shooting with Canon DSLRs, and after >50,000 frames, what I've found works best is to set the ISO high enough to get sufficiently fast shutter speed to avoid motion blur, and expose as much as possible without blowing critical highlights. Combining those principles will give you the best possible signal/noise ratio and overall image quality.

[Stephenaweiss]

Stephen,

Where you're talking about noise in the sky, are you talking about the grey wave-like noise where it looks like the program is trying to distinguish between two close shades -- almost like a brush had been dragged across it? Or are you talking about some kind of speckling effect?

JC
[a href=\"index.php?act=findpost&pid=88786\"][{POST_SNAPBACK}][/a]

Thanks, I am talking about noise in the dark sky.
st

[Forsh]

Thanks, I am talking about noise in the dark sky.
st
[a href=\"index.php?act=findpost&pid=89652\"][{POST_SNAPBACK}][/a]

I usually have the same problem.

[:Ollivr]

I thought DSLRs in general subtract a dark frame above a given exposure time?
Not sure if subtracting another one would help but I have to admit I never tried it.

Some suggestions on noise reduction from my own (rather beginner's) experience:

From a second point of view (not the photographer's), noise is often less of a problem than the artifacts you get from noise reduction/ the lost detail. If detail is first blurred and then, in part, recovered via sharpening, this may look better onscreen but make a worse print than one with a tad of noise.

You may want to try to mask the sky, convert the image to LAB and work on the single channels. On the chroma channels, try the "dust and scratches" filter. On the luma channel, use the (selective) blur tool here and there.

I have tried Neatimage but did not like it. In general I now stay away from either sharpening or blurring (i.e. noise red) of the whole picture (as long as its properly focused) and rather do it selectively. Of course its another thing when it comes to uprezzing/downsizing pics for which it may be still handy but I have only now started looking at that and cannot tell much about it.

I would think you'd roll better with keeping the ISO at its initial value and increase the exposure time to compensate for the lost stops.

Greetings,

O.

[Stephenaweiss]

You may want to try to mask the sky, convert the image to LAB and work on the single channels. On the chroma channels, try the "dust and scratches" filter. On the luma channel, use the (selective) blur tool here and there.


Thank you very much, Very interesting...Stephen