Time Domain measurements?

[73hadd]

Hi all,

I'm glad to find this place because I think measurements are important.

Am I overlooking time domain measurements in the reviews?

For reference, stereophile reviews of the Revel F228Be and B&W 802 D3 reference time domain comments like "tweeter's output arriving at the microphone before the midrange unit's, which in turn arrives before that of the woofers. " and "the output of the tweeter is very slightly too forward in time, which suggests that the B&W's optimal listening axis will be a little below the tweeter axis"

I am wondering if time domain affects perceived stereo imaging and depth.

I realize now there are two parts to my question:

1. Is time domain something that could be added to the measurements in the reviews here? Or explain why it's not important?
2. Maybe in real life sound, when a cymbal crashes, the higher frequencies arrive at the listener first before the other frequencies? So perhaps all sounds arriving at the listener at the same time, from the speaker, is not really a design advantage, and perhaps not even more accurate or realistic?

Any thoughts are appreciated, thank you!

 

[amirm]

1. Is time domain something that could be added to the measurements in the reviews here? Or explain why it's not important?

It could be but it is not important. That is why it is not in CEA-2034 standard we follow.

Research shows that we can sort of hear timing information in anechoic chamber and with headphones. In a real room though, it simply is not audible because reflections in the room all come at vastly different timing and mix together. What the speaker does at the start simply doesn't matter.

See the paper, On the Audibility of Midrange Phase Distortion in Audio Systems
STANLEY P. LIPSHITZ, MARK POCOCK, AND JOHN VANDERKOOY
University of Waterloo, Waterloo, Ontario, Canada N2L 3Gl

 

[GXAlan]

https://www.stereophile.com/content/measuring-loudspeakers-part-two-page-4

This article also provides the same plain language discussion. One expects it to make a difference but it doesn’t seem to pan out, although the authors do allow for exceptions.

”In 1990, Rodney Greenfield and Malcolm Omar Hawksford [34] used DSP-based digital filters to try to separate the audible effects of a loudspeaker's phase error from its amplitude response error. The point was made that a semi-reverberant environment will tend to mask phase effects. In addition, when typical recordings are played, which anmay have undergone many phase-altering stages during production, the audibility of phase differences becomes moot: "one is simply detecting a change in phase distortion and not a correction of it and as such preferences would most likely be personal." Nevertheless, the authors "very tentatively" concluded that equalizing a loudspeaker's excess phase error modified listeners' perception of the apparent soundstage.”

I do think that things that are detectable in anechoic environments may occasionally be detectable in some music in some environments at some times. This would be the harmony between objective research and subjective experience. The effect is extremely subtle in contrast the other more dominant effects. Room correction achieves a lot of the time domain correction which is why people value a high end transparent algorithm

 

[QMuse]

Room correction achieves a lot of the time domain correction which is why people value a high end transparent algorithm

Sure, you can make quite precise time domain correction, and it will look really nice with some FDW applied to the response measured from LP. But once you remove FDW reflections will make a mess of it and that is the reason why time domain correction is hardly audible in most of the rooms. Only if your room is pretty much reflection free you have a chance to hear the effects of time domain corrections and even then they will be subtle.

 

[tuga]

Everyone is justly raving about the performance of the Kii3 and the D&D 8c.

The Toole school preaches that everything which matters is expressed in the frequency response measurements.

Yet there is a chance that the time-domain performance of these speakers may contribute to the overall results.

https://www.stereophile.com/content/dutch-dutch-8c-active-loudspeaker-system-measurements

https://www.stereophile.com/content/kii-audio-three-loudspeaker-measurements

 

[tuga]

I dare say that group delay is also relevant to this topic.

 

[Absolute]

I did a lot of experiments with the timing correction on and off when I had Kii Three because you can switch it on and off while playing music.

I first came to the conclusion that it does change the sound a wee little bit, like it just feels a little bit less harsh with it on. So I blind-tested it.
I don't remember the exact numbers, but I think I only managed 6 or 7 out of 10 during intense listening and many switches back and forth.

When not switching, but trying to hear if I could recognize the signature it was just random like one would expect. This was with normal pop music like Simon & Garfunkel, Tracy Chapman and some track with a huge drum solo with lots of hi-hats and and dynamics.

I concluded that my room was not sufficiently revealing to reliably discern A from B during normal listening. But I remain convinced that it could very well be audible.

 

[thewas]

The audibility also depends on the listening material, for example in productions where the phase of individual mixed tracks is 'muddled" altogether the chances to hear a difference are lower. But even with good acoustics and music recordings the differences are usually just subtle, kind of icing on the cake.

 

[tuga]

I did a lot of experiments with the timing correction on and off when I had Kii Three because you can switch it on and off while playing music.

I first came to the conclusion that it does change the sound a wee little bit, like it just feels a little bit less harsh with it on. So I blind-tested it.
I don't remember the exact numbers, but I think I only managed 6 or 7 out of 10 during intense listening and many switches back and forth.

When not switching, but trying to hear if I could recognize the signature it was just random like one would expect. This was with normal pop music like Simon & Garfunkel, Tracy Chapman and some track with a huge drum solo with lots of hi-hats and and dynamics.

I concluded that my room was not sufficiently revealing to reliably discern A from B during normal listening. But I remain convinced that it could very well be audible.

May I suggest that you use PlayClassics' recordings which use only one mic per channel and no processing except "room correction" EQ.
They have an amazing recording of a drum kit.

 

[Absolute]

The audibility also depends on the listening material, for example in productions where the phase of individual mixed tracks is 'muddled" altogether the chances to hear a difference are lower. But even with good acoustics and music recordings the differences are usually just subtle, kind of icing on the cake.

If you like your stereo, do NOT try these CDs with weird stuff on it that's designed to check for phase, imaging etc. You might go from happy to unhappy in the blink of an eye!

That was just random noise to have an excuse to ask you; you're tagged as an industry insider all of a sudden. Who, what, where, when, how and why?

@tuga I would if I could, but I no longer have the Kiis. Right now I don't have anything, actually, but trying to go the DIY route.

 

[QMuse]

Here is how you can check where you stand at your room and your LP. Make a single sweep from the center of your LP at the height of your head and apply FDW of 12 cycles.

If you get something like this, chances that you will hear your perfect time domain correction are slim to none.

But if you get something like this, chances are you will be able to hear some subtle improvement:

Here's how I see it: the more reflections you are receiving at LP (due to the characteristics of your room, and with distance to your LP being larger the worse) the more they keep mixing with direct sound adding out of phase content and thus making the summation signal choppy. If the signal is not to contaminated with out of phase componenets coming from your reflections your listening environment is not so alive and thus more similar to anechoic chamber (in which you would be able to hear time domain correction). But if it is, well you can still enjoy your lively room as much you enjoy singing in your bathroom.

Btw, my room has RT60 of app 500ms, same as listening room of Mr. Linkwitz who thinks that it's just about right.
Here is what he says on his webpage:
"RT60 for my listening room is around 550 ms and pleasantly lively for conversation and socializing. I do not like a muffled cocktail lounge ambience. "

 

[DonH56]

Frequency response, magnitude and phase, can be converted to time response and vice-versa. The magnitude is not the only part of it. Group delay, which seems a worthwhile thing to plot for those who understand it, is derived from the phase response.

 

[QMuse]

Frequency response, magnitude and phase, can be converted to time response and vice-versa. The magnitude is not the only part of it. Group delay, which seems a worthwhile thing to plot for those who understand it, is derived from the phase response.

Sure. But the point here was to show how reflections are affecting phase and GD response which makes time domain correction pretty much impossible to hear if room is reflective happy.

 

[GXAlan]

Sure, you can make quite precise time domain correction, and it will look really nice with some FDW applied to the response measured from LP. But once you remove FDW reflections will make a mess of it and that is the reason why time domain correction is hardly audible in most of the rooms. Only if your room is pretty much reflection free you have a chance to hear the effects of time domain corrections and even then they will be subtle.

Perhaps, but I think one of the main advantages of FIR filters is that it can make corrections that take into account the direct sound vs. room reverberations and do some time domain adjustment.

The Pioneer MCACC Pro featured in the flagship Pioneers gets a lot of criticism but it actually has a lot of post calibration control. You can turn on or off the group delay correction and in many cases, there is a change in the sound stage particularly in a home theater environment.

Dr. Toole never said that time domain makes zero difference. He said that it has a very small effect in comparison to amplitude effects. He also focuses on mono testing. This makes sense. For every $1 spent, you are better off trying to improve the FR and the off axis sound. However, the optimization for most people in most rooms, still allows for some deviation.

As others have pointed out, a lot of highly processed music may not be sensitive to time errors. Rooms and the amount of furniture in the room, the layout of the room, etc. can all make a difference.

The greatest limitation of treating Dr. Toole’s textbooks with an appropriate level of context arises when considering exotic designs like the JBL DD55000, S2600, S3100, and 4660. These really aren’t addressed by Dr. Toole’s textbook although all of the measurements can be done and all of the same metrics can be assessed. Even with the best traditional Toole-speaker, be it a Salon2 or M2, has the issue where as soon as you move off center and are getting closer to one speaker versus another, the phantom center gets shifted as the speaker you are closer to becomes louder. Normal physics.

Speakers with asymmetrical horns are able to project the sound in such a way that as you go lateral to center, you actually get more energy from the contralateral speaker and less energy from the ipsilateral speaker, preserving the center image, even off center.

There are a lot of disadvantages to these speakers including a small envelope for the room where the precalculated directivity works (I.e. minimum and maximum separation of the speakers). In HT applications, the center channel itself can solve the center image issue). But it is something that results in a different optimization question. Which is better, the speaker that sounds perfect for one listener? Or the one that sounds pretty good for 3 listeners sitting side by side? Again, not voodoo, but measured.

 

[QMuse]

Perhaps, but I think one of the main advantages of FIR filters is that it can make corrections that take into account the direct sound vs. room reverberations and do some time domain adjustment.

Can you explain how exactly would you do such corrections in time domain?

The Pioneer MCACC Pro featured in the flagship Pioneers gets a lot of criticism but it actually has a lot of post calibration control. You can turn on or off the group delay correction and in many cases, there is a change in the sound stage particularly in a home theater environment.

You can't turn off group delay and have time correction. Time correction is correction of the phase response and group delay is 1st derivation of the phase response vs frequency.

 

[GXAlan]

Can you explain how exactly would you do such corrections in time domain?

https://www.yumpu.com/en/document/view/8042697/audyssey-laboratories

I thought that is one of the core strengths of Audyssey?
https://ieeexplore.ieee.org/abstract/document/1291970

http://www.aes.org/e-lib/browse.cfm?elib=17298

You can't turn off group delay and have time correction. Time correction is correction of the phase response and group delay is 1st derivation of the phase vs frequency.

Maybe I am confused. I consider group delay one example of a time domain issue that can be corrected and can result in audible differences. Maybe we are just disagreeing over semantics.

 

[QMuse]

https://www.yumpu.com/en/document/view/8042697/audyssey-laboratories

I thought that is one of the core strengths of Audyssey?
https://ieeexplore.ieee.org/abstract/document/1291970

This is about frequency domain correction, not time domain.

Maybe I am confused. I consider group delay one example of a time domain issue that can be corrected and can result in audible differences. Maybe we are just disagreeing over semantics.

Nothing to disagree about as it is a simple definition. As I said, GD is defined as 1st derivation of the phase response vs frequency.

When you make derivation of trip vs time you get speed vs time function. When you derive phase vs frequency you get GD vs frequency function.

 

[Matias]

In my opinion time coherence is subtle but important. That is one of the reasons I chose and still love my Thiel CS 3.7 speakers.

Current active speakers can switch this on and off to test. I have listened to a demonstration of a Goldmund system and it was noticeable. Even fixing it with convolution is worth it. After all it is still part of high fidelity to the source.

 

[QMuse]

Current active speakers can switch this on and off to test. I have listened to a demonstration of a Goldmund system and it was noticeable.

Was it a blind test?

 

[q3cpma]

It's hard to completely disregard phase/group delay when some very serious players have tried to do something about it. Some examples:
* http://www.aes.org/e-lib/browse.cfm?elib=19404 Genelec paper having for conclusion "Our results suggest that when the group delay in the frequency range from 300 Hz to 1 kHz is below 1.0 ms, it is inaudible. With low-frequency emphasis the group delay variations can be heard more easily.".
* http://www.neumann-kh-line.com/neum...g_discontinued-monitors_studio-products_O500C A forgotten gem from Klein & Hummel who did in 2001 what is done today by Hedd and KS Digital.

My opinion is that it's something that can easily be solved with today's DSP loudspeakers, transforming phase anomalies into fullband delay. Better to apply it than not, when there's no real tradeoff; that way, you don't have to ponder for centuries if it's actually audible.
Decay is also easy to solve with modern cabinet constructions technics (computer modelisation and Klippel analyzer). Stuff like Genelec's (amongst others) cast aluminium cabinet, Revel's curved enclosures, https://ggntkt.de/en/model-m1/ or https://www.qacoustics.co.uk/qacoustics-concept300.html#stand shows that we can at least try to do better than cheap MDF/plywood boxes.