what is this in my REW waterfall high frequencies

[ThxDex]

Hello community,

i am a long time reader here and as well long time hifi and music enthusiast. But i have reached my limits and you might be able to help

What do i see in my waterfall diagram in the high frequency area??
It's not only noticeable in the measurement, i also hear it during the measurement sweep. It sounds like if the tone is bouncing from left to right in the room a couple of times.
Is that a "flutter echo"? Can that be in this high frequencies? Or something similar to a standing wave but with 10kHz ... (i know that is impossible, isn't it?) Or is my measurement mic defective?

My room is slightly treated with 9 pieces of 5cm "basotect" panels on the walls and 4 thicker diy absorbers. I know that there is room for optimization in the bass region, but that is not the topic here.
I recently moved into this room from an other one. The previous room showed exactly the same thing in the waterfall

Thanks for any input and greetings from Sweden!
Ingo

 

[DJBonoBobo]

Maybe some kind of feedback in your signal chain?
I don't know what it is, but i am pretty sure it is no flutter echo or something in your room.
I would check for possible errors in setup/settings/ software/drivers.
What microphone, interface and drivers do you use?
Oh, and do you measure one channel at a time or is this a stereo measurement? If that's the case, try measuring one channel at a time.
How long is your sweep?

 

[sam_adams]

It sounds like if the tone is bouncing from left to right in the room a couple of times.

That would be a flutter echo. Best viewed in the Impulse plot tab (Slapstick impulse response shown.):

The initial impulse is a resounding 'crack' from the wooden slapstick and the flutter echo sounds like the metallic 'twang' of one of those spring-type door bumpers as it bounces back-and-forth between the walls while it decays.

 

[Newman]

That would be a flutter echo. ...as it bounces back-and-forth between the walls while it decays.

But it doesn't seem to decay in the expected manner? Instead it maintains amplitude for what must be many bounces...

 

[Sokel]

These look like HF sources, like HVAC, motors, etc.
Strange.

 

[Newman]

These look like HF sources, like HVAC, motors, etc.
Strange.

Good point. Perhaps the OP can do a baseline run with the signal generator muted, and check that waterfall plot.

 

[sam_adams]

But it doesn't seem to decay in the expected manner? Instead it maintains amplitude for what must be many bounces...

Change the mode and settings in the Spectrogram tab and lets see when it decays:

 

[Flaesh]

The first thought was that it was noise. Accordingly, I wanted to suggest taking measurements with the speakers turned off. But then I saw some attenuation in the upper octave. This is the first time I've seen this.

 

[Keith_W]

Welcome to ASR!

It is 100% a measurement artefact and not an acoustic issue. The evidence is that it does not decay. Also, the noise floor of ALL listening rooms tends to be higher at low frequencies because (1) most noise generators like refrigerators and aircond are low-frequency, and (2) rooms are low pass filters.

I agree with the suggestion that it is some kind of feedback. It is also possible that sound is bleeding from the one speaker to the other. This happened to me once when I had my mixer settings wrong. What equipment are you using to take the measurement? Are you using ASIO or Java mode?

 

[Sokel]

I have seen similar measuring a room with the AC on.
It was a constant 17kHz non attenuating spike along with it's sidebands.

It also depends of the measuring length,increasing the FFT size to max available shows things that's not visible with the usual ones.
(warning though,some amps/speakers may suffer playing for several seconds down low or up really high)

 

[Pareto Pragmatic]

I have seen something like that from loose boards at the back of my room, like right back at the very back edge. They suck up energy and spit it back with a delay, for me starting at 15khz and then getting more delay up to 20khz. This is easy to check with some floor tapping, or weighting any suspect floor sections with something and doing a sweep.

If it is on one side of the room, that might explain the left-right bounce you are hearing.

Admittedly my floor is 100 year old maple, and there are many issues with it, not just a few loose nails. But I have a feeling that many engineered wood/vinyl/etc floating floors might show something similar under the right conditions.

I discovered this when I failed to turn off a humidifier on that floor, and failed to generate white noise before turning on RTA, totally going out of my power up sequence. Huge lump at 1800hz in the noise floor, but fixing that ALSO fixed my very much NOT late reflections above 15k. Which was quite a surprise.

I will eventually nail things down, but I don't want to do that when it is super dry. For now, bigger rug, rubber backed pad, and some changed objects to control things a bit. And always a full but turned off humidifier when I am listening!

---
I now see a room change happened. So what I said is unlikely, but still possible given similar flooring and installation.

 

[ThxDex]

Hello,

thanks for the replies - i played around and made some more measurements, pretty convinced it's a measuring error now.
I am using a windows 11 machine with the latest official REW, my DAC is an RME ADI2-DAC FS with the latest firmware - the microphone is an UMIK-1.
I use the Java mode in REW, if i switch to ASIO i do not get the UMIK microphone as an imput any longer - only the inputs from the RME DAC.

What i found out, is that i had the sampling rate set to 48kHz. When i change that to 44.1kHz the effect is almost gone. So i think we can rule out the room then.
(the RME DAC showd 44.1kHz on the USB port in the display when REW was set to 48, maybe there was some mismatch thing going on)

For me it's good enough to know it's nothing that i will hear in the music. Must be related to my REW setup, drivers etc.
I'll do some reading on the optimal measuring settings and so on, seems i lack some understanding there ...

Greetings, Ingo


original settings with 48Khz

with 44.1khz

 

[Keith_W]

I do not know enough about REW's JAVA mode to know how it gets around Windows mixer. I suspect it is some kind of WASAPI Shared mode which allows non-exclusive access to the mic and speaker, meaning that the signal can be corrupted by resampling or other applications accessing audio devices simultaneously. The fact that the artefact is greatly reduced (not gone) by switching sampling rates makes me suspect that it is some kind of JAVA or Windows issue.

I have only ever used ASIO, and for that you need an interface with DAC and mic preamp built-in. Perhaps someone more knowledgeable than me can chime in?

 

[sam_adams]

I am using a windows 11 machine with the latest official REW, my DAC is an RME ADI2-DAC FS with the latest firmware - the microphone is an UMIK-1.
I use the Java mode in REW, if i switch to ASIO i do not get the UMIK microphone as an imput any longer - only the inputs from the RME DAC.

What i found out, is that i had the sampling rate set to 48kHz. When i change that to 44.1kHz the effect is almost gone.

The UMIK-1 is a 48KHz sample rate device. That means that all devices need to set to the same sample rate. Additionally, you are using a USB dac and a USB mic to do your measurements. Each one of those USB devices will have its own clock. The clocks will not be in sync with each other. You may receive warnings from REW regarding timing errors when doing measurements. You should consider purchasing a measurement mic that can be powered from the RME's XLR connectors.

 

[Keith_W]

You should consider purchasing a measurement mic that can be powered from the RME's XLR connectors.

He has an RME ADI-2 FS which does not have analog inputs. The RME ADI-2 Pro does have analog inputs, but do not have phantom power so you can not use an XLR mic.

 

[sam_adams]

He has an RME ADI-2 FS which does not have analog inputs. The RME ADI-2 Pro does have analog inputs, but do not have phantom power so you can not use an XLR mic.

So that leaves the settings that still need to match for both devices and the fact that he will probably encounter USB clocking errors during measurement. Having chosen a device that does not have input capability is unfortunate, but the RME ADI-2 Pro's lack of phantom power on the input can easily be solved.