You can't use a sweep outside the test room nor does the ISO 18233 standard state that. They measured those levels at microsoft because the noise floor of a mic is not constant and they make lab mics that are that quiet. The standard method relies on 1 octave or 1/3 octave RTA, so there is no noise canceling that can be achieved. There is no magic the user can do to fix this, you can't get rid of that noise. My experience has been that noisy mics are just not adequate (same with SPL meters). My walls were built to an STC 80 spec, that means I need a 120dB sound source to reliably detect that above the noise floor of the receiver room. If you can't measure to below the noise floor of the receiver room, then you would need the sound source to be VERY loud to be loud enough to be detected over the mic's noise. Cheap gear has a very high noise floor, typically unable to measure much below 25-45dB depending on what we are talking about. I recently learned it can directly import a direct impulse measurement, so you can use a clapper or balloon pop.Īs for cheap gear, that isn't my experience. The Article I linked discusses the purpose of the impulse measurement and REW can absolutely do that. This is true of the AudioTools app as well however. I then calculate STC based on that relative to the STC curves.Īt the moment I’m still measuring noise at most frequencies above 300hz so the tests haven’t been successful enough to confirm everything is working. I export the values at each 1/3 octave value and the difference between the two is the TL. I also take 1/3 octave noise floor measurements of both to ensure I am never measuring just noise.
I take 1/3 octave RTA measurements for both. Then I move the measurement rig out of the source room and into the receiver room and measure again. I take a spatial average over about 30 seconds. First I measure in the room while running the mic out of the room and wearing hearing protection. I then use my laptop with a Motu 828x and a NIST traceable mix with 15dB of self noise. I need this because playing 120dB for long periods of time introduces compression after a while. One is the generator and the other is the remote. I use a mix of pink noise and pure tones. I use two iOS devices with audio tools for the test signals. I’m still having problems but am closer than I’ve ever been. I finally had to breakdown and spend a small fortune on new hardware because the transmission loss of my partitions was so high that I was unable to produce a loud enough test signal to capture that signal through the wall. I’m curious what hardware you are using to do this. I’ve been developing my own protocol for doing this using REW and an external excel worksheet I created. This paper provides the procedure you need to follow and puts AudioTools method into context. Click to expand.It’s not automated but it can be done manually.
0 kommentar(er)
