Is transient response the most important thing for the perceived audio quality in a system ?

[RayDunzl]

Concerning a speaker physically emitting a square wave:

The signal voltage would have an initial jump to the level of the square top or bottom.

The cone/diaphram would have to move out/in to make the initial jump.

The voltage levels off and sustains.

The cone/diaphram holds its position at the top/bottom of the square.


Problem:

The top of the square would represent a steady-state increased pressure "in the room".

What is making that pressure in the room when the cone/diaphram is stationary during the peak of the wave?


My observation:

When I've sent a square to my speakers, I'll see the jumps as the cone/diaphram moves in or out.

Then, whether or not the rest of the wave resembles a square is very dependent upn the fundamental frequency of the square.

At some frequencies it "gets lucky" and manages to resemble a square, at other frequencies it's a mess.



For example, a short segment of a swept square wave, showing the digital signal and what the microphone picks up in-room, as it crosses a frequency where it looks "pretty good".

As the frequency changes during the sweep, the in-toom output cycles through bad/better/worse comparing the in-room wave to the signal applied.

The example above is sweeping through a range near 50hz.





And, just for grins, a measurement at the output of the preamp from long ago:

Original digital signal sent to the system
Preamp output without DRC - 25 foot noisy cable to PC sound card for capture
Preamp output with DRC (room correction) for MartinLogan speakers with subs
Preamp output with DRC for JBL LSR 308

The in-room measurement is not shown here (as not the topic for that moment), though the MartinLogan "corrected" wave is likely in use in the first example above.

The room correction is AcourateDRC creating the filters for a miniDSP Opender-DI.

 

[fpitas]

Then, whether or not the rest of the wave resembles a square is very dependent upn the fundamental frequency of the square.

At some frequencies it "gets lucky" and manages to resemble a square, at other frequencies it's a mess.

That has always been the case when claims about square wave response are made. This is nothing new. Long ago it was a feeble excuse to use first-order filters, and a cult grew up around it.

 

[Tangband]

Regading transients i find the insight of Roger Nichols (audio engineer of Steely Dan) interesting. He claims that transients al already dissepearing when recorded on analog tape within houres to a certain degree. Did read this article maney years ago a started to collect older CD'd (if possible first pressings) that captured as close as possible the existing transients at the time producing the CD.

Nichols,:
if you record something on a piece of analog tape and play it back later the same day, the same program is not on the tape. And there’s nothing so far that anybody’d been able to do about that, you know, like those little magnetic particles are made to be able to wander around and they do so by themselves while the tape is just sitting there. I’ve made DAT copies when I’m cutting tracks, and then have an automation snap shot of the mix and then later that evening put the tape back on, play it back, compare it with the Dat, and there’s already starting to be a difference. And by the time a week or two weeks go by and it’s time to mix, a lot of the transients have started to disappear.

Link: http://pieralessandri.com/ROGER-NICHOLS-interview

Last week i could get hold for just 1euro Carole King Tapastry album released in 1971 an remasterd in 1996 i guess from the original analog tape so no Vinyl master. It contains transients who are atleast preserved till 1996. Priceless..

Very interesting !

 

[Snarfie]

Very interesting !

What i also found remarkable his outspoken disgust for Vinyl in the beginning of the article.

Nichols:

And I hated clicks and pops that were on records, you know, I wanted to sit down, and I was trying to listen to everything. And I couldn’t ignore the surface...the grinding surface noise of clicks and pops.

 

[Cbdb2]

What i also found remarkable his outspoken disgust for Vinyl in the beginning of the article.

Nichols:

And I hated clicks and pops that were on records, you know, I wanted to sit down, and I was trying to listen to everything. And I couldn’t ignore the surface...the grinding surface noise of clicks and pops.

The truth about analog reproduction, its lossy, including studio tape machines. Do people think recording studios all went digital (with machines 2 or 3 times the price) because they sounded worse?

 

[Snarfie]

The truth about analog reproduction, its lossy, including studio tape machines. Do people think recording studios all went digital (with machines 2 or 3 times the price) because they sounded worse?

I was cured from Vinyl when i compard my Vinyl version of Ahmad Jamal - Rossiter road with the dead silence CD. The delicate sounds came alive.

 

[danadam]

It doesn't need to respond faster than the fastest rate of rise/fall of a 20KHz sine wave. And the flat frequency response - if it can do 20Khz at full power - shows that it can do that.

Just for reference (files in attachment), here's 100 Hz square (coincidentally consisting of 100 harmonics), 2800 Hz square (consisting of 4 harmonics), impulse and 20 kHz sine:

And here's a comparison of previous squares to "flat" squares as generated by Audacity with option "Square, no alias" (clearly it does something more than just "no alias"):

 

[theREALdotnet]

That has always been the case when claims about square wave response are made. This is nothing new.

You say that as if it was ok. Shouldn’t this particular waveform be reproduced faithfully, and predictably, like any other?

Within the bounds of available bandwidth, of course. Something like the below. The red curve is a square wave up to the 23rd harmonic, like a 1kHz square wave generated by REW at 48kHz sample rate. The green curve is limited to the 19th harmonic, and this I believe is what the output of HiFi systems should resemble.

I believe it does, to a large degree, up to the amp output, but what comes out of many speakers is a hot mess. There are so many phase errors introduced there that transient-rich output is often distorted beyond recognition, regardless of how well the speaker measures otherwise. Unless we take this into account and measure both magnitude and phase response of speakers (and all components, really), we’ll never be able to come up with a measured fidelity ranking that matches actual sound quality.

Every time I read “and here is the step response, for those who like to see it”, I say to myself ”hell yeah!”

 

[fpitas]

You say that as if it was ok. Shouldn’t this particular waveform be reproduced faithfully, and predictably, like any other?

Within the bounds of available bandwidth, of course. Something like the below. The red curve is a square wave up to the 23rd harmonic, like a 1kHz square wave generated by REW at 48kHz sample rate. The green curve is limited to the 19th harmonic, and this I believe is what the output of HiFi systems should resemble.

View attachment 288215

I believe it does, to a large degree, up to the amp output, but what comes out of many speakers is a hot mess. There are so many phase errors introduced there that transient-rich output is often distorted beyond recognition, regardless of how well the speaker measures otherwise. Unless we take this into account and measure both magnitude and phase response of speakers (and all components, really), we’ll never be able to come up with a measured fidelity ranking that matches actual sound quality.

Every time I read “and here is the step response, for those who like to see it”, I say to myself ”hell yeah!”

I think you'll always be defeated by room reflections, if nothing else. And few speakers, even in an anechoic chamber, will be able to produce a square wave except at one sweet spot.

 

[levimax]

You say that as if it was ok. Shouldn’t this particular waveform be reproduced faithfully, and predictably, like any other?

Why would you want to spend extra engineering energy and money to accurately reproduce a waveform that does not exist in music? There is no reliable evidence that accurate square wave reproduction has anything to do with perceived sound quality. The best use for them is to quickly diagnose amplifier malfunctions using an Oscilloscope but even that use is becoming obsolete with devices like the AP analyzer which has much more powerful tools.

 

[fpitas]

Why would you want to spend extra engineering energy and money to accurately reproducing a waveform that does not exist in music? There is no reliable evidence that accurate square wave reproduction has anything to do with perceived sound quality. The best use for them is to quickly diagnose amplifier malfunctions using an Oscilloscope but even that use is becoming obsolete with devices like the AP analyzer which has much more powerful tools.

Yes, this. It was a massive feel-good exercise when I was young. It hasn't changed.

 

[dualazmak]

Unless we take this into account and measure both magnitude and phase response of speakers (and all components, really), we’ll never be able to come up with a measured fidelity ranking that matches actual sound quality.

Thank you, fully agree! This is why I always stick to 5. objective measurement and subjective listening to actual room air sound at my listening position (which also reflects room acoustic environments), even though I also intensively measure FR as well as "time domain = relative delay" (transient characteristics) in;
1. upstream digital domain within PC,
2. analog line-level output from multlichannel-DAC,
3. amplifiers' SP high-level output before protection capacitor, and
4. SP high-level signal after protection capacitor.

I believe that such measurements in various "stages" in our audio setup, throughout upstream digital domain down room air sound, would be critically important for our observation and understandings aiming towards optimal tuning of "air music sound" in our audio system in our listening environments; of course tuning of room acoustic is another really critical factor, as all of you may fully agree...

Just for example and for your reference, please refer to my post here for the summary of my FR measurements in these five stages.

 

[gnarly]

If you think phase matters with speakers, then yes, .....good acoustical square wave performance matters too.

Everybody can accept frequency response matters, setting phase aside.
Folks clearly have widely varying views on the importance of phase.

Listing some measurement attributes of flat phase, when added to flat frequency response:
- excellent impulse response
- excellent step response
- very good square waves within the audio range that provides sufficient odd harmonics to fill in the fundamental
- excellent transient response
- excellent waterfalls, spectrograms, etc.
- elimination of group delay other than from bottom end low frequency rolloff

All these measurements form an identity....get one right, the others have to be right too.
All due to phase being flat, as well as what's commonly called frequency response being flat.

I dunno guys...is phase important?
If you don't think so, you probably need to throw out all those measurements as being important too....

 

[RayDunzl]

What is a speaker cone doing during the flat part of a square wave?

 

[Doodski]

What is a speaker cone doing during the flat part of a square wave?

For the most part it is stationary and the position dependent on the voltage that is applied or not.

 

[RayDunzl]

For the most part it is stationary and the position dependent on the voltage that is applied or not.

If it is stationary, what happens to the pressure wave in the air?

 

[Doodski]

If it is stationary, what happens to the pressure wave in the air?

I have no idea. I've watched a lot of solenoids operating and they do hold position if the voltage and any offset is steady. That's how they work.

 

[theREALdotnet]

Why would you want to spend extra engineering energy and money to accurately reproduce a waveform that does not exist in music?

How do you know that? This is just one of an infinite number of possible signal shapes. If the speaker cannot reproduce this signal, what other signals will it be incapable of reproducing?

 

[theREALdotnet]

If you think phase matters with speakers, then yes, .....good acoustical square wave performance matters too.

I’m not even sure that phase (alone) is the issue here. There are no transients in a sustained square wave, this is 100% steady state, for all the constituent frequencies.

 

[gnarly]

What is a speaker cone doing during the flat part of a square wave?

Moving