Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Chris Grimshaw]

Hi ladies and gents,

I was thinking about sound equipment earlier (I believe it's a disease), and had a thought - could we just do a sine sweep measurement of the monitor using the vocal mic (ie, everything in position), look at peaks in the frequency response, and knock those down with EQ, instead of the traditional ringing out the monitors?

Here's an example - two frequency response curves. Speaker is a Yamaha DBR10 in its monitor position, at the base of the mic stand, mic is an EV N/D967. REW is the software.

One of the traces is just the mic on a stand, near a reflective surface (think small room stage near a back wall), the other trace I've put my hand flat over the end of the mic. You can probably guess which is which.



Now, when I increase the level of the mic going to the monitor, the blue trace gives me feedback at 535Hz (according to my phone's RTA app). When I've got it so it just rings at ~500Hz, and then put my hand over the mic, it screams. Feedback everywhere. My phone shows lots of peaks from 2.7kHz upwards. Even just going near the mic with my hand starts things ringing. Looking at the traces, that's to be expected - there's up to 15dB more sensitivity in the kHz range when you put your hand over the mic.
Looks like we can directly correlate gain-before-feedback with the frequency response of the monitor "seen" by the mic.

Here's another one - the third trace is with me standing by the mic, holding my breath (can't put any "me" signal into the measurement).



Things haven't changed much - the 500Hz peak has gone to nearly 600Hz, but the rest follows pretty closely, so I think its safe to take readings without anyone at the mic unless you want to wring every last bit of SPL out of the monitors.

The sine sweeps don't have to be particularly loud to get useful readings, just so long as they're comfortably above the noise floor. Next time I have a gig and plenty of setup time, I think I'll try this. It seems much better than standing at the mic turning the levels up, getting a load of feedback, reading out the frequency and taking it out on the EQ.

Thoughts?

Cheers
Chris

[Scott Helmke]

Some people like to use SMAART between vocal mic and wedge for just that purpose. Knock down the peaks as seen by the vocal mic.

[Kevin Maxwell]

I use Smaart, put pink noise into the monitor, use the vocal mic as the measurement mic (aim it at the monitor) at a little bit below mouth height. You can put it closer if you want. Be sure you are not putting the mic back into the monitor while doing this. The problem with EQing monitors without exciting them with some sort of sound source is it leaves a lot of things out of the equation. I take an output of the console from the solo bus so I can easily route a mic to Smaart for this purpose. Besure that your channel EQ is flat or bypassed when doing this. Don’t get carried away and always check to see how they sound to you. I have found that even though it may not be sticking up in the measurement I usually have to take a bit out around 200hz or 250hz or there is too much low end in the monitor, probably due to the proximity effect of the mic that doesn’t show up in the measurement. I don’t know why but it seems to work best if you start your cuts from the low frequencies to the high. Also if after doing this it doesn’t sound right flatten everything out and try again. Sometimes you get carried away on the first pass and you really need to be more conservative when you do it again. BTW I have found that pink noise varies depending on what pink noise device you are using, it is not necessarily flat. So an RTA is not a good choice to do this with. But when using Smaart where you are comparing the input to the output it works.

[Kevin Maxwell]

I forgot to mention, a good quality EQ is very important and the better the monitors the easier it is to get them to behave. 

[John Roberts {JR}]

The non-flat frequency response is not the cause of feedback but contributes to which feedback frequency will dominate first.

The several frequencies where feedback will occur for a given physical alignment are determined by path length, and transit time (speed of sound in air), where the multiple transits over the same acoustic path reinforce the sound arriving at the mic.

If the entire path gains at those specific feedback frequencies are >1 the signal will continue to build with each transit and run away.

Eqing your path to be more flat may help reduce feedback modes that fall within response bumps, thus the advice that better mics and better monitors help. A perfectly flat system will still feedback given enough gain, just not as soon as one with response bumps on top of the feedback modes.

JR

[Art Welter]

..could we just do a sine sweep measurement of the monitor using the vocal mic (ie, everything in position), look at peaks in the frequency response, and knock those down with EQ, instead of the traditional ringing out the monitors?

Even just going near the mic with my hand starts things ringing. Looking at the traces, that's to be expected - there's up to 15dB more sensitivity in the kHz range when you put your hand over the mic.

Thoughts?

Chris,

You can do as you have, but it will not account for the different face, mouth and hat cavities, and other nearby reflective surfaces that don't have the same directivity index as your hand.

To properly "ring out" monitors requires using the mic in all the (stupid) places it is likely to be put, while exciting the resonant frequencies with your voice of all the reflective surfaces. And then, stand by to find out that you can't reproduce everything the performer will manage to do, but after a while, you can figure out about 95% of the "stupid human tricks".

If the simplistic approach of sine sweeps worked even remotely well, we'd still be doing it  .

Art

[John Chiara]

Chris,

You can do as you have, but it will not account for the different face, mouth and hat cavities, and other nearby reflective surfaces that don't have the same directivity index as your hand.

To properly "ring out" monitors requires using the mic in all the (stupid) places it is likely to be put, while exciting the resonant frequencies with your voice of all the reflective surfaces. And then, stand by to find out that you can't reproduce everything the performer will manage to do, but after a while, you can figure out about 95% of the "stupid human tricks".

If the simplistic approach of sine sweeps worked even remotely well, we'd still be doing it  .

Art

Been trying to explain to some that you can't preset anything that has variable like different gains, different users, different mic and monitor placements...Etc. I wish....

[ProSoundWeb Community]