So, my understanding is that the colloquial use of RT60 is incorrect because sound bouncing in a small room does not become effectively random before it totally decays. You can still identify individual reflections in the Energy Time Curve for most of the decay time. What you have in small rooms is 'modal decay' and not reverberation. The sound field needs to be totally jumbled into randomness for it to qualify as true reverberation.
RT60 is defined in terms of real reverberation and
critical distance so it doesn't technically apply to small rooms where the direct sound is louder than reflected sound everywhere.
I think this is a real and important distinction, but what I'm not sure about is how this changes the recommended approach to treating a small room. I think you still want to control decay time and modes regardless of what you call them.
Since I was criticised earlier in this thread for using the RT60, and I invited my critic to post his reply with no response, I guess I'll give my explanation on RT60. RT60 has two components - "reverberation" and "time to decay 60dB after EDT". Neither makes sense in a small room.
First, the 60dB decay requirement. EDT is the "Early Decay Time", which is the rapid decay in sound after the speaker's impulse. It is excluded because it is not room reverberation since the sound hasn't had time to travel and reflect. Also, the speaker itself may still be producing sound, e.g. cabinet resonances, driver ringing, etc. It is conventionally defined as 5ms. The 60dB requirement is problematic because the typical noise floor of a listening room is about 40dB, meaning that the speaker has to sweep incredibly loud (> 100dB) if we want to observe a 60dB decay. Fortunately, SPL decays linearly, so we measure the time to decay by 20dB or 30dB and extrapolate it to 60dB by simple multiplication.
"Reverberation" refers to multiple overlapping room modes. ALL wavelengths form room modes, but short wavelengths form thousands of them and they overlap so much that the modes even out and form a reverberant field. We are interested in the decay in the reverberant field. As the wavelengths get longer, the modes start to separate out until we can see them distinctly below the transition frequency. So if you see a huge peak in the RT60 below the transition frequency, you are not looking at decay, you are looking at a room mode.
For this reason, in small rooms, we drop the "reverberation" and "60dB decay" requirements and use the T20 and T30. Both T20 and T30 are acceptable substitutes for the RT60, but only down to a certain frequency defined by the transition zone. I calculated the Schroder frequency to help the OP determine his transitional frequency (4x Schroder). It is important to remember that this is an approximation.
It is difficult to look at decay on a waterfall graph if we are looking at low frequencies. This is because peaks create longer decay, and nulls shorten them. This is why I flattened the peaks using my DSP software so that it is easier to look at the decay. Even after doing this, it appeared that bass decay was still prolonged. This is why DSP can help a little with LF ringing, but it can not ameliorate it. What
is effective are large pieces of furniture.
Re: critical distance in a small room. It is untrue that the direct sound is louder than the reflected sound "everywhere". In small rooms, reflections are earlier and louder so this
shortens the critical distance.
I hope this clears it up, and if I am still wrong about something, please pull me up on it.
www.lydogakustikk.no
