TOPIC:

Newbie to this forum ... your enlighten is much appreciated.

Just got my UMIK-1, and I am not sure why I get unexpected high spectrum reading on low freq ( < 30 Hz ) ?

To reproduce the scenario :
1) Run the REW tone generator : 1 KHz
2) Run the REW RTA
3) Accumulated spectrum attached.

The RTA clear shows the 1KHz fundamental, as well as the integral harmonics.

What I don't understand is why are there so much energy at the low freq : < 30 Hz ?

@ RKwan

Not everything we don't understand mean that something is faulty.. Sometimes, It just mean understanding a bit what we're looking at.. Check out a little charge just pulled out from internet..



As you can see, even from a quiet room, 50dBSPL is actually quite common noise floor. Say you have your air con or some fan or some low noise rumble from outside, that would show a similar response. In other words, nothing to worry about as you'd always want to measure above your room noise floor. i.e. so you actually read the measurement of your speaker rather than the overall noise floor of your room.

Hoping this info helps and have fun!

DevTeam

Repeat the same experiment on a desktop ( last one was done on a laptop )
- with better speakers
- tone spl set at 75dB
- UMIK is about 4' from the speakers
- in a small room ( 10' x 10' x 7' ), full of "diffusions"



Similar result : much spectral energy at the low end ... almost as much as the 1Khz tone !!!

The reason that I want clarity on the low freq end is that I trying to characterize the acoustic of a hall : 88' x 50' x 30' , which has little sound abortion. The resonance freq of such room go down to single-digit-Hz.

I just wonder if my UMIK-1 is defective ?

PS : I have done the same experiment with Audacity + UMIK-1 ...
- similar result : unexpected high spectral energy below 30 Hz ?!
- Audacity can generate a 1Khz tone ( without the mic ) into the recording track, and the FFT on it is very clean.

@ Rkwan

Would it be possible that you maybe try to use our app note to get started?
Check out here "www.minidsp.com/applications/acoustic-me...mik-1-setup-with-rew"
That graph you just posted is unfortunately a bit useless as it's dBFS (i.e. zero reference to SPL). As mentioned before with the link (hopefully you maybe read a bit more on google to learn a bit more of SPL readings of a typical room), it's very common to have low frequency energy in a room. Your umik-1 isn't faulty.

That 1KHz is here because you most likely have tone generator enabled?

DevTeam

Hi devteam,

I followed the exact link that you provided to setup the REW + UMIK-1.

I am not sure why you said that the RTA plot ( in dBFS ) is useless ?
Does it not show the relative power level of the fundamental ( 1 Khz @ 75 dB ) against the unexpected low ( < 30 hz ) freq ?
- the 1 Khz was set at SPL = 75 dB
- I would suggest the low freq, with roughly the same dBFS in the RTA, would have roughly the same 75 dB

I understand that there is always the full spectrum of sound everywhere, the debate is at what level ?
My question : why is there so much at the low freq ... roughly the same level provided by the Tone Generator : 1 Khz @ 75 dB ?

PS : I tried checking/unchecking the Preferences/Soundcard/High Pass ... makes no difference.

RKwan,
I usually start any measurement session by recording a “noise floor” measurement, with the excitation turned off. With the signal generator off, you will likely see that the background acoustic noise is your low frequency culprit., as has been described by devteam. You will likely be able to hear this. With the tone generator on again, measure an averaged spectrum with say 20 averages, and your noise floor may be 12 dB or more lower.

I don’t recall how to do synchronous averaging with RTA, but devteam can perhaps pipe in with help.

Charlie

RKwan wrote:

Newbie to this forum ... your enlighten is much appreciated.

Just got my UMIK-1, and I am not sure why I get unexpected high spectrum reading on low freq ( < 30 Hz ) ?

To reproduce the scenario :
1) Run the REW tone generator : 1 KHz
2) Run the REW RTA
3) Accumulated spectrum attached.

The RTA clear shows the 1KHz fundamental, as well as the integral harmonics.

What I don't understand is why are there so much energy at the low freq : < 30 Hz ?

Look up the definition of pink noise and a 1/f spectrum....
Then compute the power difference between 100-Hz and 0-Hz

I use REW:RTA for the following scenarios :
1) Noise Floor : just the existing ambient noise ( looks 1/f )
2) REW : pingNoise @ 75dB ( full range : 10Hz - 20KHz, rather flat ? )
3) REW : toneGen : 1Khz @75dB

It seems to me that the UMIK-1 provides a constant 50dB between 2Hz - 20Hz, which I should have felt ... but I don't.
- is the UMIK-1 providing a segment of the spectrum that doesn't exist ?
- RTA calculation artifacts ?

Noise Floor spectrum


Pink Noise spectrum ( generated by REW )


1KHz spectrum ( generated by REW )

Do you live in a city?

Yes, I live in a city : Ottawa Canada.

As for environmental noise :
- I live in the suburb
- about 2Km from the highway
- 5+Km from railway lines.
- 15+Km from the airport
- the area is sandwiched by farms and natural swarms