the same is with ACR/LR - in its processing workflow/path (codewise), raw data is demosaicked as soon as possible and color transformed into a regular colorspace (colorimetric - isn't that the word ?) just with linear gamma...
Hold on, you are still not clear enough for me to understand. Are you saying that as soon as I import the image into LR (or open in ACR), the full resolution raw data is demosaiced? And from that point on, if I want a 1mb or 30mb file, the ACR engine takes this processed (or particularly processed data) of full resolution data and then provides the size/color space/bit depth I've asked for?
that happens before you see any rendered (from raw data by ACR) image on screen... and then, from that point on, it is not different from working with .TIFF opened in ACR/LR
Again, I'm having difficulty with perhaps our native languages. When you say
open, do you mean the preview seen as soon as I select "
Open" on a raw file for ACR OR open when I expect ACR to provide the data size, color space, bit depth etc shown to me IN Photoshop?
Perhaps if you outlined how you think the processing path operates I'd understand better:
1. Select Open in ACR
2. ACR access raw data and builds (you have to fill in the rest)
3. Parametric edits from user are applied
4. User exports 1mb image in sRGB for web (you have to fill in the rest in terms of what happens next).
Here's what I *think* happens and I'm wondering if you feel this is true or not:
I '
open' a raw file in ACR or import into LR.
A small preview is generated (in LR we have preferences to select the size).
I edit the image by building a list of parametric instructions. I suspect there is no reason this has to be applied to anything but this small preview but maybe you are saying the engine has processed data using the full resolution raw data.
The raw data file is
accessed and with the edits I've built, an image is now rendered as I desire. Size, bit depth, color space etc.
Where above am I off?
If this sounds correct, then the where we are not communicating clearly is the very last sentence above, where it
seems that the raw data
plus the instructions have to both be accessed so rendering can take place. If that's the case, then the raw has to be used here no? Or are you saying the raw data was processed the very first time I accessed the data (or build a preview), that data lives somewhere waiting to be rendered? It seems that would make a huge overhead in processing and storage just to view a raw you might not edit.