1. Read the disk in DTC preservation format
2. Convert track 0 to an FM sector image
3. Convert the rest to an MFM sector image
4. DD this to a single mixed sector image
5. Use a C program to recreate a DMK image
This is neither completely automatic nor close enough to the original formatting of the disks for special cases such as copy protected disks which use fake CRCs, custom sector sizes, or otherwise unreadable data to identify the original software. The DMK format should be able to support any of these but I'd need a way to recover the original FM/MFM stream of bytes from the KryoFlux stream files. The sector image format looses all meta information which I have to recreate in a standard way. I've done this only for the format actually used by my disks but the TRS-80 operating systems supported a plethora of variants thereof.
I would dream of built-in DMK file support in DTC, but a raw FM/MFM read mode would help to create my own converter. This is what the KryoFlux is all about, isn't it?
WayBack link to the DMK file format description by David Keil (the original page is long gone )
Ira Goldklang's TRS-80 Revived Site
Marcus von Cube
The only exception to this rule is G64, but even for that DTC generates a custom, extended G64 file that holds more information than a normal file does.
There is no plan at the moment for formats specific to certain emulators only, unless it really benefits many KryoFlux users (like G64 does).
However, it is planned to have a serialized format specifically for FM and MFM disks that collects data from a track during analysis; this is especially useful for comparison, but I think it's exactly what you need as well.
You could either use this serialized track data which will be obviously easier to process than flux reversal signals (what DTC does) - once available. The content will be very similar to what you'd have via the various Track Read commands on FDCs like the WD-177x, except for the fact that the original track data won't be altered, like e.g. the Address Mark detector modifying certain bitcell sequences on WD chips.
Alternatively, you could use the stream files directly as well, but correctly translating flux reversals into data is not trivial to get right.
Should you choose to use the planned serialized format, your feedback would be more than welcome on exactly what you'd need.
Is IPF a supported output format? I was thinking it is meant for authoring or archival purposes in order to reproduce salvaged disks as closely as possible.
IPF files generally require an authoring tool, so dumps submitted for preservation to SPS will have "official" IPF files. There is an alternate encoder out there though written by one forum member here for the purpose of creating the file itself from a stream dump - obviously not authoring the content, but for "home use".