PDA

View Full Version : A couple of observations on ISO levels and printing.



I @ M
24-02-2016, 12:47pm
I have witnessed the extraordinary advances in technology over the years that camera makers have managed to squeeze out of their products when it comes to iso levels.
Many times I have been somewhat underwhelmed by images posted on the 'net that have been taken at high iso levels and even more disappointed by the "rescue" attempts with noise reduction software.
The balance between non smeared detail and noise seems to be a fine line when viewed on high contrast monitors and even more so when viewed at pixel peeping levels.
Going back in history I took a photo at iso 400 with a Nikon D200, that model was never renowned for having good high iso ability but after processing the image with very moderate noise reduction and deciding it should be printed I was quite pleased that the resultant 16x24 inch monochrome print showed hardly any noise.

Fast forward a few years and whilst I try to stay at the camera's base iso of 100 for the majority of my images I found myself wanting to produce an image under the lights of our kitchen which consist of a 13 watt LED light in a shade which gives mostly downward light, a couple of 50 watt halogen spots above the sink and a very dim rangehood light.
To get an acceptable shutter speed for both hand holding, allowing 1/2 stop overexposure and to keep subjects from displaying any slight movement I needed to have 5000 iso at the aperture I wanted.
Processing the image and after experimentation with varying levels of noise reduction resulted in noise that was, to my eyes at least, somewhat ugly in the darker areas of the image. That is when viewed on a monitor at up to 100% magnification.
I went ahead and had the image printed as a 16x24 glossy and to say I was surprised at the finished image would be a huge understatement.

Looking closely at a print that is meant to be looked at from a reasonable viewing distance shows a little noise where I KNOW it is but in keeping with the slightly oddball subject of the photo it is barely noticeable to anyone else.

Moral of the story is ----- don't fear high iso levels, stop pixel peeping at pictures on monitors and head forth and print ALL your images. :D

The current cameras here are a couple of generations old now so the newer models with even more stratospheric iso levels must produce better again.

Aint technology marvelous.

jev
24-02-2016, 8:52pm
Moral of the story is ----- don't fear high iso levels, stop pixel peeping at pictures on monitors and head forth and print ALL your images. :D
Hear-hear!

I still have a large (75x125) print hanging in the hallway that was shot on a darkish day somewhere in the winter of 2004 using a Canon EOS D30. Yes, you read that right: D30, not a 30D. A full 3 MPixel, 400 ISO image. The print has been seen by several fellow photo-enthusiasts since than and not a single one commented on either the resolution or the noise. Until I told them, that is. :2smile:

Pixel peeping and "noise" are issues for monitor staring, not photo printing.

John King
24-02-2016, 9:15pm
Monitors are low resolution output devices. My 1920x1200 IPS aRGB panel is composed of some 27,000 sub-pixels per square inch.

Printers are high resolution output devices. My Epson R3880 is laying down some 2,880x2,880 ink droplets per square inch (~8,294,400 droplets per sq.in. ... ), more or less regardless of the dpi at which one chooses to print (it also has a lower ink usage 1440x2880 mode that will be lower resolution ... ;) - I never use that mode).

This explains most of the differences to me. My entire monitor panel has fewer 'sub-pixels' than one square inch from my printer.

I recently took an image at ISO 12,800 using my E-M1 and the somewhat questionable quality 14-42 EZ collapsible pancake lens. Taken at 1/6th second from a moving footbridge (the Southbank, Melbourne pedestrian bridge). After a bit of noise reduction, the RAW printed up very nicely at A4 size. The JPEG was pretty ordinary though. This did surprise me. I really did not expect it to scrub up as well as it did, figuring that this was well and truly happy snap territory - 5x7" at best. I could probably get away with printing this at A3 size, which really does surprise me.

I have also printed fine art prints at A2 size from my 5 MPx E-1, but not using anything over about ISO 800. That wasn't the E-1's strong point, shall we say :lol:.

tduell
25-02-2016, 7:46am
If noise is a real concern and you are shooting a static scene, one solution is to use the "zero noise" approach.
A search should turn up some useful info on this, it isn't hard to do, and does work.
In essence, shoot a number of exposures, some say the highest should be +4 EV.
These can then be 'fused', similar to HDR to get your result.
The software 'enfuse' is ideal for this as it is particularly good at selecting the lowest noise pixels.
It can all be done automagically with hugin-2016.0.0, which is currently in beta testing.

Cheers,
Terry

I @ M
25-02-2016, 11:50am
If noise is a real concern and you are shooting a static scene, one solution is to use the "zero noise" approach.
A search should turn up some useful info on this, it isn't hard to do, and does work.
In essence, shoot a number of exposures, some say the highest should be +4 EV.
These can then be 'fused', similar to HDR to get your result.
The software 'enfuse' is ideal for this as it is particularly good at selecting the lowest noise pixels.
It can all be done automagically with hugin-2016.0.0, which is currently in beta testing.

Cheers,
Terry

Yep, for totally static scenes where you can use a tripod to "bracket" images so to speak I am sure the modern day software does an admirable job.
The thing that the high iso capability seems to lend itself to me more is the ability to hand hold and maintain shutter speeds for a single exposure.
And from there, processing and printing the image brings out the true value of the tech.

Thor
27-02-2016, 10:59pm
Good advice I @ M. I will very rarely print and would never think of printing a pic that exhibits high noise. Guess you proved that one wrong ;).