Discussion in 'Panasonic Cameras' started by Manu-4Vendetta, Feb 14, 2013.
GH3 gets DxOmarked! Is on par with the E-M5. | 43 Rumors
I think that was expected because GH3 is rumored to be using same sensor. The only interesting thing I see in the E-PM2 has higher score for Low light ISO than both GH3 and OM-D. How is that possible? Isn't E-PM2 using the same sensor as the other two cameras?
I have a problem with DXO numbers. How come that 3 cameras with the same sensor are getting a gap of more then 100 in the ISO. In other two testings the numbers are marginal and can be caused by implementation or different calibration of test equipment.
Does any body knows what is their methodology.
I guess it will not be just the sensor, you will have to do something like this processor and developed each camera, the sensor itself will not have the same behavior in a Sony and an Olympus.
The biggest difference is between the PL5 and the OMD, they re both Olympus.
I am quit sure that Olympus did not rewrite the all ASIC for the PL5, PM2.
The PM2 is even better. take a look at the attached link
DxOMark - Compare cameras side by side
One thing that hasn't been verified is if the OM-D and GH3 actually use the same sensor.
If you look at history, Panasonic has always implemented the "premium" sensor first in their products with Olympus following suit sometime later.
So "if" the OM-D and GH3 have the same sensor, why would Panasonic give it to Olympus first?
The OMD and GH3 use a Sony sensor.
Panasonic has a problem in their high end sensor development. We need to see what sensor Panasonic will introduce in their next line of cameras especially the GX2 and G5 replacement.
Where has this been verified?
Okay...Now I'm done...
President of Oly, last July.
Olympus E-M5 sensor made by Sony [updated/corrected] - 1001 Noisy Cameras
(among other places)
(Never say "never" again)
A lot of rumors go around in the camera world but one that I found interesting was that Panasonic is apparently going full steam ahead and making good progress in developing a Foveon type non-Bayer sensor. Now THAT would be epic if it ever became a reality and was implemented in m43 bodies and we retained all of the other nice things we are used to like fast AF etc.
There was a news when 2 months ago, Sony is investing and taking part of Olympus shares, due to Olympus scandal, Olympus needs money to run their business. 2 weeks later, there is another news Olympus will make some lens for NEX system.
I think this is why Olympus gets Sony's sensor.
PS: the news was heard from the short news on Asia TV.
Many Olympus M4/3 lenses reviews from DXO is using Panasonic GX1, GX1 sensor only scored 55, I'm wondering if all these Oly's lenses are using OMD, will they get higher better score?
The relation must be longer then this. It takes at list two years to develop a sensor and about 1 year to develop camera. Sony have decided to develop the :43: sensor at list 2-3 years ago. I do not believe that they have developed it before they had a contract with Olympus.
Still no verification (from Panasonic themselves) as to what they're using in the GH3?
None that I know of yet, but when a Panasonic product manager (or engineering head, I can't recall) was asked point blank if it was a Sony sensor at a press conference, the answer was "no comment"
Which if you ask me is pretty much the same thing as confirmation :tongue:
There's also other tips like the fact that it's almost identical performance-wise to the OM-D sensor, right down to showing purple flare and reflections - which no other m4/3 camera has done until the OM-D EM-5.
I wouldn't be surprised if GH3 does use the same sensor as E-M5. There aren't many manufacturers out there to buy from.
I knew my E-PM2 had way better Image Quality(tm) than my E-M5. Good thing I sold that old tech.
But seriously though, I think the lesson here is that a difference of 100 in the DxO testing is small enough to be within their testing error. When you look at their SNR curves and see that they are all pretty much on top of one another, it's safe to assume that the ISO SNR is, for all practical purposes, the same.
The other thing I've learned from looking at lots of DxOmark curves is that when the curve changes direction at high ISO like the E-PM2 curve does above (see how it follows the same trend as the others but then separates more above ISO 3200), that is either measurement error or on-chip smoothing (cooked RAW like the Nikon 1 Series). No sensor is going to really kick in and give you something extra in the super high ISO range.
Looking at this SNR graph we see that for each jump in ISO we get 3db in SNR.
This is nice since 3db means that the SNR value is doubled which conform is doubling the ISO.
In their explanation regarding this test they state "A difference in low-light ISO of 25% represents 1/3 EV and is only slightly noticeable."
From this I can conclude that ISO test unless gives a big delta is not that important. It got large error margins and the numbers unless have a large gap have small significance.
Yep, I give it the eye test and don't consider anything significant unless I see ~3db separation. Also, SNR should degrade at a similar or higher rate as ISO increases, otherwise there is testing error or on-chip smoothing at play. The biggest problem with their summary scores is that they give credit for on-chip smoothing. May as well test the in-camera JPEGs instead of RAW if you're gonna do that.
Separate names with a comma.