E-M5 has DR 13!

Started Apr 13, 2012 | Discussions thread
OP Sergey Borachev Veteran Member • Posts: 3,936
Re: Mini DR test here

Timur Born wrote:

Here are DxOmark results versus TechRadar results both scaled to the same graph, dotted lines are DxOmark.

  • The TIF values measured by TR are considerably higher than the values measured by DxOmark!

  • There is a plateau in the TR reading at low ISOs while the DxOmark readings go straight down.

  • Once TR and DxOmark run appr. in parallel the DxOmark values are around 0.75x the TR values.

  • The GX1 and E-P3 values between TR and DxOmark do not differ by any practical means!

My interpretation is that TR's TIF values do not account for the NEX-7's non-linear raw curves (compression). And taking this as a basis I suspect that the E-M5 may also use a non-linear approach.

Going one step further I looked at the part of the E-M5 curve where the TR graph declines in a fashion more similar to what we see in DxOmark graphs and calculated the values with a x0.75 multiplier. This would leave the E-M5 at 8.625 EV for ISO 3200, a value higher than the NEX-7, very close to the "screen" (per pixel) value of the D800 and after subtracting some doubt-offset generally in a more "real-world" area.

Doing similar math at ISO 200 results in about 11.6 EV for the E-M5. I assume this will be a bit less in practice.


I am so confused but I was thinking that the enhancement from such compression/decompression or some other way Olympus or Sony might have used to get 14bit like data representation out of 12 bits should still result in more usable DR in the final image and is therefore a benefit. No?

Post (hide subjects) Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow