REW Noise floor impact on Frequency Response and Waterfall interpretation

[dwillis60]

I'm not clear exactly how REW handles the calculation of the noise floor as it relates to frequency response and the waterfall plot. Is the measured noise floor taken into account in the FR/waterfall plot? Or is this left to the interpretation of the user?

My most recent measurements suggest a FR peak and some ringing at 60hz, and the distortion tab suggests that this may be a function of something going on in the noise floor. Perhaps the 60hz mains frequency here in the states. When I ran a RTA analysis, the same 60hz peak was prominent and steady.

Based on the attached plots, is this a reasonable interpretation?

Thanks.

 

[Wayne A. Pflughaupt]

Not sure what you mean by “calculating” the noise floor. It will show up on the graph as what could be called “grass.” If you set the lower limit of the graph lower then you have it now, you’ll see what I’m talking about.

I don’t see any ringing at 60 Hz. In fact there really isn’t any significant ringing in your graph. But there is some low-level steady-state noise from something in the vicinity – typically 60-cycle noise from an appliance.

Regards,
Wayne

 

[dwillis60]

I wasn't clear if REW took noise floor measurements and then produced plots after controlling for that noise. In other words, the peaks are SPL readings minus the noise floor. Put simply, what you see is what you get, and there is no need to interpret FR or Waterfall through the lens of the noise floor.

You're right. There isn't much ringing there. But I see bumps and I get paranoid. And the further down you go in the Waterfall, the worse things appear. I try to be reasonable and set the lower limit at 40-50db just above the highest noise floor reading. Most of my SPL peaks are around 80db. So I'm assuming that a 40db range is reasonable.

I also don't want to get too focused on the Waterfall. My concern is also with the SPL/Frequency Response. Specifically the peak around 60hz. I'm assuming that the peak here is an actual peak related to room characteristics and not simply a transient peak related to the noise floor.

Thanks

 

[Wayne A. Pflughaupt]

In other words, the peaks are SPL readings minus the noise floor. Put simply, what you see is what you get, and there is no need to interpret FR or Waterfall through the lens of the noise floor.

Exactly. As long as you’re well above the noise floor, any function REW performs is meaningful.

That said, there is a function that does look at the noise floor specifically – I forget “where” it is since I’ve never used it. Perhaps John will comment.

I also don't want to get too focused on the Waterfall.

Beyond “FYI,” IMO waterfalls are mainly useful for determining the effectiveness of low-frequency treatments (such as bass traps), utilizing “before” and “after” measurements. Beyond treatments, taking care of the frequency response (via equalization or other means) will also take of the waterfall, as our moderator Matthew Poes is fond of noting.

My concern is also with the SPL/Frequency Response. Specifically the peak around 60hz.

Low frequency graphs are best analyzed using no smoothing.

I'm assuming that the peak here is an actual peak related to room characteristics and not simply a transient peak related to the noise floor.

As far as I know, there are no room acoustics issues affected by, or related to, the noise floor.

Regards,
Wayne

 

[dwillis60]

Exactly. As long as you’re well above the noise floor, any function REW performs is meaningful.

That said, there is a function that does look at the noise floor specifically – I forget “where” it is since I’ve never used it. Perhaps John will comment.

Beyond “FYI,” IMO waterfalls are mainly useful for determining the effectiveness of low-frequency treatments (such as bass traps), utilizing “before” and “after” measurements. Beyond treatments, taking care of the frequency response (via equalization or other means) will also take of the waterfall, as our moderator Matthew Poes is fond of noting.

As far as I know, there are no room acoustics issues affected by, or related to, the noise floor.

Regards,
Wayne

Thanks. I appreciate the insight.

 

[John Mulcahy]

Noise floor is captured at the beginning of a measurement and shown on the distortion graph, where it helps to distinguish distortion from noise. It isn't subtracted from anything.

 

[dwillis60]

Thanks. I think I'm catching on slowly but surely.