ISO Test

[sparkie1263]

I will git it a try again.

Thanks
Frank

[Rooz]

why bother Frank, its clearly much, much better.

besides which iso 6400, f8, 1/4000th is a poor iso test as it is not low light so is not testing the sensitivity of the sensor to light which is exactly what iso is supposed to be.

[DonSchap]

Frank,

Again there seems to be a creep up the scale as you arrive at 12800. Maybe that's the secret of the better response of the sensor, as you go higher in ISO, it seems to be washing the image out ... or in-camera brightening/desaturation of it.

Thanks for the second sweep.

[Peekayoh]

Confusion reigns, methinks.

Don, any washing out is down to exposure. Frank has kept a constant aperture/shutter combination above iso800 and, in 'P' mode, relied on the flash to calculate and supply the correct exposure while changing ISOs.

Frank, you screwed up and, as Rooz says, it's useless to try and assess the high ISO performance of a camera when there is plenty of light around, that simply puts the signal well above the noise floor.
It's possible to produce meaningful results using flash but you need to be in full control of the lighting. That means manual mode with the camera on a tripod and in a dim room.
Start at f/2 or f/4 with the flash off camera on full power. Move the flash back, bounce it off a wall, put on a diffuser, anything to get the shutter up to 1/2sec or even 1sec.
Once set, you can lower the flash output in steps (1/2,1/4,1/8,1/16,1/32) whilst increasing the ISO (400,800,1600,3200,12800)
Mind you, I'm rarely bothered with a lack of ISO as I tend not to take pictures in the dark. An overrated deficiency in my book.

In the interest of clarity, I have to pick up something Rooz said of your test that it ... "is not testing the sensitivity of the sensor to light which is exactly what iso is supposed to be".
What you are actually testing is the dark noise inherent in the gain circuitry behind the sensor and noise introduced by the A/D conversion.
The Sensor itself has a single, fixed sensitivity. AFAIK, every Sony camera since the A100 has a Sensor with a base ISO of 200 giving a 1 stop advantage over the A100 with it's base ISO of 100.
At base ISO, no gain (amplification) is required so if Sony were able to produce a Sensor with a base of ISO400 we would benefit by another stop in terms of noise.

[DonSchap]

I had just assumed a "constant" light source ... and completely forget to mention it.

(slappin' the ol' forehead ... D'oh!)

You have to conduct the test under identical lighting conditions, for it to be valid concerning exposure. Anything else is a loose cannon.

If it is just a test of the ... sharpness of the image ... well, exposure is less important, but still cannot be ignored. Again, it is nice to have a constant light source to shoot with, hence hot lights, studio strobes or the God, Apollo at your beck and call. The elex-pop-up flash is a real roll of the dice, in my estimation and as witnessed.

You will have to forgive me, as the α850 does not have Built-in Flash. I am usually forced to consider the light source, all the time, as I either have to add the:

• SONY HVL-F20AM,
• HVL-56AM,
• HVL-F58AM or
• Metz mecablitz 76 Handheld arc-welder,

if the indoor lighting is problematic or mother Sun has hit the sheets.

Carry on.