I started writing this post asking 

Are there any headset/mics that do a good job of muffling or quietening my voice to prevent it from disturbing others?  And possibly also provide me with a bit more privacy?

E.g. asking if anyone had tried to do speech recognition using the many motorcycle helmets that contain speakers and microphones, e.g. to do speech control of a GPS system, possibly with a heads up display inside the motorcycle helmet visor.

While I haven't seen anybody asking that question, I eventually found  other posts talking about   trying to reduce the amount of noise that a user of speech recognition inflicts upon others in the neighborhood. Things such as a stenographer mask -  although the consensus seems to be that stenographer masks have lousy sound quality, if for no other reason than your breathing is constrained, and are quite uncomfortable. The consensus seems to be per people like Lunis that it is better to have a headset/mic that you can talk to very quietly.

Too bad.

 I'll post the rest of my  already written message below, but I suspect that most of it is moot.  Although if somebody has tried to use a motor cycle helmet with speakers and mics for speech recognition with Dragon, and not just GPS,  I'd like to how good or bad it was.

---+ DETAIL (TL;DR)

Are there any headset/mics that do a good job of muffling or quietening my voice to prevent it from disturbing others?  And possibly also provide me with a bit more privacy?

E.g. I imagine that a fancy way to do this would be to have ANC both for my outgoing voice as well as for incoming sound.  Although passive noise cancellation, i.e. just muffling via a plate or mask in front of my mouth (i.e. in front of the speakers's mouth), might be sufficient.

E.g. I know that there are many motorcycle helmets with speakers and microphones, to permit communication between riders, phone connectivity, and more and more voice control features.   Some wired, some bluetooth.  Some helmets allow speakers and microphones to be installed that are not actually sold by the helmet manufacturer.  This might allow one of the higher end speech recognition mics to be used,

(I sometimes joke about wanting the "motorcycle helmet of silence", riffing on Maxwell Smart's "Cone of Silence".)

Q: has anyone tried using such motorcycle helmets with speakers and mics for speech recognition with a PC ... with Dragon?

Note: I am not asking about speech recognition while on a motorcycle at speed.  I am asking about wearing such a motorcycle  helmet with speakers and  mic in a stationary environment. I'm sure it's uncomfortably hot, and probably quite embarrassing. Nevertheless, might be worth a try. Especially if could  cut away  most of the helmet.

 as I was writing this post I eventually found other posts in the KnowBrainer forms on

Stenomask https://talktech.com/stenomask-steno-sr/

https://talktech.com/shop/stenomask-store/sr-pro-dragon-bundle/

While stenographer mask microphones (cover your nose and mouth) are extremely quiet and won't disturb coworkers, they are expensive, difficult to use and feature less than desirable accuracy because of how you have to control your breathing. We can only recommend Martel or Sylencer stenographer mask microphones after exhausting all other options.

We recommend speaking softly into a hands-free USB microphone. Our preference, in this situation, is the speakerless FlexyMike DEC with a USB soundcard.

From <https://www.knowbrainer.com/forums/forum/messageview.cfm?catid=6&threadid=30688>

We truly believe that our SpeechWare FlexyMike DEC + SpeechWare USB MultiAdapter will pick up the softest speech with maximum possible accuracy.

From <https://www.knowbrainer.com/forums/forum/messageview.cfm?catid=6&threadid=30688>

 I am not sure that  speaking softly into the SpeechWare FlexyMike DEC + SpeechWare USB MultiAdapter  would solve the privacy problems, although it might  reduce annoyance of one's neighbors to an acceptable  level.

---+  BACKGROUND (TL;DR)

Returning to using speech recognition this summer, after years away, has greatly helped my "computeritis" pain.

But: speech recognition has also greatly limited where I can use my computer. Leading to greater social isolation, etc.

---++ Social isolation

For example:  for much of the last ten years I have worked from home, remote from the rest of my team who are a long way away in different states, and sometimes different continents.  The remote work is not so bad, text communication essential, email, slack, etc., but even with phone/video I can go days without talking to people in person during the workday.

Like many, I found that going to a restaurant or coffee shop helped.  Even if I don't know the people, the other customers, around me - although the waitstaff grew to know me by name.  Also, several folks nearby work(ed) for the same company as I do(did), in related areas, but not on the same team, and were in the habit of working at coffee shops, where I would join them. Plus, there are several Meetups in the area for remote workers - folks get together, work on their own stuff, but talk to each other over breaks.  Occasionally sharing advice, e.g. about programming problems and other computer issues, if company secrecy allows it.  (Since my work recently is more open source, I have fewer constraints in this respect than in the past.) 

But, using speech recognition has two problems when using it outside of my private home office:  (a) my voice disturbs others around me, and (b) in public I am much more worried about other's overhearing sensitive information, whether company or personal private, than I am typing.  Especially with a privacy screen.  (Nano-drones with cameras will make typing less private, but still...)

Similarly, even when there was a company office where I could sit in a cube, I would be reluctant to use speech recognition because it disturbed my neighbours.   In fact, this is why I gave up using speech recognition circa 2002 -   my cubicle neighbors complained.

---++ Family time(?)

Similarly, before I (re-) started using speech recognition, I would often work in the living room on my laptop while my wife was working on her lesson plans and my daughter was working on her homework.  Understandably  the sound of my speaking to my computer distracts them.

(It's not exactly a family activity like playing a game together, but it's better than nothing.  There is often occasional conversation, me helping them with their computer problems, etc.)

---++ Working while waiting for  food/coffee

Before I (re-) started using speech recognition, I was often in the habit of going to a restaurant and reading my email while waiting for food to arrive. And similarly, working on my PC at the dining room table while waiting for the water to boil for tea  or  cooking something in the oven.

Again, this is something I am less likely to do, for fear of  annoying other  members of my family at home,  or  annoying other customers of the restaurant  and/or speaking aloud sensitive information.

---+ Working on an airplane or train

Enough said.

DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design.