As a home player, I have used only cheap dynamic and large diaphragm condenser. I guess people here have experience with other.
 
I try to build more or less functioning internet conference system. And I have not seen even half way acceptable solutions. The only usable devices I could observe was from Polycom, with one extender per table. And custom build systems with personal mics.
 
What I want is sound pickup from the whole room. For the moment ~6x8m but next project is for ~8x15. The is not for music. It just has to pick the speech with constant level on output.
 
I currently test 2x large diaphragm condensers throw DM for gate, EQ and compression. The result is almost satisfactory, remote end get acceptable signal from any position in the room.
 
But there is one aspect I want to improve: the upper part of the room is almost covered by metal tubes, the room was constructed as a labor... Not only that horribly resonate and produce echo, the cooling system is super noisy (air moving throw uncovered metal tubes), especially at warm days.
The mics I use (at the moment not covered at all) take all sounds, including the noise, corridor, steps. So I am thinking about a possibility to limit the direction and/or distance.
 
I will appreciate any advise. May be another mic type, may be some kind of caves for existing. Modifying the room (I mean cover tubes and beton walls)  unfortunately is not an option.

Polar pattern is what cheaper large diaphragm microphones lack. Using the existing microphones, you can try alleviating some of the reflections by mounting the microphone fairly high on the wall (about 2/3 up from the floor) with the logo pointed downward toward the far corner. This will maximize the reflection length to what is coming into the front of the condenser so that a gate could hopefully shave off (most of) the rest. The back of the condenser is the bigger challenge, and would need some sort of shielding to attenuate the shorter reflections hitting the wall behind it. For most phone systems, the microphones drastically attenuate anything that is not in a 300Hz-3.4KHz band, so doubling or even tripling the EQ to shelf off everything outside that range may also help with reverb (if a gate is not adequate). The biggest hurdle is to shield the back of the microphones. I am pretty sure that PolyCom's expansion microphone are condenser, but the housing is rigid plastic, so the primary sound entry is through the front vents on them. Something similar might work, but I cannot think of something simple offhand (something rigid with a foam spacer mounted on the back side of the microphone).

Thank you for the reply. I will try high positioning mics.
Which polar pattern would you recommend for my scenario? I am using cheapest mics just to check the concept is working in general.
 
 

You might try a dynamic mic.  I have been very pleased with my purchase of the Heil PR 30B Large-Diaphragm Dynamic for my church choir.  Very reasonably priced ($249), good gain before feedback, coverage, and rejection.  Use code SUMMER18 for 15% off.
 
Is seating around a table or lecture style?

Is everyone sitting at a table or is this an open room with just chairs?  If everyone is sitting at a table, you might try boundary microphones.  They are designed to pick up the sound hitting the surface and are common in conference type settings.  For the smaller room you might need 2 or 3.  For the larger room you will need more depending on the size/layout of your table.
 
Also, I believe the Polycon type systems use boundary mics.

fwrend
You might try a dynamic mic.  I have been very pleased with my purchase of the Heil PR 30B Large-Diaphragm Dynamic for my church choir.

I guess the choir produce more loud and consistent sound. Does that mic has reasonable output from a single person speaking normally from 3-5 meters?
 

  Very reasonably priced ($249), good gain before feedback, coverage, and rejection.  Use code SUMMER18 for 15% off.

I am in EU... PR30 is €350 here. The price by itself is ok, but (since not private) ordering is complicated and returning is almost impossible. I will have to find a place to test it first, the nearest store does not have it. Otherwise, if I have success in the first room, I think I can waste some money testing equipment for the second...
 

Is seating around a table or lecture style?

People re-organize tables. Normally that is lecture style, 10-40 persons. But sometimes they make a round table from the first 1-2 rows, for 10-15 persons. The same in the second room, just 1.5 - 3 time more people. But the second room is acoustically "normal", with usual office plastic ceiling and no constant noise.
 
Fortunately, the sound quality does not really matter. And current system seems like does the job even in case the person has turned his head 180° away from the mic (I plan to make real test next week, so far all tests was with 2 persons in the room changing positions and one another on remote end). If I can manage to limit vertical angle to ~20-30°  (exclude the floor and ceiling) and/or horizontal angle to 45-60° (exclude the biggest noise generating corner), it will be almost perfect (for the purpose). 

If the room was less problematic, the simple solution would be a LDC with "omni" polar pattern.  (Boundary mic would be similar)
 
Since the room itself is full of reflections and negatively affects the sound, you don't want to capture the "room as a whole". 
 
Given your circumstances, I'd look toward what radio stations use when dealing with multiple individuals in a room.
RE-20, RE-320 (modern take), and SM7b are very popular for this purpose.
All are large diaphragm Dynamic microphones... and will capture less of the room.
The RE-20/RE-320 are popular because they're extremely rugged... and don't color the sound a lot if the source is off-axis.  If you want to capture as little of the room as possible, go with the SM7b.
 
Many folks (in less than ideal home studios) use the SM7b precisely because it doesn't capture much of the room.
 
All three mics are useful for many recording tasks (vocals, kick, guitar/bass cab, etc).

So I see that large diaphragm Dynamic is not only a guess but also the industrial approach for my case. And I see SM7B in the local store. RE-320 is in another authorized store (RE-20 is 2 times more expensive). I have also found PR31 and PR40 in authorized store. In case my test setup is accepted, I will order something from that list.
 
Also thanks for mentioning boundary mics, I have completely overseen they exist (mentioned in "Instrument mics" section). They are less attracting me at the moment. It seems like, as with Policom, they are good when the table configuration is fixes as round/C-like, there is a possibility for cabling in the middle  and the number of tables is not large. Also I am not sure how good that works with speakers near/inside ceiling, so effectively pointing toward the table surface. But I can give Beta 91A a try on yet another location.
 
Thanks! I knew I can learn a lot there
 
BTW the controlling part of the test setup:

* I have mentioned "Not a Mouse" controller in Sonar forum, but there is no interest (I use it as cheapest and smallest remote controller for REAPER). In this configuration it is switching microphones (will be wireless mouse laying on the first table...). But the next setup will use IR/RF remote to the display (loosing PC dependency, much longer remote battery life).
* The display is self made from ESP8266 and MAX7219. Electronic part was connecting 5 wires, no soldering required. Communicates with the mixer (XAir) throw WiFi/OSC. LEDs are well readable from 5-10m, getting attraction by "scroll" animation on mode change. I have compared the result with a dedicated small LCD monitor and personally have found that solution better (especially in this self cut wooden box). Such display cost ~$10, use USB for the power (and initial programming) and can work with any OSC capable devices/software (X-Air mixers, RME Totalmix, REAPER, Sonar with AZ Controller) after several lines code modification (words to display and OSC addresses to use). 
 

The RE-320 is similar to the RE-20... but uses a Neodymium magnet (hotter output than the RE-20) and has an upper-mid presence boost. 
 
The SM7b was used for the vocals on MJ's Thriller.  

Dynamic mics are great for noise reduction, but that also includes the front. When you say "constant level" from a room, you can literally scream into a dynamic and not peg it at 1m, so at 3-4m things can easily drop off the radar.

As far as polar pattern on a condenser, cardioid is the most common that will give you the least grief from sound entering the back (non-emblem) side.

http://www.akglife.com/po...irectional--technology