High end "echos" during test sweeps

[Sound_plan]

So, I have REW setup so that it sends to my DAW, and my DAW sends the signal back to REW, this allows me to make EQ changes to the sine sweep, and see how those changes affected the graph it records.

I know the ideal thing to do is to treat the room and not do corrective EQ, but I unfortunately need to go the EQ route.

What I'm noticing, is that when the sweep gets up in the higher end, idk, maybe around 10-13k? And up, it kind of loops back, and I hear the sine sweep, only in that range, repeat, as the original climbs up to the top.

Is this normal? I would guess not, but I also can't figure out why this might be happening.

 

[John Mulcahy]

Usually caused by something in the chain operating at a lower sample rate and aliasing.

 

[Sound_plan]

Usually caused by something in the chain operating at a lower sample rate and aliasing.

Thanks for the reply. You think it's a samplerate thing? Would oversampling usually fix that sort of thing?

So, I'm using reaper as my DAW. REW appears to have output set at 48khz, I usually work at that sample rate, because I make videos, and YouTube prefers it. But I went into the project settings of the tab I was using for REW, and I set the project samplerate to 48khz. Then I made sure the audio device settings were set to 48khz also, and that the interface was set at 48khz, but I'm still getting the same thing.

I only have the one EQ plugin in the chain, which has no settings on it. It doesn't have any oversampling features. After I changed all the settings to make sure they were all at 48khz, I opened a new instance of the EQ.

Any ideas of what else I could check for?

EDIT: something else kind of weird, my EQ is picking up some 10khz-20khz noise that my ears can't hear when I monitor the input. So, that looks like it's a steady noise stream coming from REW somehow. Idk if that be a clue at all or anything.

EDIT, another thing occured to me, I'm seeing higher up harmonics of sorts when the sweep sweeps, but, it should just be that one pitch where the sweep is at, which seems odd to me. It does seem like it's passing through something. Could it be the routing software I'm using? VB-Audio Virtual Cable?

 

[John Mulcahy]

If you are using the Java drivers REW has no direct control of the sample rate, it sends data to/receives data from the OS and the OS will resample it as required. On Windows the device will run at the rate configured for it on the Windows audio properties, Advanced tab, default format. Using ASIO drivers on Windows means REW can directly control sample rate. On macOS the sample rate will be the rate configured in Audio Midi setup.

You also need to make sure there is no input monitoring, if any of the input gets routed back to the output it creates a loop that means what gets captured is a summation of multiple delayed copies of the stimulus and it is only fit for the bin.

 

[Sound_plan]

If you are using the Java drivers REW has no direct control of the sample rate, it sends data to/receives data from the OS and the OS will resample it as required. On Windows the device will run at the rate configured for it on the Windows audio properties, Advanced tab, default format. Using ASIO drivers on Windows means REW can directly control sample rate. On macOS the sample rate will be the rate configured in Audio Midi setup.

You also need to make sure there is no input monitoring, if any of the input gets routed back to the output it creates a loop that means what gets captured is a summation of multiple delayed copies of the stimulus and it is only fit for the bin.

So, Windows will use 44100? I am using VB audio cable. I think I'm going to try and set everything to 44100 for this, since that's the most likely standard for anything I might not set.

For reaper though, I need input monitoring to be on, because that's how the sound is going to leave out of my speakers, but the microphone should not be being used by reaper at all. The microphone should only be used by REW to make the test graph. So, I don't think I'm getting any sort of feedback like that.

I am getting some harmonics up in the really high end though that build as it sweeps up, and then you get the bounce back, it looks a lot like aliasing as you say.

 

[Sound_plan]

I tried replacing VB audio cable with another program the same guy made, which seemed to be another audio routing piece of software but I think somehow compatible with asio, or marketed as higher fidelity. I'm honestly not sure of the difference, really, but, I tried that, and the original problem I was having went away, and was replaced instead with some clicking, that looked like just a high shelf in my EQ. The clicking was teh sort of clicking sound you get like if you make cuts not at zero crossing. I also discovered the little bit of noise I had up in the high end must have been sort of truncating noise from differing bit depths. When I made sure everything was in 24 bit, it went away.

 

[Sound_plan]

If you are using the Java drivers REW has no direct control of the sample rate, it sends data to/receives data from the OS and the OS will resample it as required. On Windows the device will run at the rate configured for it on the Windows audio properties, Advanced tab, default format. Using ASIO drivers on Windows means REW can directly control sample rate. On macOS the sample rate will be the rate configured in Audio Midi setup.

You also need to make sure there is no input monitoring, if any of the input gets routed back to the output it creates a loop that means what gets captured is a summation of multiple delayed copies of the stimulus and it is only fit for the bin.

You were exactly right! It took me forever to figure out how to get VB cable to be consistent for sample rate with everything, but I finally got it lol. I'll explain it in case anyone else finds this that could use the info.

So, I needed to operate on 24 bits for starters, that part was pretty easy. Next up, I needed to set things up so that VB cable registered the same samplerate on input, on output, and internally. Now, in java mode, as stated above, you can't alter the samplerate. So, you've got to have 48khz. So, you set vb cable to be internally to 48 khz. I had major issues trying to get all 3 to match up, if I changed one thing, another would change, but I discovered that you can go into control panel/sound/advanced/ and find the vb cable "device" and from there alter what the output is. so, you can select 48khz and 24 bit from there. Now, if your DAW is set to 24 bit 48khz as well, REW will be using that as well, and you're off to the races!

Thanks for your help!

 

[John Mulcahy]

Good to hear. For others' benefit, it isn't necessary to use 48 kHz but the rate in the default format of the advanced audio properties for the devices being used should match the rate in REW.

 

[Sound_plan]

Good to hear. For others' benefit, it isn't necessary to use 48 kHz but the rate in the default format of the advanced audio properties for the devices being used should match the rate in REW.

Oh, I misunderstood. I thought you meant in java driver mode you couldn't alter the sample rate REW was using.

At any rate, no pun intended, I think I had trouble getting vb cable to get 44.1 on the input. Maybe I just fudged it up. I got them all the same eventually though, is the important part, as you say.