I like the dynamic range of RAW files, but I don't want raw sensor data, but I don't want 8-bit jpegs. What can I do to get the desired result?
You should use the most neutral RAW processor to preserve the best qualities of an image. There is no other option: the last cameras capable of writing 16 bit TIFFs are discontinued - for a reason.
There are two programs which I have personally verified to preserve as much data as possible:
- dcraw. With this program you may exclude any processing alltogether (option
-D) or enable just the debayerisation (option -d). You may also set the native white balance to preserve channel data which gets cut otherwise (option -r 1 1 1 1 -W) but the handling of WB will be different.
- RawTherapee. This program is just as good as dcraw in preserving image data (with correct settings) but also includes better debayerisation algorythms.
Both of them may not have the colour profiles which you would like but RawTherapee handles DCP and ICC input profiles.
Other programs (including Adobe programs and programs from OEMs) may introduce colour shifts and undesirable effects.
Your comments reveal that you want OEMs to make cameras output 16 bit TIFF images. As I said, there are no such cameras produced any more, this is not the feature which customers liked and wanted to pay for.
RAW output is superior in every way except requiring additional workflow steps if you SPECIFICALLY want raster TIFF in the middle of your workflow (for which I know no reason). Uncompressed 16bit TIFF file of a 24 MP image would be 144MB large and would offer no advantages over a RAW file combined with RAW processor of choice. It happens so that people are fine with using additional software to get the best results.
You may try asking OEMs output 16 bit TIFFs but I wonder what will be the result if there are still OEMs which do not output DNG.
DNG may be both 16 bit raster and totally RAW depending on what camera puts in it. No camera puts 16 bit raster data into DNG.