Importance of SPL-Response Linearity

[Mr. Widget]

I'll raise the (heretical?) thought that speakers that compress slightly at normal listening levels may sound more laid-back.

I imagine the sound character change will vary with the speaker type/design.

The graphs hugely amplify the differential. Between 86 and 96 dB, there is less than 0.5 dB change in Wharfdale above 60 Hz! Even at 50 Hz it is just changing 1 dB.

Very true, but I remember reading that the engineers at Harman saw frequency response differences between LEAP measurements where the signal was a swept sine wave and MLSSA which uses a short pulse that is deconvoluted. Ultimately they attributed the difference to the heating of the VC when using the swept sine wave. If that short frequency sweep can affect the performance of a loudspeaker, just think what 20 minutes of spirited playback may do?

 

[LTig]

In my experience that "stressed" sound is mainly a result of high multitone distortion which often exceeds 10%.

It was also lots of compression.

 

[MaxwellsEq]

I'll raise the (heretical?) thought that speakers that compress slightly at normal listening levels may sound more laid-back.

Compression almost always increases the energy at any given frequency, so it tends to sound more intense, not more laid back.

 

[Avp1]

I keep looking at these graphs in Erin’s reviews and thinking these measurements deserve more discussion. Doesn’t a lot of difference hear suggest that the speaker’s dynamic response is poor?

View attachment 249088
View attachment 249093

These two graphs simply point the fact than Litton runs out of juice at around 90dB, which is rather low. Kef Reference is not shy at 100dB and likely can do some more, which makes it appropriate for playing quality records with high dynamic range. Though Kef will still benefits from adding a subwoofer.

 

[ahofer]

These two graphs simply point the fact than Litton runs out of juice at around 90dB, which is rather low. Kef Reference is not shy at 100dB and likely can do some more, which makes it appropriate for playing quality records with high dynamic range. Though Kef will still benefits from adding a subwoofer.

If these measurements were done more broadly, then it might be simple to point out. But it seems like Erin is the only one who does them. Given the amount of argument on this site about completeness of our suite of measurements, this seems an important one that gets short shrift.

I’m guessing horns do better on this dimension.

 

[valerianf]

It would be interested to get the same measurement for a JBL or a Cerwin Vega speaker designed to output high SPL.
We could compare the result with those of smaller speakers.

 

[Plcamp]

The graphs hugely amplify the differential. Between 86 and 96 dB, there is less than 0.5 dB change in Wharfdale above 60 Hz! Even at 50 Hz it is just changing 1 dB.

Is the spl level at which the onset of such behaviour starts a useful indicator/differentiator?

Audible or not, it does indicate the beginnings of something, worsening with spl.

 

[ahofer]

Here’s JBL 4305P. A small powered monitor with a horn (waveguide?)

And the M2:

 

[Avp1]

Is the spl level at which the onset of such behaviour starts a useful indicator/differentiator?

Audible or not, it does indicate the beginnings of something, worsening with spl.

It does. This is SPL above which speaker becomes significantly non-linear. When you play music, you better keep average SPL at least 6dB below that point. 90dB limit means that speakers are not good for anything other than near field listening (like placing them on your computer desk).

 

[valerianf]

Professional monitor from JBL.
Speaker drivers and amplification are optimized to get the same response curve whatever is the SPL level.
It is a different design that has some additional qualities: low volume or high volume listening give you the same result.

 

[Plcamp]

It would be fun to see a surface plot (z dimension spl) of this in 1 db increments over a standard range for a couple of closely competing speakers.

There is no replacement for displacement?

 

[Thomas_A]

Shouldn’t those stress tests have a more music-like power signal?

 

[Holmz]

Shouldn’t those stress tests have a more music-like power signal?

Pretty in pink…

 

[Mr. Widget]

Shouldn’t those stress tests have a more music-like power signal?

Meyer Sound thinks so. They developed M Noise which they claim is a closer match to actual music than any other test signal.

 

[Deleted member 36000]

Some observations….

The graphs posted by the OP are comparing a $1500 set of speakers to an $8000 set. Ive compared a lot of these graphs from Erin’s various reviews and the mass majority more closely resemble the Linton’s graph, whereas graphs like the KEFs are the exception and generally are reserved for more expensive speakers.

I also appreciate that Amir pointed out that the graphs give a pretty exaggerated sense of scale. For sane levels of listening, the compression is pretty minor and for my ears, would probably be inaudible.

For the record, I have the Lintons and I love them. They are in a pretty big room (our main living area) and they get plenty loud for my tastes without sounding strained.

 

[ahofer]

Some observations….

The graphs posted by the OP are comparing a $1500 set of speakers to an $8000 set. Ive compared a lot of these graphs from Erin’s various reviews and the mass majority more closely resemble the Lintons graph, whereas graphs like the KEFs are the exception and generally are reserved for more expensive speakers.

I also appreciate that Amir pointed out that the graphs give a pretty exaggerated sense of scale. For sane levels of listening, the compression is pretty minor and for my ears, would probably be inaudible.

For the record, I have the Lintons and I love them. They are in a pretty big room (our main living area) and they get plenty loud for my tastes without sounding strained.

The point is to discuss the importance of this measurement.

 

[Sancus]

Meyer Sound thinks so. They developed M Noise which they claim is a closer match to actual music than any other test signal.

M Noise is just pink noise up to 700hz and practically all domestic HiFi speakers bottleneck below 700hz.

M Noise is more relevant and useful for PA and cinema horns and stuff like that.

The big problem with SPL tests is the failure condition. THD doesnt correlate with audible issues so you need a more complicated composite of distortion components. There are a lot of proposed alternatives but nothing that has really stuck so far. Sub measurements have standardized around CEA2010 at least.

 

[Deleted member 36000]

The point is to discuss the importance of this measurement.

Yeah, I tend to go off on tangents a lot. I did mean to add that I like seeing this kind of measurement, personally. I don’t see it as an end all, be all kind of measurement, but I think it speaks to the quality of the tweeters/drivers and how well a speaker is engineered.

 

[Avp1]

I also appreciate that Amir pointed out that the graphs give a pretty exaggerated sense of scale. For sane levels of listening, the compression is pretty minor and for my ears, would probably be inaudible.

Same is commonly said by low power tube amplifiers owners. They got used to compression. If you are not trained, you may not recognize it. But when you compare speaker which compresses sound with one that not, the difference is usually clear.

 

[kemmler3D]

Shouldn’t those stress tests have a more music-like power signal?

I think using tones is fine, if you fed more energy in at once (e.g. pink noise or something) you wouldn't necessarily find anything different, unless it's an active speaker with weak amps. So maybe?

I think these tests should be standard at LEAST for any active speaker. As Amir has said, he hates speakers that run out of power, and since we need to test the amp and speaker as one, this is a straightforward way of testing the amps within actives.

I used to sell cheap Bluetooth speakers, and compression was almost always easy to hear. In fact, most of them include DSP-based dynamic compression. I think higher-end speakers aren't categorically different. But hopefully the compression is not as easy to hear.