Are there multi-voice speech recognition libraries?

Question

C++ / C#

Good day!
I’ll start with an approximate statement of the problem: speech recognition of many voices (up to a dozen voices per session, most of which were not heard in previous sessions, and most likely will not be heard in the next ones), they will speak in turn, simultaneous speech will be rare and can be neglected. They will speak in Russian (60-90% of the time), recognition of foreign speech is optional.
The session lasts up to 4 hours, strictly indoors, the number of microphones on which recording is made is from one to 5-6. Some voices may be quiet, especially at a distance from the microphones, and vice versa, loud comrades may be near the microphones. Extraneous noise: city noise, indoor noise (coughing, calls, etc.).
There is no and will not be access to the Internet, the program will work on the desktop (mainly Windows 7).
Conversion to text is not necessary directly from the microphones, after the end of the session, recordings will be available (by the number of microphones). Record processing time is not critical, but if possible no longer than the duration of the record itself.
So far I have chosen PocketSphinx, but maybe there are options that are less laborious and more suitable for the task?