Help understanding group delay

[klettermann]

Despite my efforts understand GD I'm still baffled. So, rather than waste people's time here I invoked ChatGBT to get jump startged. That actually made it worse, so maybe we can start over. Here are a few nOOb questions.

• What scans do I take with REW? I can use REW just fine, but it's not clear what to measure. Individual subs (I have 2)? Individual mains? Everything everywhere all at once?
• Then how to interpret? ChatGBT suggested that just a little sub movement would do a lot. I'm sceptical.
• Subs are ported. Seems like some plug the ports. Is this really a thing, and how sealed is sealed? Really airtight or just block with foam or something?
• What else to be aware of?

To be clear, I don't even know if I have a problem. I used REW for EQing the systemn, XOing subs with mains and setting delays. Honestly it sounds pretty damn good, but the idea of leaving some performance on the table grinds me. Thanks for any and all comments, suggestions, links, tutorials etc etc. Cheers,

 

[Chromatischism]

Subs are ported. Seems like some plug the ports. Is this really a thing, and how sealed is sealed? Really airtight or just block with foam or something?

Your manual will give instructions on this. It is how they achieve variable port tuning.

For example for my Rythmiks, it will say if you plug one port, switch the sub to 14 Hz extension.

 

[Beave]

Group Delay Threads on ASR:

https://www.audiosciencereview.com/forum/index.php?threads/what-is-group-delay.30289/

https://www.audiosciencereview.com/forum/index.php?threads/group-delay-101.39333/

https://www.audiosciencereview.com/...y-vs-group-delay-common-misconceptions.39591/

 

[Chris A]

What scans do I take with REW? I can use REW just fine, but it's not clear what to measure. Individual subs (I have 2)? Individual mains? Everything everywhere all at once?

All measurements using swept sine upsweeps will generate a phase and group delay curve. So every measurement that you've saved will have group delay data.

Never try to measure more than one loudspeaker or subwoofer at a time.

If you've been exposed to any differential calculus, group delay is nothing but the first derivative of phase with respect to frequency. The group delay curve tells you the instantaneous slope of the phase curve at each frequency. It's not any more complicated than that.

It turns out that the human hearing system can actually detect group delay changes at levels above ~1 millisecond (the unit of measure for group delay is in time units) at frequencies above 200-400 Hz. That's why it's important to plot group delay at no more than 1 ms minor divisions at higher frequencies. I often see people plot group delay at units approaching 50 ms minor vertical divisions (which are totally worthless, except at frequencies below ~30-40 Hz).

For subwoofers that play very deep frequencies (below 30 Hz), you have to keep in mind the period of the sound waves (i.e., 1/frequency). For very low frequencies, very large group delay values tend to be normal, especially for subwoofers that do not have response very, very deep (below 5 Hz). "High passed" subwoofers and bass bins all show high levels of group delay. But you must use smoothing (I use psychocacoustic smoothing) to see and interpret phase and group delay measurements.

Measuring subwoofers is a bit more involved than full-range loudspeakers. The problem is that room reflections are an issue in reading phase and group delay at higher frequencies (above ~100 Hz), as well as setting the datum for phase. And one must correct the datum used by REW in each of your phase and group delay measurements.

Chris

 

[Keith_W]

Despite my efforts understand GD I'm still baffled. So, rather than waste people's time here I invoked ChatGBT to get jump startged. That actually made it worse, so maybe we can start over. Here are a few nOOb questions.

• What scans do I take with REW? I can use REW just fine, but it's not clear what to measure. Individual subs (I have 2)? Individual mains? Everything everywhere all at once?
• Then how to interpret? ChatGBT suggested that just a little sub movement would do a lot. I'm sceptical.
• Subs are ported. Seems like some plug the ports. Is this really a thing, and how sealed is sealed? Really airtight or just block with foam or something?
• What else to be aware of?

To be clear, I don't even know if I have a problem. I used REW for EQing the systemn, XOing subs with mains and setting delays. Honestly it sounds pretty damn good, but the idea of leaving some performance on the table grinds me. Thanks for any and all comments, suggestions, links, tutorials etc etc. Cheers,

Think of a race where there are fat people and fit skinny people. They all start off at the same time. But when they reach the finish line, all the fat people are grouped together and they finish later. It's the same with sound - low frequencies have a longer period, so they naturally take longer to be reproduced if they all start at the same time.

How linear-phase DSP fixes this: imagine that fat people are allowed to start first. After a certain amount of time, skinnier people are allowed to start, and the skinniest and fittest people start the last. At the finish line, the fat people and skinny people all arrive together.

How to measure with REW: take a sweep of whatever you want to look at with a static microphone (i.e. NOT an MMM). Every minimum-phase device has group delay - mostly subwoofers, but also tweeters. And even reflections, crossover components, and minimum-phase DSP. If you want to look at the pure group delay of a subwoofer with no room influence, you will need special procedures - maybe a ground plane measurement, CTA 2010, etc. An in-room measurement of a subwoofer still gives valuable information, but you have to make sure that the information is valid over a listening area by taking multiple measurements. Extract the minphase response (in REW, right click and "Min Phase Copy"). Then look at the "GD" tab in REW.

What it means: there are psychoacoustic detection thresholds for group delay. For e.g. if your subs are delayed by a minute, you would surely hear that. But what is the actual threshold? 10ms? 50ms? 200ms? The problem is, nobody really knows. There are studies with test signals and headphones, but most of these studies do not extend down to low bass and the detection thresholds seem to be shooting up exponentially. I would say it is easily about one period at whatever bass frequency you are looking at (i.e. about 20ms for 50Hz). This is my guess by extrapolating the data.

Ported vs. sealed: Sealed means airtight. Anything that is not airtight is ported. It doesn't matter if the port is opened or partially closed by stuffing it, it's simply a ported loudspeaker with altered port tuning. And, as you can imagine, you can have an unintentionally ported subwoofer if you construct it very poorly!

 

[kyuu]

First thing I would ask: are you experiencing an issue that you're trying to diagnose? Why are you worried about group delay?

 

[klettermann]

First thing I would ask: are you experiencing an issue that you're trying to diagnose? Why are you worried about group delay?

Thanks to all for the clarifications. They'll take a spell to get my brain around. Meanwhile, the above is really the most importent question for me. ChatGBT dragged me into all this. He/she/it/they "looked" at some GD scans and suggested that audible improvements were possible. I don't necessarily believe it, but I don't dismiss it either. So how to evaluate and decide? I'm happy to post traces here but I'm not even sure what to post or if was measured in a way that provides useful info. In any case, the system sounds pretty great to me as it is.

 

[thcdru2k]

In REW, open RTA, use Pink PN noise, set about 1/24 octave, and turn on the phase trace. While you adjust delay or the phase knob, watch the phase curve across the crossover band like 60 to 100 Hz. A straighter, more linear phase curve means lower, more uniform group delay, and a twisting curve means worse group delay.

 

[amirm]

Your main focus in measuring your room should be to correct frequency response errors, especially in low frequencies (below 200 Hz). There is no reason for you to worry about such things as Group Delay.

 

[klettermann]

Your main focus in measuring your room should be to correct frequency response errors, especially in low frequencies (below 200 Hz). There is no reason for you to worry about such things as Group Delay.

Thank you Amir, you have set me free! FR and sub integration seem fine. So I'm letting this go. Thanks and cheers,

 

[middlemarch]

Think of a race where there are fat people and fit skinny people. They all start off at the same time. But when they reach the finish line, all the fat people are grouped together and they finish later. It's the same with sound - low frequencies have a longer period, so they naturally take longer to be reproduced if they all start at the same time.

Absolutely the best description ever of GD! Hats off to you sir! If we could only get that into Wikipedia...

 

[Chris A]

Just a note about terminology used above...from this site:

In signal processing and electronics, the frequency response of a system is the quantitative measure of the magnitude and phase of the output as a function of input frequency.

So phase is important when talking about "frequency response", but the thresholds of audibility for group delay (the first derivative of phase) for very low frequencies remains one of the great unknowns in psychoacoustics.

The only reference to audible subwoofer group delay audibility seem to be from 1990--using headphones--not subwoofers: G. J. Krauss, “On the audibility of group distortion at low frequencies,” in Proc. Audio Eng. Soc. 88th Conv., Montreux, Switzerland, Mar. 1990.
___________________________________________________________________________________

The following is from page 420 of Toole's first edition (Sound Reproduction...):

Craven and Gerzon (1992) stated that the phase distortion caused by the high-pass response is audible, even if the cutoff frequency is reduced to 5 Hz. They say it causes the bass to lack “tightness” and become “woolly.” Phase equalization of the bass, they say, subjectively extends the effective bass response by the order of half an octave.

Group delay is present due to the changing amounts of phase delay at subwoofer frequencies--due to the high pass nature of subwoofers. But the threshold of audibility of group delay at subwoofer frequencies is still not quantified via peer reviewed study/report.

Chris

 

[Flaesh]

integration seem fine

Can you upload .mdat?

but the thresholds of audibility for group delay (the first derivative of phase) for very low frequencies remains one of the great unknowns in psychoacoustics.

and depends on something (what?))
Recently, one buddy compared MATT test to music, playing through speakers and headphones. With the MATT test, anomalies are audible in room, but with music, they're always inaudible to him. Simple 2.0 stereo, Dirac equalised, IIRC something like about 40 ms delay near 100 herz.

 

[Mort]

Absolutely the best description ever of GD! Hats off to you sir! If we could only get that into Wikipedia...

I dunno seems like he's fat shaming bass response.

 

[Chris A]

Simple 2.0 stereo, Dirac equalised, IIRC something like about 40 ms delay near 100 herz.

My 5.1 array, violet trace represents the left-right loudspeakers (Jubilees), blue trace is the center K-402-MEH, and peach trace are the surrounds (stacked AMT-1s on top of Belle Klipsch bass bins--which roll off at ~60 Hz due to their short horn path lengths):

Phase response of the above:

Chris

 

[Mart68]

I dunno seems like he's fat shaming bass response.

Phat shaming.

 

[klettermann]

Can you upload .mdat?

and depends on something (what?))
Recently, one buddy compared MATT test to music, playing through speakers and headphones. With the MATT test, anomalies are audible in room, but with music, they're always inaudible to him. Simple 2.0 stereo, Dirac equalised, IIRC something like about 40 ms delay near 100 herz.

Here goes.

 

[klettermann]

Thank you Amir, you have set me free! FR and sub integration seem fine. So I'm letting this go. Thanks and cheers,

After a good nights sleep I realize that I'm not as free as I thought. Per amir's comment about not worrying about GD - if it's not worth worrying about it then why does it get so much attention?

 

[kyuu]

After a good nights sleep I realize that I'm not as free as I thought. Per amir's comment about not worrying about GD - if it's not worth worrying about it then why does it get so much attention?

Group delay is a real thing, but it's also one of those audiophile bugbears where people go chasing ghosts. It's like asking why do cables get so much attention.

That said, if you look at is as a learning opportunity there's nothing wrong with taking some measurements and looking at it. Who knows, it's not impossible that there is something abnormal going on.

 

[Keith_W]

After a good nights sleep I realize that I'm not as free as I thought. Per amir's comment about not worrying about GD - if it's not worth worrying about it then why does it get so much attention?

It doesn't get much attention?

It's one of the things you can measure. And you can even do something about it with DSP. But you have to ask yourself whether you are improving your listening experience, or just the measurement? It's like DAC's. Some DAC's measure better, but it won't improve your listening experience. My take on it is: yes, I know the thresholds are unknown. But it doesn't cost me anything to fix it with DSP, so that's what I do.