Does Aperture always apply white balance to RAW files?

Question

Hi Aperture users,
I have been an Aperture user for many years.  Aperture handles the white balance of my Canon XTi RAW files just as I want it to: it interprets the RAW data by applying the white balance selected at the time of the photo.  I.e., when Aperture decodes the RAW file, it makes it look more-or-less like the in-camera JPG with respect to the white balance.
However, in the past several months, many well-known contributors of this forum have stated "RAW is RAW -- Aperture will not apply any camera settings."  This statement was made in response to many "novice" questions about why a RAW photo in Aperture did not look anything like the JPG from the camera.
So, my questions:
Does anyone have a camera to whose RAW files Aperture does not apply in-camera white balance settings?
Does everyone else agree that this exception to "RAW is RAW" is acceptable?
thanks,
nathan

Answer 1

Hi Nathan -- I was hoping someone much more informed than me would weigh in with some engineering truth.  All I got on offer is the thin gruel of observation.
Mr Endo wrote:
So, my questions:
Does anyone have a camera to whose RAW files Aperture does not apply in-camera white balance settings?
Does everyone else agree that this exception to "RAW is RAW" is acceptable?
1. I don't.  RAW is sensor data.  Most cameras record the White Balance camera setting as a metadata tag attached to the RAW data  (The WB has no effect at all on the captured data.)  Most RAW converters use this setting as the default WB.  Some RAW converters perform "colorimetric interpretation" -- they basically compare the data captured to a generic color space and assign a WB.  (This is, I assume, the same thing a camera set to "Auto WB" does.  My camera is superb at assigning a WB for any mixed scene; it's repeatably bad, though, whenever the actual scene is truly colored (meaning "away from average").)  In both cases -- 1. using the attached-to-the-RAW-data WB, and 2. analyzing the captured data to intelligently predict the WB -- a WB is applied to the RAW data.  It has to be, in order to produce an image from the data.
2.  The problem here isn't that this is an exception -- it's that different people mean different things when they say "RAW".  The RAW data is RAW data.  As above, the WB setting has no effect on it.  But the image derived from the data -- any image produced from the data -- much have a WB set (and, for human use, a tone-curve applied).  Some people mistaken think that RAW is an image format, or that the image produced by the RAW converter is a "RAW" image.  It is not.  RAW data must be converted to an image format in order to be displayed as an image.  The conversion must apply a WB.
I have found it useful (though perhaps mistaken) to conceive of my data-gathering to picture-making as following these steps:
. Capture a pattern of light.  This is just data.
. Convert the data to an image.  RAW conversion done by Aperture.  (NB: the settings for this can be changed.)
. Adjust the image to my artistic satisfaction
. Convert the adjusted image to an image format file
. Publish the image-format file in one or two now standard forms: to be viewed on a computer monitor, or as a print.
I mention only because at some point I consciously added the "Convert data to image" step in order to better grasp the process.  RAW data is RAW data -- but that step identifies an intermediate that many overlook.  At that step the information recorded along with the RAW data and stored in the RAW file is applied to the RAW data.
To me, "RAW is RAW" is correct.  "Aperture will not apply any camera settings," is wrong.
I am happy for any clarifications or corrections.