Can speaker harmonic distortion decrease to 1% or less if SPL is lowered enough?

[audiodog]

As an example, I have a pair of the JBL 305P speakers. From Erin's measurements, you can see the harmonic distortion graphs here at 86 dB and 96 dB: https://www.erinsaudiocorner.com/loudspeakers/jbl_305pmk2/

Can the speaker harmonic distortion decrease to 1% or less if SPL is lowered enough? Say 65 dB.

Or do speakers have some minimal harmonic distortion level even when the SPL is very low? For example, high quality speaker X will never achieve below 1% harmonic distortion even when played at a very low SPL.

 

[NTK]

One of Purifi's selling points for their drivers is low distortions. Better than -40 dB (1%) at >100 Hz for 96 dBSPL. To measure very low distortion levels, you'll need to test in a very low noise environment, which means good anechoic chambers. Also, distortion levels of measurement microphones are usually not given in their spec sheets, and the system noise levels impose a lower limit on the measurement range. For typical distortions levels, see: https://www.bksv.com/media/doc/be0089.pdf, figures 6.29 - 6.32.

Here are Erin's distortion measurements of the March Audio Sointuva AWG:

March Audio Sointuva AWG Bookshelf Speaker Review

March Audio Sointuva AWG

www.erinsaudiocorner.com

 

[kyuu]

@NTK I'm not really sure that answers the OP's question...

 

[Deleted member 58670]

Sure it can... here's a Magnepan 1.6 with DSP crossover, measured at 75 dB SPL at 2 meter distance. That's equal to 81dB at 1 meter.

Red line is 2nd harmonic, yellow is 3rd.

I dont know for sure if its the speaker or the Umik-1 I'm actually measuring here...

 

[DonH56]

Everything has a lower bound on distortion, though for electronics it is usually below the noise floor when signals are tiny. For speakers, there are mechanical as well as electrical effects, but achieving distortion of below 1% is very common with small signals. Check out the speaker reviews on this site for numerous examples. Here is one, the Revel F206: https://www.audiosciencereview.com/forum/index.php?threads/revel-f206-tower-speaker-review.53691/ Midrange distortion is ~0.1% (-60 dB) for a moderately loud output.

You can go to the master review page below then click on "Speakers" at the top for a list with links to the reviews.

Master Audio Review Index

www.audiosciencereview.com

 

[voodooless]

Everything has a lower bound on distortion, though for electronics it is usually below the noise floor when signals are tiny. For speakers, there are mechanical as well as electrical effects,

There is also just the room noise floor that will inevitably be in the way and will swamp distortion at some point. It really isn’t that different to electronics.

 

[DanielT]

As an example, I have a pair of the JBL 305P speakers. From Erin's measurements, you can see the harmonic distortion graphs here at 86 dB and 96 dB: https://www.erinsaudiocorner.com/loudspeakers/jbl_305pmk2/

Can the speaker harmonic distortion decrease to 1% or less if SPL is lowered enough? Say 65 dB.

Or do speakers have some minimal harmonic distortion level even when the SPL is very low? For example, high quality speaker X will never achieve below 1% harmonic distortion even when played at a very low SPL.

Can the speaker harmonic distortion decrease to 1% or less if SPL is lowered enough? Say 65 dB.

With your JBL 305P, 1% distortion at 65 dB, at X frequency?You don't know until you've tested, measured.

Good speakers JBL 305P, but like some other active speakers in roughly the same price range, for example the Kali Audio LP-6v2, they unfortunately have rather high distortion.

Regarding when, at what level, you hear distortion. A quote from another forum that I pasted here on ASR:

 

[Juhazi]

I have added extra horizontal bar inside the box and measured my MkI speakers indoors at 1m. Noise floor was around 40dB. REW with Umik-1.

Yes it can go below 1%

 

[Ageve]

Yep, less than 1% THD >50-100 Hz is common for good speakers, at 86dB / 1m.

One example, measured with Umik-2 in my quiet living room:

Less than 1% deep bass distortion is not as common, but there are exceptions, like Revel 328Be measured by Amir:

Revel F328Be Speaker Review

This is a review and detailed measurements of the Revel F328Be floor standing speaker. I purchased a single unit through our company (Madrona) with kind discount from Harman Sales (I asked for a review sample from the PR group but was told "don't call us, we will call you."). The retail cost...

www.audiosciencereview.com

 

[DanielT]

Yep, less than 1% THD >50-100 Hz is common for good speakers, at 86dB / 1m.

One example, measured with Umik-2 in my quiet living room:

View attachment 429598

Less than 1% deep bass distortion is not as common, but there are exceptions, like Revel 328Be measured by Amir:

View attachment 429601

Revel F328Be Speaker Review

This is a review and detailed measurements of the Revel F328Be floor standing speaker. I purchased a single unit through our company (Madrona) with kind discount from Harman Sales (I asked for a review sample from the PR group but was told "don't call us, we will call you."). The retail cost...

www.audiosciencereview.com

Fortunately, with increased distortion in the lowest frequency registers, our ability to detect (and get annoyed by) distortion then also decreases.

I wonder how many can detect for example 5% distortion at 50 Hz? Are there studies on that? It doesn't have to be research studies, it can also be fun to read hobby blind distortion subwoofer tests or something like that.

 

[thewas]

I wonder how many can detect for example 5% distortion at 50 Hz? Are there studies on that? It doesn't have to be research studies, it can also be fun to read hobby blind distortion subwoofer tests or something like that.

If its low order like h2 I am quite sure its not audible, while possibly 5% h5 is, there exist quite a lot of studies on this topic, some can be found here:

Audibility of low frequency distortion

@ctrl To discuss your post in the other thread. I'm travelling atvthe moment so please excuse dodgy phone typos and ability to access all the info I need to. Fully accept the point about masking but let's look at another graph So here we see H2 needs to be about 15dB below the fundamental...

www.audiosciencereview.com

On the other hand the typical harmonic distortions are not enough to sufficicently show audible problems (or their absence), mulitone distortions are also needed for that (which usually relate to more audible problems when a woofer not just not plays lower bass):

Low frequency harmonic distortion is almost inaudible. So what’s the point of low distortion drivers?

Putzeys, Dec. 2019

purifi-audio.com

 

[NTK]

I wonder how many can detect for example 5% distortion at 50 Hz? Are there studies on that? It doesn't have to be research studies, it can also be fun to read hobby blind distortion subwoofer tests or something like that.

The table below lists the distortion thresholds ANSI/CTA-2034-B uses to determine the loudspeaker's peak SPL rating. For 50 Hz (max of band 1), the limits are:
H2: -12 dB (25.1%) or less
H3: -18 dB (12.6%) or less
H4: -21 dB (8.9%) or less
... etc
Unfortunately, I couldn't find a reference to where these distortion threshold numbers came from.

 

[RickS]

Like that @thewas referenced Bruno P's paper. There are some others that am guessing @Lars Risbo would add. In any case, there are many factors here that make this answer more complex than the posted speaker measurement indicates. They include driver materials, driver design, enclosure and driver mounting just to name a few. Another complexity is accurate measurement too. Room noise has already been mentioned as potential limiting factor. As usual, consistent conditions matter, and so meaningful comparisons are often limited. Also, not many manufacturers share detailed distortion data. However, Purifi does and one in particular gives us an interesting view of the relationship between SPL and harmonic distortion...

Am not sure of all the measurements conditions (enclosure and environmental) and should emphasize this is ONE specific driver. If you check the Purifi website for others, you can compare just how different each driver is. These are some of the better drivers on market with regard to low distortion so typical mass market drivers are likely much worse. Yes, there are audibility considerations but has been mentioned already but this hopefully (somewhat) addresses the OP's overall question. The above graph suggests a roughly linear relationship between SPL and distortion but pick another driver and you would be (rightfully) hesitant to make generalizations.

Notably, this is just one variation with Purifi's (very consistent) offerings, would expect even greater variation between different vendors and different driver designs. Bottom line, the answer is not simple but may be more so if some of the variables are constrained much more than posed by the OP.

 

[DVDdoug]

There is also just the room noise floor that will inevitably be in the way and will swamp distortion at some point. It really isn’t that different to electronics.

Right... And, at low levels distortion is less audible simply because it's quieter, and it can drop completely below audibility.

With electronics, the noise generally remains constant so as a percentage it's worse at low levels. Ideally, the noise will be below audibility even with silence. (Of course passive speakers don't generate noise unless they are buzzing or rattling, etc.)

...If you take a 16-bit file and reduce it by -48dB digitally, you are down to 8-bits of resolution and it sounds rather bad, but only if you re-amplify it. When music fades-out, the distortion (quantization noise) remains below audibility and it sounds just-fine!

 

[DonH56]

High'ish levels of bass distortion is often perceived as "good" or "fuller bass" since the harmonics are more readily heard than the fundamentals.

 

[kemmler3D]

Extending the original question a bit... Harmonic distortion can probably be reduced almost to zero in drivers if tech gets good enough.

I think Doppler distortion (which appears as a type of modulation distortion) has a minimum above zero.

It happens when high frequencies and lower frequencies are coming from the same cone. The high frequency motion is forced to "ride" on the cone as it moves for the lower frequencies, applying the familiar Doppler effect to the high frequency stuff. This produces a kind of frequency modulation and is unavoidable in a given frequency band.

I think you can reduce it at a given SPL with larger cone area, maybe this is an underappreciated feature of why large speakers are often said to sound "effortless"...

 

[DonH56]

Reducing FMD usually requires smaller, not larger, drivers...

 

[dfuller]

Reducing FMD usually requires smaller, not larger, drivers...

I am assuming the reason for reduced FMD with a larger driver is less required excursion for a given SPL. Otherwise it wouldn't make sense, and yes, more ways means less IMD all other things being equal.

 

[kemmler3D]

Reducing FMD usually requires smaller, not larger, drivers...

My thought is that if you kept SPL and bandwidth for the driver in question constant, but used a larger area, the cone would be moving more slowly for a given signal, and therefore FMD would decrease. But I might have confused myself here.

 

[dfuller]

My thought is that if you kept SPL and bandwidth for the driver in question constant, but used a larger area, the cone would be moving more slowly for a given signal, and therefore FMD would decrease. But I might have confused myself here.

Not so much slower but moving less.