Step response?

[moedra]

Does anyone know why our step response looks like this? I’m helping someone measure his room and this is not something I’ve seen before.Usually, the step response just goes straight across but here we’re getting a rise in the response that doesn’t make sense to me. What is causing this and what does it mean? What effect will this have on the measurement and subsequent processing actions? Is there something wrong with the microphone?

 

[John Mulcahy]

There is a DC offset on the input, select the High pass option in the Input options section of REW's soundcard preferences to exclude it.

 

[dcibel]

DC or at least very low frequency noise, like LP rumble.

 

[moedra]

Thanks guys. So that could mean the mic is doing it, right? We used a UMIK-2. Tried measuring in two machines and got the same result.
Is there a way to remove the offset from an existing measured response like we have, or do the measurements need to be redone with the high pass filter on the input?

 

[John Mulcahy]

The measurements would ideally be redone, but you could apply a high pass filter on the EQ window, save the predicted response as a measurement using the action near the bottom of the filter tasks pane then do your analysis on that. It is very odd to have an offset with a UMIK, 1 or 2, that should be reported to miniDSP.

 

[sm52]

but you could apply a high pass filter on the EQ window, save the predicted response as a measurement using the action near the bottom of the filter tasks pane

Could you show the equalizer window settings for this to work?

 

[John Mulcahy]

There's not a lot to show, it's just a high pass filter.

 

[potzie]

There's not a lot to show, it's just a high pass filter.

View attachment 58409

hi, sorry if this a dumb question, but what version of REW has this fancy EQ filter window? I have been following a YT tutorial which also has this, but i cant find any verson online with it.

 

[moedra]

hi, sorry if this a dumb question, but what version of REW has this fancy EQ filter window? I have been following a YT tutorial which also has this, but i cant find any verson online with it.

V5.20.14 early access build

V5.20.14 has moved to a new look and feel, FlatLaf, for most components, to modernise the appearance and make it easier to adjust. That involved a lot of changes, as many REW interface components were heavily customised, so I'm making early access builds available to help pick up any interface...

www.avnirvana.com

 

[sm52]

John Mulcahy​

I tried using a high pass filter. The Step Response of some measurements ended near zero with the 11.5 hertz filter. But in other measurements, I went up to 31 hertz to get the same result. I understand that this is not necessary. It was an experiment. If it is not possible to measure the driver again using the high-pass filter checkbox, what is the maximum filter frequency in the equalizer window, after which it makes no sense to correct the measurement in this way?
Second question. Why can't the filter be set below 10 hertz?

 

[potzie]

many thanks

 

[John Mulcahy]

I tried using a high pass filter. The Step Response of some measurements ended near zero with the 11.5 hertz filter. But in other measurements, I went up to 31 hertz to get the same result. I understand that this is not necessary. It was an experiment. If it is not possible to measure the driver again using the high-pass filter checkbox, what is the maximum filter frequency in the equalizer window, after which it makes no sense to correct the measurement in this way?
Second question. Why can't the filter be set below 10 hertz?

The less low frequency extension a response has, the sooner its step response will decay. To deal with DC offsets it doesn't make sense to use anything above the minimum frequency offered by the filter.

The 10 Hz minimum for the EQ filters was chosen to exceed the range of equaliser offerings (most do not allow frequencies below 20 Hz) and avoid numerical issues which can occur at high samples rates and high filter orders with very low centre frequencies.