Home Theater Forum and Systems banner

1 - 6 of 6 Posts

·
Registered
Joined
·
4 Posts
Discussion Starter #1
Have untreated spare bedroom with only a desk and desk -- for voiceover/tutorials. Room has hardwood floors with a rug, bare walls.

Just purchased a dbx RTA-M mic and used it with my laptop speakers (buying a small pre-amp to run speakers from laptop today). For setup, I ran XLR from the mic into a Focusrite ISA one and then Scarlett 2i2 for the USB interface. Mic was on a stand at ear height placed about a meter from desk (toward center of room). In REW, I was able to get enough gain for receiving -12 to -20 db on the input ribbon gauge.

The resulting waterfall, RT60, and spectrograph are attached. It appears from the attached waterfall image that for a sampling length of 500 ms, the sound never diminishes and rather just bounces. Obviously since the room is untreated (empty spare bedroom, hardwood floors) - it is presently a disaster for sound. Is there anything wrong with the waterfall, indicating measurement issues? What does the RT60 plot indicate for the decay, and can you make out anything from the spectrograph as well?

Last, it seems that room EQ for voiceover (not mixing, home theatre) presents an issue when sitting at the desk, using laptop, and recording while generating tutorials. What is recommended for RTA mic placement and speaker placement to EQ a room for voiceovers? The issue is that in order to duplicate the recording process, the speaker should really be positioned where you mouth is, and the RTA mic needs to be where a typical voiceover mic is located (e.g. 4-6 inches from mouth after a pop filter). Any suggestions?
 

Attachments

·
Registered
Joined
·
4 Posts
Discussion Starter #3
Thanks for the response. I finally got my laptop sound card output to run through a pre-amp then to stereo speakers, and then re-made the calibration measurements with mic positioned ear-height at an equilateral triangle to the speakers on the edges of the desk. In the attached waterfall plot, I didn't limit the bottom to 35 db but you can see the pattern that emerged -- whihc indicates a low fre problem below 200 Hz. It also looks like the spectrograph tells the complete picture, which is that the lows don't decay or attenuate but rather continue to bounce off the walls.

I am thinking about possibly installing a GIK room kit or Primacoustics London 10 or 12?
 

Attachments

·
Premium Member
Joined
·
6,711 Posts
The low frequency part of that measurement is just noise, pushed up in level by your soundcard cal file which is not a soundcard cal but a room measurement. Best start by removing the soundcard cal file from the REW Soundcard preferences and using Clear Cal on your measurements to remove it from there, then you'll get a better view of what your measurement actually shows.
 

·
Registered
Joined
·
4 Posts
Discussion Starter #5
Okay, I selected clear cal (there wasn't a sound card cal file linked to REW) in the sound card tab of preferences. I think I have to remeasure and just ensure I don't use levels or cal in the soundcard tab of preferences. I really only want to use the "measurement" button on the upper left(?), and then save to disk.

Something I noticed when taking some measurements was that I once made a measurement for which the top peaks of the waterfall were all near zero db and the bulk of data were shown a negative db in the range -80 to 0. What causes all the waterfall peaks and troughs to be negative? Is there anything being done incorrectly or set incorrectly when the top peaks at near zero db?
 

·
Registered
Joined
·
4 Posts
Discussion Starter #6
These REW measurement results should probably work(?). The SPL plot is in 1/12 octave, and the Waterfall is in 1/6 octave (600 ms). I was reading about the spectrograph and recalled it's the result of an FFT, so the plot is just a power density heat map. From about 2700-4200 Hz in the spectrograph there are three major spikes, and since the room is 13' x 13', is this a room mode?

What about distortion and RT60? In the distortion plot, is there an integer multiple that breaks away at higher frequencies? Is there a long decay or resonance apparent causing voice echo?

What frequency do I have problems at -- for which traps or absorption are likely helpful. Would a diffusor be helpful, given these data?

Thanks in advance!
 

Attachments

1 - 6 of 6 Posts
Top