DxOMark still silent on the E-M5

Started Sep 1, 2012 | Discussions thread
3DrJ Senior Member • Posts: 1,027
Re: Dxo is a questionable "organization" and . . .

Detail Man wrote:

The fact that the image-data in RAW image-files is typically organized within a structure that is called a "TIFF wrapper" is insignificant in relation to the nature of the actual image-data itself ...

The "wrapper" isn't exactly insignificant. It contains numerous headers defining how the data in one or more subtrees are organized and possibly compressed. It also contains info about camera settings, and so on. Many do contain proprietary information, that is, info pertinent to the particular camera. The headers are mostly in a predictable order, such as CFA arrangements. Otherwise, it would be impossible for non-manufacturer's converters to work for so many different RAW formats.

Conversion between RAW and TIFF data formats involves profound changes to that image-data including black-level subtractions, de-mosaicing (which may also include spatial-frequency convolution-filtering and sub-sampling operations), mapping of the image-data to a particular RGB color-space, application of RGB tone-curve transfer-functions (including shifted black-points), all of which can and do modify the output Dynamic Range and Signal/Noise Ratio of the converted data.

The fundamentally necessary change is demosaicing the Bayer grid into RGB channels. If it can be accomplished without scaling for brightness, color balance, etc., you'd have an RBG image in a viewable format, if not a satisfactory, finished state, but one reflecting the raw data.

The different image-data formats (even in non-commercial converters) are in no way equivalent ...

That doesn't make too much sense. As said, the stored raw data is "decoded" per info in various headers, which are mostly written in a way converters can use to decipher the specifics of the way the raw data was recorded. The raw data organization varies camera to camera but as such are "equivalent" but not "interchangeable" (which is probably what you meant).

One easily available, and free as well, that can do this is dcraw, written by Dave Coffin. dcraw can output 16-bit linear unscaled TIFF, PPM or PGM files, when the program is invoked with the appropriate switches.

Other than reading about such things, have you ever attempted to actually implement such ?

I don't understand the question. How does it matter?

Have you "implemented", written such program yourself?

With a bit of effort you can even find on the Internet methods to use a digital camera as an instrument to measure luminescence of light sources. The method uses software including dcraw and imagej, and when properly calibrated, a digital camera can perform measuring tasks otherwise requiring very expensive lab equipment.

Other than reading about such things, have you ever attempted to actually implement such ?

Have you?

I still don't understand what you are asking.

So there you have it, TIFF is just fine, and if you don't believe it, by all means look it up, check it out, be enlightened.

It could happen to anyone, particularly if/when they might gain more than a cursory understanding

Show us the code you've produced, then we'll talk...

JRA

 3DrJ's gear list:3DrJ's gear list
Olympus E-1 Olympus E-3 Olympus PEN E-P3 Olympus OM-D E-M5 Olympus E-M1 +5 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow