PDA

View Full Version : ISO 64 dynamic range


drmarkf
13th June 2017, 09:41 PM
Having just had an idle dip in to the m4/3 forum on dpreview (I know, I know...) I noticed some discussion of the Low ISO mode on the 1ii having around 1-stop lower dynamic range than the ISO 200 base.

I know a dip is sometimes seen for older cameras below base ISO, but I thought the graph just plateaued on the 1ii - e.g. http://photonstophotos.net/Charts/PDR.htm#Olympus%20OM-D%20E-M1%20Mark%20II

This, also: https://www.dxomark.com/Reviews/Olympus-OM-D-E-M1-Mark-II-sensor-review-New-standard

A bit of googling hasn't turned up anything else I'd trust: has anyone got a definitive answer?

Certainly noise is good at 64, and having the option is great for blurry water.

David M
13th June 2017, 09:47 PM
Well Kodachrome 64 had more dynamic range than Velvia. :D

drmarkf
13th June 2017, 10:12 PM
I was more of a K25 man...

It's scarcely a major issue (like working at ISO 400 rather than 200, perhaps): a classic dpreview subject!

David M
13th June 2017, 10:27 PM
I had an OM body calibrated/dedicated to K25, K64 and K200 (after Kodak launched K200).

Then Fuji introduced Velvia and editors were looking for over saturated colours and blocked shadows. :rolleyes:

chorleyjeff
13th June 2017, 10:33 PM
I was more of a K25 man...

It's scarcely a major issue (like working at ISO 400 rather than 200, perhaps): a classic dpreview subject!

Wimp.
I managed with ASA 10 Kodachrome. With a Retinette 1a.
That ASA 25 stuff was wish washy.
Jeff

drmarkf
13th June 2017, 11:09 PM
Mmmmm. I also liked Agfa CT18.

My Avatar is a scan of a K25 slide taken by my Dad on his Voigtlander rangefinder.

David M
13th June 2017, 11:36 PM
I shot CT21 as my high speed stock at times when the cool colour balance of Ektachrome 200 wouldn't suit the subject.

pdk42
14th June 2017, 06:33 AM
Back to Mark's question..

I've found that ISO64 works very well. There's a tendency to blow highlights more easily so you'll need to drop the exposure a little (half a stop it so). That means you'll need to boost the shadows later which defeats some of the advantages so I personally only use it when the DR of the scene is low.

Michael Sewell
14th June 2017, 07:20 AM
I use ISO64 a lot, particularly when HDR'ing Trucks with flash etc.

drmarkf
14th June 2017, 07:28 PM
I've found that ISO64 works very well. There's a tendency to blow highlights more easily so you'll need to drop the exposure a little (half a stop it so). That means you'll need to boost the shadows later which defeats some of the advantages so I personally only use it when the DR of the scene is low.

I'm following a relevant discussion on the m4/3 dpreview forum which is really interesting.

There's clearly some differences in the way the sensor is read/driven between 64 and 200 - for example, under-exposing by 1/3 stop is recommended for 64 (bearing out what Paul says) while up to 1.3 stops over exposure is recoverable at 200.

I haven't used 64 much yet (some slow-shutter water shooting, plus propellor-aircraft in very bright light) but I certainly agree the noise floor is fantastic.

My own experience at 200 is certainly that much more raw recovery is possible than with the mki, but I'm not sure I'd yet say 1.3 over was reliably good. I guess there will be differences introduced by how you set your metering (currently I'm using centre-weighted matrix more than before) and maybe other variables, and presumably the raw converter will also affect things.

pdk42
14th June 2017, 09:37 PM
It's a pity that rawdigger is no longer free - that is a great tool for looking at exposure impact on the real raw data. I'm no skinflint but €14 for something with very specialist use is really a bit much.

pdk42
14th June 2017, 09:39 PM
I'm following a relevant discussion on the m4/3 dpreview forum which is really interesting.

There's clearly some differences in the way the sensor is read/driven between 64 and 200 - for example, under-exposing by 1/3 stop is recommended for 64 (bearing out what Paul says) while up to 1.3 stops over exposure is recoverable at 200.

I haven't used 64 much yet (some slow-shutter water shooting, plus propellor-aircraft in very bright light) but I certainly agree the noise floor is fantastic.

My own experience at 200 is certainly that much more raw recovery is possible than with the mki, but I'm not sure I'd yet say 1.3 over was reliably good. I guess there will be differences introduced by how you set your metering (currently I'm using centre-weighted matrix more than before) and maybe other variables, and presumably the raw converter will also affect things.
Can you post a link Mark? I'm no fan of DPReview's forums, but there are some threads on there that are useful and don't descend immediately into the usual flaming and rudeness.

drmarkf
14th June 2017, 11:06 PM
Here's the thread - it's still developing:

https://www.dpreview.com/forums/thread/4168237

Ricoh
15th June 2017, 08:21 AM
Pardon me as a non-EM1 Mk2 owner, but a digital camera work best at its base ISO. Sensors only have one response (actually many, but it it's a function of wavelength), and twiddling the ISO dial only alters the gain setting of the electronics once the photons have done their work on electron potential.

Phill D
15th June 2017, 09:02 AM
That's an interesting link Mark thanks. The sunset references and follow on link are really interesting too.

pdk42
15th June 2017, 09:42 AM
Pardon me as a non-EM1 Mk2 owner, but a digital camera work best at its base ISO. Sensors only have one response (actually many, but it it's a function of wavelength), and twiddling the ISO dial only alters the gain setting of the electronics once the photons have done their work on electron potential.

I think that's an over simplification Steve. There are lots of moving parts, algorithms and decisions that go into ISO determination. That starts off with ISO calibration itself which is not as straightforward as it seems. Then you factor in analogue gain circuitry (before digitisation), tone curve mapping etc.

Whilst in the real world you're probably right about base ISO being the best for IQ generally, extended low ISO, for the right scene, will deliver lower noise and increased apparent sharpness.

Ricoh
15th June 2017, 10:02 AM
As for A/D gain, before importing to the wizardry of digital manipulation, in theory lower gain does improve S/N, so there could be something in this ISO 64 being discussed.

In most cases it's better to keep the gain at optimum setting, as per the design. Of course I've not seen any design data so I'm just talking theoretically.