Unable to get timing reference for non-subs

[tjcinnamon]

If I run an acoustic reference for my subs (ch4), I get a distance/timing measurement. However, for all of my other speakers I get values close to 0 or slightly negative. I've tried 7 different acoustic references and all of them result in the same distance/timing issue.

I've tried the latest build of REW 5.20.14ea15 and the latest stable 5.20.13. I've tried with FlexASIO and ASIO4All. I've tried in pure-direct mode and standard multi-channel with and without Audyssey

Attached is an MDAT of measuring my left channel with the acoustic reference on each of the 7 channels (ch4 can't go high enough to be an acoustic reference). Then I included one measurement of the reference for the sub which yields a result.

I'd be surprised if all my speakers had bad impulse responses that couldn't be measured but perhaps there's a setting? Both Audyssey and Dirac were able to get accurate/similar timing results. I'm using a Umik-1 calibrated by spectrum labs. Although my room isn't ideal it's fairly well treated.

Any ideas on why my non-subs won't yield a time?

 

[John Mulcahy]

What are you expecting to see? Delays with an acoustic timing reference are relative to arrival of the reference at the mic. If all the speakers are properly aligned all the delays would be zero.

 

[tjcinnamon]

What are you expecting to see? Delays with an acoustic timing reference are relative to arrival of the reference at the mic. If all the speakers are properly aligned all the delays would be zero.

oh... I figured it would give me a number like the subs where it would be 16ms (or 8ms for my mains). How do I get distances with REW for non-subs?

For example, if I moved the mic 1ft to the left for my left speaker the timing should be 1ms difference (I think)?

I used Cross Correlation time alignment to vector average my subs (which was AWESOME!). I was hoping to do the same with the mains but the impluse responses looked odd relative to the subs.

 

[John Mulcahy]

For distances I recommend a tape measure

There isn't any way for REW to know absolute distances, it only knows the differences in arrival times relative to the reference. Timings will shift according to the change in distance relative to the distance to the reference. Cross-correlation alignment should work OK with your left and right, for the other speakers the reflections are so large the result may not be meaningful.

 

[tjcinnamon]

For distances I recommend a tape measure

There isn't any way for REW to know absolute distances, it only knows the differences in arrival times relative to the reference. Timings will shift according to the change in distance relative to the distance to the reference. Cross-correlation alignment should work OK with your left and right, for the other speakers the reflections are so large the result may not be meaningful.

Would acoustic treatments reduce these reflections? My room is carpeted and fairly treated with treatments, so I'm surprised there are so many reflections. Good to know that the numbers are normal and what to use for treatments.

 

[sm52]

tjcinnamon​

I looked at the mdat. All graphs are similar to each other 1 to 1. Are these graphs of different speakers?

 

[John Mulcahy]

Apart from the sub the latest set of measurements were all made using output 1 on ASIO4All, so they are all the right channel.

The "spl matching" set in the other thread have separate measurements from each speaker. You would need to track down the reflection sources in each case, some (e.g. the center channel) are very close to the direct signal so from something near the speaker. The channels 5,6,7,8 responses are more difficult to reconcile, particularly 5 and 6 which look to have some processing applied to them to produce such delayed reflections.

 

[tjcinnamon]

For distances I recommend a tape measure

There isn't any way for REW to know absolute distances, it only knows the differences in arrival times relative to the reference. Timings will shift according to the change in distance relative to the distance to the reference. Cross-correlation alignment should work OK with your left and right, for the other speakers the reflections are so large the result may not be meaningful.

so then just use a vector average to get a "sphere" of multiple measurements? Leave the timings alone?

 

[tjcinnamon]

tjcinnamon​

I looked at the mdat. All graphs are similar to each other 1 to 1. Are these graphs of different speakers?

https://www.avnirvana.com/attachments/spl-matching-mdat.58195/

 

[John Mulcahy]

What is it you want to do with the average?

 

[sm52]

tjcinnamon​

It seems to me that during the measurements, not one measured speaker plays, but two. And so it is in many measurements.

 

[tjcinnamon]

What is it you want to do with the average?

EQ the result to flatten it

 

[tjcinnamon]

tjcinnamon​

It seems to me that during the measurements, not one measured speaker plays, but two. And so it is in many measurements.

that is really odd! I'll take a look.

 

[ddude003]

Those front L and R speakers are rather an unorthodox design... And I would not expect them to perform well setup in the room as you have them...

Still wondering what your full signal chain, read kit, looks like...

 

[tjcinnamon]

Those front L and R speakers are rather an unorthodox design... And I would not expect them to perform well setup in the room as you have them...

Still wondering what your full signal chain, read kit, looks like...

They perform okay. They took a lot of finessing and I know they aren’t operating to their full potential. They are too far apart and I’m not far enough back for that distance. That said, I’ve done some absorbtion and angling and the SDA effect is definitely noticeable between on vs off.

2 channel music sounds way better in two channel vs up mixed and that wasn’t the case before.

Set up the long way they’d be too close to the side walls. I experimented with that and it didn’t add a lot to it. Ideally, my seat would be about 2’ back and I’d have an additional 5 feet behind me.

But the effect is noticeable when on vs off

 

[ddude003]

As I have asked several time across a few of your threads... Without knowing what you full signal chain looks like there is not much more I can offer other then suggesting you look at Serkan Gur's method of Inversion EQ...

 

[tjcinnamon]

As I have asked several time across a few of your threads... Without knowing what you full signal chain looks like there is not much more I can offer other then suggesting you look at Serkan Gur's method of Inversion EQ...

My signal chain is HTPC/AppleTV to Denon AVR to Anthem MCA325 to Polks

Denon has Audyssey MQX and I use REW to create the PEQ filters.

What is a read kit?

I’m all ears if you have ideas!

 

[tjcinnamon]

As I have asked several time across a few of your threads... Without knowing what you full signal chain looks like there is not much more I can offer other then suggesting you look at Serkan Gur's method of Inversion EQ...

I’ve been watching his YouTube’s

Is it filter 2 in this video:

 

[John Mulcahy]

EQ the result to flatten it

It doesn't make sense to do that on an average of multiple different channels, if that is what you were suggesting.

 

[tjcinnamon]

It doesn't make sense to do that on an average of multiple different channels, if that is what you were suggesting.

All the same channels. Like taking 4 measurement positions for the left channel and then averaging and EQ’ing the average

 

[ddude003]

What is a read kit?

AKA signal chain, kit, system...

I’ve been watching his YouTube’s

Is it filter 2 in this video:

The entire method... And not using Audyssey...

 

[John Mulcahy]

All the same channels. Like taking 4 measurement positions for the left channel and then averaging and EQ’ing the average

That may be problematic until you determine the source of the late reflections on some channels. Should be fine for L/R.

 

[tjcinnamon]

AKA signal chain, kit, system...


The entire method... And not using Audyssey...

I’ll give it a shot soon. It will be a bit. Any recommendation: equalizerAPO, Roon, or Jriver?

 

[ddude003]

There are a number of ways... Players that include a convolver like Roon & JRiver... Then there are players and DAWs that have VST, AU, AAX plugins capability like Audirvana+ or Reaper for instance, then there are convolver plugins and even some Reverb plugins have convolution engines imbedded in them... Lots of ways to mix and match... I have Roon, Audirvana+ and Reaper along with LiquidSonics Reverberate as well as Camilla DSP all work and will take the same FIR.wav files...

 

[tjcinnamon]

There are a number of ways... Players that include a convolver like Roon & JRiver... Then there are players and DAWs that have VST, AU, AAX plugins capability like Audirvana+ or Reaper for instance, then there are convolver plugins and even some Reverb plugins have convolution engines imbedded in them... Lots of ways to mix and match... I have Roon, Audirvana+ and Reaper along with LiquidSonics Reverberate as well as Camilla DSP all work and will take the same FIR.wav files...

Oh man. I used to be big into DAWs and VSTs. I was a big electronic music production hobbiest. It was all consuming (kind of like my quest for Audio/Video now)