DPReview has finally posted some goodies from their Olympus OM-D/ E-M5 and they are impressive. Indeed, the camera outperforms the Panasonic GX1, G3, and GH2. The noise levels are apparent, with the Oly outperforming its Panny cousins by about half-a-stop, and in areas like contrast, saturation, and greys, the new Olympus really shines. It's much better than the GX1.
Unfortunately, the tests don't tell everything on their face. The Olympus was recently tested, revealing that its ISO settings were incorrect. ISO-200 was actually ISO-120, and all further ISO settings were similarly inaccurate. This means that when the new Olympus performs as well as the Sony NEX-5n at ISO 3200, it's actually performing that well at closer to ISO-2000. This is pretty bad ISO inaccuracy, certainly the worst that I've seen recently, and is either a gross accident or it was done purposely by Olympus to game comparisons.
But to say that the new E-M5 underperforms the APS-C as regards noise is a "duh" statement. The sensor is smaller. What I'm curious about is color depth and and dynamic range. It's for this reason that I await the DxOMark tests, which generally align to what I experience when manipulating RAW files. Moreover, even with the ISO inaccuracy taken into account, the new Oly outperforms its direct competition from Panasonic, while being about even with the Canon G1X.
Still, none of this jibes with Focus Numerique's comparison where the E-M5 noticeably underperformed the NEX-5n, 7, and the Canon G1X. That test was ISO-to-ISO, meaning that the ISO of that E-M5 was probably accurate. If you check the EXIF data on the photos from FN, you'll find that they are all f/5.6 with a shutter of 1/125 for the ISO-3200 test. This is a better comparison than the DPReview tests, since those tests assume that the ISO levels are generally accurate.
I await the DxOMark score, because it will tell us whether this sensor is indeed the same one as found in the G3/GX1, which I suspect. Here's hoping that dynamic range gets a boost, because I can deal with the noise as long as the range of sensor is increased.
No comments:
Post a Comment
All posts are moderated, so it may take a day for your comment to appear.