A sine wave sweep is useful for helping choose the right speakers

[steve_1979]

When auditioning different speakers I listen to a range of different types of music to hear how they sound (obviously). I also find it very useful to play a sine wave sweep too. Clicky

A sine wave sweep will show up any variations in the frequency response, phase changes and any other funny stuff that often occurs around the crossover much more clearly than just listening to music alone. Every time I hear a speaker that sounds pleasant with music they are always the ones that also perform well with a sine wave sweep and every speaker that sounds wrong with music are always the ones that perform badly with a sine wave sweep. It's a very useful tool when comparing speakers.

It's hard to describe what phase distortion sounds like but if you compare a few speakers with both music and a sine wave sweep you'll notice an unnatural sound that comes and goes at certain frequencies which sounds and feels very uncomfortable.

With music on its own you hear a mixed mish mash of lots of different frequencies all playing at the same time. Depending on what speakers you're using some of these frequencies are in phase while others are out of phase and all are at different and constantly varying volume levels which drown each other out to varying degrees. This can make it difficult to isolate any problems that the speaker has and you end up with the situation where you have an unnatural or irritating sound but you can't quite put your finger on exactly what's wrong with it.

With a sine wave sweep you can isolate each individual frequency one at a time and hear very clearly how a speaker sounds at this particular frequency. Using a sine wave sweep together with music will tell you much more about why a speaker sounds the way it does and what are its shortcomings are (that may not otherwise be immediately noticed with a quick listen when auditioning them in a shop) than just using music on it own to compare speakers.

While clarity, a flat frequency response, dynamic range, a good stereo image and the absence of clipping are all very important factors for sound quality, having it all in phase will help make speakers so sound much more pleasant and easy to listen too. It's makes such a huge difference to how enjoyable a speaker sounds and yet most audiophiles are too obsessed with worrying about the relatively insignificant cables and bit-rates they use with their massively phase distorted hifi systems to realise.

Phase distortion is more audible than variations in amplitude and yet it's completely ignored by most people so they end up disatisfied with the way their hifi sounds and waste lots of time and money trying all sorts of different solutions to fix it without understanding what the problem is to start with. Their instinct tells them that something is unnatural with the sound they're hearing but they don't know why.

When phase distortion is described on forums many people don't fully understand what effect it has on the sound. If people took the time to play a sine wave sweep as well as music through several speakers in close succession it would be immediately obvious what a big effect that changes in phase has on the sound.

Ok, rant over.

What are your thoughts on this?

 

[Thompsonuxb]

Lol.....

I have a laser lens cleaner from Discwasher with 8 optimisation tools on it.

A frequency sweep test, phase test, pink noise test and the notorious demagnitising test tone.

I run them everytime I clean the lense on my cdplayer. Even the de-magnitising test which halts magnetism building up in circuit boards, crazy?.....maybe, but hey, I believe in Jesus.

Back in the day before I became a discerning listener I ran my tweeters out of phase ( I believe in bi-wireing too) which gave me a much higher frequency response, but will burn out your tweeters eventually.

But all the above I have checked and re-checked on my system.

The things we do.....

 

[Vladimir]

Brilliant write up Steve.

Regarding Phase Distortion, a picture speaks a thousand words (left in phase, right out of phase).

 

[matt49]

Sure, phase distortion is one thing to worry about. But rather than listening for it in an amateurish way, I’d treat it digitally. Devialet’s SAM technology deals with it cleverly. There are other DSP technologies that do the job, e.g. DEQX.

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

 

[tonky]

Sigh 'n' wave*bye*

 

[steve_1979]

Brilliant write up Steve.

Thanks.

It's a shame this damn forum software won't let me edit all of the bad grammar in that post though. It just keeps saying "Access denied" when ever I click on the edit button. *dash1*

Regarding Phase Distortion, a picture speaks a thousand words (left in phase, right out of phase).

Hahaha! Those graphs look like a birds eye view of drunk footsteps in the snow.

 

[Frank Harvey]

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

That was a point I was going to make...

 

[Phileas]

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

That was a point I was going to make...

This isn't a FR sweep. It's listening to the sounds the speaker makes as the frequency input changes.

 

[Leeps]

Sure, phase distortion is one thing to worry about. But rather than listening for it in an amateurish way, I’d treat it digitally. Devialet’s SAM technology deals with it cleverly. There are other DSP technologies that do the job, e.g. DEQX.

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

Hi Matt, could you explain what this means in layman's terms? I'm amateurishly interested in acoustics, but I'm certainly no sound engineer.

Thanks.

 

[steve_1979]

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

That was a point I was going to make...

I agree that a sine wave sweep isn't going to show you everything but it will tell you a lot about how a speaker performs. For example it will show up variations in amplitude at specific frequencies, highlights any phase changes and lets you hear clearly how well the mid/bass driver crosses over to the tweeter.

When auditioning speakers in a HiFi shop it's an easy to use and useful tool that can help highlight a speakers strengths and weaknesses much quicker than just using music alone. You could either use the YouTube video that I posted a link to or there are free phone/tablet apps available that allow you to do a sweep manually.

It's a tool that can be used in addition to using music and like I mentioned in my first post, speakers that sound best with music are inevitably the ones that also perform well with a sine wave sweep.

 

[matt49]

Sure, phase distortion is one thing to worry about. But rather than listening for it in an amateurish way, I’d treat it digitally. Devialet’s SAM technology deals with it cleverly. There are other DSP technologies that do the job, e.g. DEQX.

But for me the elephant in the room when choosing speakers is transient response, and an FR sweep won’t tell you much about that.

Hi Matt, could you explain what this means in layman's terms? I'm amateurishly interested in acoustics, but I'm certainly no sound engineer.

Thanks.

This page explains it (see the paragraph on phase distortion etc).

The Devialet SAM system is designed to get rid of phase distortion in speakers. It does this by measuring (i) speaker output and comparing it to (ii) speaker input; then by comparing (i) and (ii), with the help of some hugely complex maths, it generates a correction file. The correction file is applied by the amp's DSP engine in real time during playback. N.b. SAM doesn't correct for room acoustics; the original measurements are done anechoically. All it does is try to correct the distortions inherent in the design of a given speaker.

The DEQX system (described here) addresses a wider range of issues, including room acoustics. There are other systems that do similar things.

All interesting and worthwhile stuff, though as I noted earlier, I'd be at least as (if more more) interested in the transient response of my speakers, and the use of a sine wave sweep, as advocated by Steve, won't tell you anything about that.

 

[steve_1979]

Interesting stuff Matt. I'd like to hear them in action.

 

[matt49]

Devialet's SAM system is being very well received. As you can turn it on and off with one click, it's simple to do direct comparisons of your system with SAM and without. So far I've only heard SAM with the B&W 805Ds, where it made a considerable improvement. It seems to have the greatest effect on smallish 2-way speakers, where it hugely improves bass response and integration of the drivers.

Since my main speakers have no crossover in the 'presence zone' and have active woofers, they won't get the SAM treatment, but it wouldn't benefit them anyways.

My point about transient response is that it's at least as important (if not more so) than the other issues you mention (FR and phase anomalies). Or to put it another way, the differences in TR performance between speaker designs can be enormous and dwarf the differences in respect of FR. For instance, all dynamic speakers (active or passive) introduce loads of TR distortion by design, and this affects their ability to sound "in the room with you". If you've heard a decent pair of panel speakers, you'll know what I mean.

To find out how good a speaker's TR is you need program material that includes transients, e.g. music. A sine wave sweep won't help.

 

[Native_bon]

Now this is a very interesting topic. Phase distortion can cause all types of problems. Form imaging to loss of instrument timbre to timing. I remember reading an article about this subject a while back & it may just play a very major part in HIFI which is still over looked. Due to crossovers on speakers sending frequencies to the bass, mid & treble, to hifi racks to speaker stands to even room acoustics. They may not on their own cause a major problem, but cumulatively you begin to run into problems very quickly.

Steves piont is a very Valid piont, but just listening to sound waves does not solve the problem. Unless you listen to all same equipment & in same room. A lot can affect Phase distortion & can be very hard to pin dwn. This may also account for systems sounding good in shop & not so good at home.

Also even trying to find the sweet spot to standing & seating, to listening at ear level to speakers. All this plays a major part. The problem for most speaker manufacturers is the problem of two people listening to music from different positions in the room & getting very different results. I think this may be one of the philopsophy behind the design of kef & tannoy speakers to minimise phase problems.

 

[Vladimir]

Would I be able to apreciate from a sine wave sweep the differences in timbral character when comparing speakers? Example, with B&W speakers that have kevlar driver, compared to paper cone drivers, some harmonics are gone due to harder selfdamping of the kevlar material. Where I hear "tzungggg" from a drum tom-tom on a paper driver, I may hear a "thungg" on the kevlar, or "tziinggg" or metal cone driver.

Most have seen this video, but here it is again for those who haven't. RMAF11: Loudspeaker Measurements Explained, John Atkinson, Stereophile Editor

 

[andyjm]

A frequency sweep is (by definition) a single frequency at any one time. It is therefore impossible to have any phase distortion - the phase of what to what exactly?

Listening to a frequency sweep is more likely to show up room effects than it is to allow the speaker to be evaluated. Dips and dives in perceived response will almost certainly be down to the room.

One area however where it may be useful is at the crossover point (and to be fair, this is where the relative phase of the two drivers will matter, but not in the 'phase distortion' context of this thread).

 

[steve_1979]

A frequency sweep is (by definition) a single frequency at any one time. It is therefore impossible to have any phase distortion - the phase of what to what exactly?

You've already answered this question with your next paragraph.

...One area however where it may be useful is at the crossover point (and to be fair, this is where the relative phase of the two drivers will matter,

The crossover between tweeter and mid/woofer is often right slap bang in the frequency range where our ears are the most sensitive. With many speakers the phase in the crossover frequency range is often all over the place and speakers that perform badly in this area sound awful IME.

Phase destortion in the crossover frequency range can be caused directly by the crossover itself, cone breakup when they're operating out of their optimum frequency range or by other EQ/DSP that's been implemented.

Listening to a frequency sweep is more likely to show up room effects than it is to allow the speaker to be evaluated. Dips and dives in perceived response will almost certainly be down to the room.

True.

But if you sit very close to the speakers while doing a sine wave sweep you can minimise how much you hear the room effect (to a degree) and concentrate on the sound coming out of the speaker which will be (relatively) much louder than the sound coming from the room at this distance.

 

[matt49]

Listening to a frequency sweep is more likely to show up room effects than it is to allow the speaker to be evaluated. Dips and dives in perceived response will almost certainly be down to the room.

True.

But if you sit very close to the speakers while doing a sine wave sweep you can minimise how much you hear the room effect (to a degree) and concentrate on the sound coming out of the speaker which will be (relatively) much louder than the sound coming from the room at this distance.

I think Andy's making a different point, viz. that a sine wave sweep will usefully show up room responses but won't say much of interest about direct speaker effects. You need to bear two things in mind:

1. most phase distortion in the strict sense (i.e. as distinct from phase anomalies between drivers) is inaudible.

2. even where it's subtly audible in an anechoic chamber, in real-world situations room acoustics will tend to mask the cues we use to pick it up, so it'll be inaudible.

 

[steve_1979]

I think Andy's making a different point, viz. that a sine wave sweep will usefully show up room responses but won't say much of interest about direct speaker effects. You need to bear two things in mind:

1. most phase distortion in the strict sense (i.e. as distinct from phase anomalies between drivers) is inaudible.

2. even where it's subtly audible in an anechoic chamber, in real-world situations room acoustics will tend to mask the cues we use to pick it up, so it'll be inaudible.

Differences between the phase of the tweeter and mid/woofer in the crossover frequency is easily noticable IMO.

Also if the mid/woofer and tweeter are badly matched you can hear the transition between the two even if they are in phase. This is something else that is easly picked up with a sine wave sweep.

 

[andyjm]

A frequency sweep is (by definition) a single frequency at any one time. It is therefore impossible to have any phase distortion - the phase of what to what exactly?

You've already answered this question with your next paragraph.

Nope. Vladimir's pretty graphs and the common understanding of phase distortion is where the phase delay of different frequencies causes a very different shape waveform to be produced, in spite of having the same spectral energies of the original signal.

The phase issues at the crossover point where both drivers are being driven, and depending on the relative phase of the driver's output result in constructive or destructive interference taking place is something else entirely.

A frequency sweep will not allow you to detect phase distortion. As I posted before, distortion of the phase of what to what?

 

[steve_1979]

Just to add. I only mentioned using sine wave sweeps because they are so easy to use by anyone. You can play a sine wave sweep as easily as a music track when auditioning speakers in a shop.

You don't need to be an expert or take measurements or anything like that. You can just listen to the sound and very quickly learn some useful information about how a speaker will sound. As I said it won't show you everything but it can show you a lot that might otherwise be missed in a brief listen using only music.

 

[steve_1979]

A frequency sweep will not allow you to detect phase distortion. As I posted before, distortion of the phase of what to what?

One area however where it may be useful is at the crossover point (and to be fair, this is where the relative phase of the two drivers will matter,

 

[matt49]

Maybe the word "distortion" is the wrong word to use here [...].

Yes.

I don't think anyone would argue that using a sine wave sweep when auditioning speakers would be utterly useless. It's just that it wouldn't be nearly as useful as your OP tried to claim.

 

[steve_1979]

If anyone reading this thread is interested in trying it out for themselves here are a couple of links that may be useful.

Android apps link: Clicky

Bare in mind that any issues highlighted by a sine wave sweep will be particularly noticable if you compare several different speakers using both music and a sine wave sweep.

 

[matt49]

I don't think anyone would argue that using a sine wave sweep when auditioning speakers would be utterly useless. It's just that it wouldn't be nearly as useful as your OP tried to claim.

I'm not sure why you don't think that a sine wave sweep would be extremely useful when auditioning speakers?

[...]

I didn't say it wouldn't be useful. As for "extremely", I'm not so sure. I find a familiar well recorded piece of classical piano is more useful for highlighting the issues you've talked about.

The reason I responded on this thread is that your OP made one mistake (re. phase distortion) and left out one of the most important aspects of speaker performance (transient response). My responses and andyjm's have merely pointed out the error and the omission.

What's not to like?