I always found it kind of strange of how every person on the ship knows voice and face of every other crewmember. While fixing it completely seems impossible to implement properly, i suggest changing how people hear masked crewmembers with stolen ids.
In next examples let's us X as original identity, Y as stolen identity
========
Example1:
Scientist stole engineer's id and is now walking around masked. He speaks something, every person sees his phrase as "X as Y says:", even assistant which just arrived on station and never seen or heard neither X or Y guy. Scientist gets in trouble.
How i see it work with new system:
Scientist stole engineer's id and is now walking around masked. He walks into a bar in with bunch of people sitting there, doing bar stuff. Scientist orders drink and since nobody in a bar neither is from science or engineering departments, they see his phrase as "Y says:" If there was another scientist in bar, he would see impersonator's speech as "X as Y says:" since he remembers the voice and is able to see stolen id.
========
Example2: Guy from engineering is a syndicate agent with QM assasination mission. He steals identity of cargo technician, and moves into a room with another technician and CE which is there for paperwork/ordering BSE parts/etc. Syndicate agent asks that cargo technician about location of QM, and both CE and that cargo tech percieve that guy differently.
Since Y works in cargo, cargo technician knows how people in his department sound like, and he hears agent as "Unknown as Y says:"
CE knows his people, and sees stolen id, which agent is wearing and recognizes him by voice and hears him as "X as Y says:"
========
So with new system, how you hear masked person changes depending if you are:
In same department as impersonator, you will hear "X as Y says:" since you remember voice of your colleagues, and see his id, you understand who is it.
In same department as impersonator's target, you will hear "Unknown as Y says:" since you remember voice of your colleagues, and you understand that this is someone else using your colleague's id.
Security with voice recognition device (from now on VRD) in your pocket , device will constantly correlate speakers's voices around you to their equipped id's data. If inconsistencies are found, device begins to *beep*.
Neither in same department as X or Y, you will hear impersonator as "Y says:"
Implying that arriving people are late for work, this system should work on them as well.
========
Then:
Captains would know voice of every head of department, voice of NT rep, their blueshield
Heads would know voices of each others plus NT rep and blueshield.
Heads would know voices of people in their departments.
HOP would know voices of every civillian role plus people from service department
Since janitors are silent eyes of station, they would know voices of every crewmember on station.
Same applies to mime. Chief knows voice of IAA.
Everyone would know annoying voice of clown.
Everyone knows voice of their department's head as well as captain's voice.
Noone except for HOP will recognize voices of civillians, making sneaking easier for them for drawback of having no access, and not knowing any voices.
Civillians don't know any voices beside for captain's and hop's (maybe add bartender to this list)
========
About VRD(voice recognition device) i mentioned earlier:
In order for security forces to be still a threat for impersonators, i want them to have some kind of device, simular to universal voice recorder in size and design, but different by function.
What it does is it will constantly record samples of people speaking around device owner, correlate voice to id card equipped, and if there is something wrong it will *beep* loudly.
(Maybe add function of adding problematic sample to little storage, which will be shown during transcript print with fluff information from id card and problematic statement itself, but i think that would be op, since it will make stolen id unusable, due to security knowing that it've been stolen.)
EMP's will break VRD's completely, making them always ignore problematic speakers, and beep randomly anytime someone talks around it.
I imagine detective having one as well, which will have its own skin, and will work against syndicate voice changers too.
In turn, syndicate voice changers won't trigger usual VRD's, and will get ability to chameleon into different types of masks, and for badassery, change into blood-red cigar.
======== I hope i won't screw up images.