Did some quick low low light testing to find the highest 'usable' ISO. 6400 is the highest that is usable IMO for general use. I think that you can get good size prints from 6400 with plenty of detail and relatively high color fidelity. I found ISO 16,000 to be the highest you can go before you start getting heavy magenta color shift, which gets bad at 20,000 and worse at 25,600, though can be usable for web or small prints if processed right. 16,000 could be used for small prints or web use when needed. The shot below printed at 4x6 shows very little noise in the print with lots of detail. 5x7 I think would be fine at this ISO and 8x10 might even be OK if you don't mind some noise in the print.
This is ISO 16,000 in very dim light (stray light at night from the next room over) with some Lightroom noise reduction to taste (balance noise and detail), which I find about luminance 40 to be good for this high ISO.
Jman13 wrote:
Did some quick low low light testing to find the highest 'usable' ISO. 6400 is the highest that is usable IMO for general use. I think that you can get good size prints from 6400 with plenty of detail and relatively high color fidelity.
Thank you a lot for the image and info. This kind of "report from the field" is often more valuable than a formal review.
Just a question about high ISO. To put your words in context, what do you consider your "usable" high ISO in other cameras?
I am asking this because each of us has a kind of personal "threshold", for example I consider ISO 800 for Canon 450, ISO 1600 for 7D and 3200 for 5D Mk II, I go above those values only in an emergency.
By comparing what's your "usable" high ISO on other cameras I could get an idea how it compare to my personal idea of "usable".
Usable to me means 8x10s that look good...maybe a little visible noise in a print that size, but not so much as to dominate the image. From the sound of your comparisons, you are more noise-averse than I am. I haven't used most of the cams you've referenced, but I think you'd probably be fine with 1600 and could use 3200 in a pinch.
Due to this thread I am seriously considering OM-D as a secondary system to my Nikon D700 and full set of ZF lenses.
I am thinking about a small package:
12/2 for a good quality wide angle
20/1.7 for a pancake-take-everywhere-lens and unobtrusive people and street lens (might add a 25/1.4 later as the better but bigger alternative).
other than that I would use lenses I already have, like my Zeiss ZF lenses and especially:
60/2,8 Macro Leica R
280/4 Leica APO Telyt. I have bought this recently and want a crop body for it for extended reach - so far I thought about a D 7000.
It seems that with the latest sensors m4/3 has reached IQ of APS-C sensors of the last generation (like D 7000), which is enough for me.
Do you guys think the EM-5 would support the Leica 280/4 with enough pixel goodness to still shine? Is it an alternative for the D7000?
Due to this thread I am seriously considering OM-D as a secondary system to my Nikon D700 and full set of ZF lenses.
I am thinking about a small package:
12/2 for a good quality wide angle
20/1.7 for a pancake-take-everywhere-lens and unobtrusive people and street lens (might add a 25/1.4 later as the better but bigger alternative).
other than that I would use lenses I already have, like my Zeiss ZF lenses and especially:
60/2,8 Macro Leica R
280/4 Leica APO Telyt. I have bought this recently and want a crop body for it for extended reach - so far I thought about a D 7000.
It seems that with the latest sensors m4/3 has reached IQ of APS-C sensors of the last generation (like D 7000), which is enough for me.
Do you guys think the EM-5 would support the Leica 280/4 with enough pixel goodness to still shine? Is it an alternative for the D7000?
The better the glass, the better for the super high pixel density of these cameras. I will tell you that you'd probably enjoy the enlarged manual focus and the 3-4 stops of image stabilization you'd get with the E-M5.
Jordan -- That's really some unbelievable IQ wide open at f/1.8, and nearly so at f/2! So far, I'm really impressed with Olympus' exposure and color balance starting with the E-LP1 on to this E-M5 model. You really know how to get the most out of your twilight and night colors. Great stuff -- thanks!
Jman13 wrote:
Did some quick low low light testing to find the highest 'usable' ISO. 6400 is the highest that is usable IMO for general use. I think that you can get good size prints from 6400 with plenty of detail and relatively high color fidelity. I found ISO 16,000 to be the highest you can go before you start getting heavy magenta color shift, which gets bad at 20,000 and worse at 25,600, though can be usable for web or small prints if processed right. 16,000 could be used for small prints or web use when needed. The shot below printed at 4x6 shows very little noise in the print with lots of detail. 5x7 I think would be fine at this ISO and 8x10 might even be OK if you don't mind some noise in the print.
This is ISO 16,000 in very dim light (stray light at night from the next room over) with some Lightroom noise reduction to taste (balance noise and detail), which I find about luminance 40 to be good for this high ISO.
Yeah, not too bad at all. That 16,000 looks like the GH1 at 2,000 or 2,500 but with a finer grain - which usually means easier to clean with NR tools. Schweeet!
So does this make the OM-D the king of low-noise for µ4/3 ? I think it does yeah!?! And I wonder how far it's pushing into the APS-C world at that... It's certainly better than any APS-C camera released prior to 2010 I think.
Jman13 wrote:
The better the glass, the better for the super high pixel density of these cameras. I will tell you that you'd probably enjoy the enlarged manual focus and the 3-4 stops of image stabilization you'd get with the E-M5.
Apparently IBIS doesn't work as well with long lenses, probably because the relative amounts the sensor would have to be moved by go up rapidly, due to the magnification. I don't know if 3-4 stops would be realistic with a 280mm lens. Testing would show.
Wouldn't the sensor movements be smaller for a longer lens?
carstenw wrote:
Apparently IBIS doesn't work as well with long lenses, probably because the relative amounts the sensor would have to be moved by go up rapidly, due to the magnification. I don't know if 3-4 stops would be realistic with a 280mm lens. Testing would show.
Joseph Marney wrote:
Wouldn't the sensor movements be smaller for a longer lens?
Gut feeling says the opposite.
When comparing IBIS to optical stabilization systems you can often read (the opinion) IBIS is better for shorter focal lengths while long tele lenses are better stabilized optically.
I say "opinion" as I don't know of many (if any?) controlled test.
Joseph Marney wrote:
Wouldn't the sensor movements be smaller for a longer lens?
I think they would be larger. Generally, your hands would move a certain amount when you are being unstable, and this could be measured in degrees turn of the camera. A 1 degree turn of the camera, for example, would turn into a much larger shift in the image for a long lens.
Jonas B wrote:
Jordan, have you done any reasonably controlled tests taking images at faster shutter speeds leaving the IBIS on?
Controlled? No. But I have gone out and taken pictures. I haven't turned IS off yet, and there were no issues at any normal shutter speed. Tack sharp at 1/2000s, and 1/4000s with IS on.
As to longer lenses, it might be slightly less effective, but I have been able to get consistently sharp shots with my 40-150 at 150 and 1/25 sec, which is around 3.5 stops, which is mighty impressive for an IBIS system.
I might (depending on work) get to try it out with my 50-300L tomorrow. Due to focal length changes, I'll leave it off most of the day, but I'll turn it on at 300mm and let you know how it fares.
Joseph Marney wrote:
Wouldn't the sensor movements be smaller for a longer lens?
Larger. Longer focal length a given angle subtends a larger fraction of the sensor than for a short focal length (i.e. the FoV of a long focal length is much smaller than for a short focal length). Thus for a given angular movement of the camera the sensor will have to move more with a long focal length lens.
Jman13 wrote:
I might (depending on work) get to try it out with my 50-300L tomorrow. Due to focal length changes, I'll leave it off most of the day, but I'll turn it on at 300mm and let you know how it fares.
I gave it a try with a 500/8 reflex and it was definitely at its limits at that point. Not much difference on or off for exposures. However, focusing was much, much, much easier with the stabilized live view!
Yeah, having a stabilized viewfinder and IS this good makes this the first camera I can honestly say I'd rather have this than a bunch of optically stabilized lenses (except perhaps for those that are really long like your 500mm....a compact 1000mm FOV...gotta love it )
Jman13 wrote:
The better the glass, the better for the super high pixel density of these cameras. I will tell you that you'd probably enjoy the enlarged manual focus and the 3-4 stops of image stabilization you'd get with the E-M5. ...
I might (depending on work) get to try it out with my 50-300L tomorrow. Due to focal length changes, I'll leave it off most of the day, but I'll turn it on at 300mm and let you know how it fares.
I am very interested in your results. Can you try to manual focus with the 300 as well, just to see if this is managable at all?
Since I will be shooting wildlife with the long lens, I don't need 1/25 of a second, it'll be rather between 1/250 at low light and 1/1000 and up anyway. Wonder if the IBIS will still help at higher speeds, even with a lens that long. Or will it just plain give up or even worsen things, because it isn't able to follow the movement and make jerky adjustments.
I like your out of life tests. Very usefull! thanks for posting them and keep it up!
kwalsh wrote:
I gave it a try with a 500/8 reflex and it was definitely at its limits at that point. Not much difference on or off for exposures. However, focusing was much, much, much easier with the stabilized live view!
ok, that's interesting. 500 is rather long though, when doubled. With that 50-300 it would be possible to figure out up to what focal length the IBIS will reduce any shake
