Complaint thread about speaker measurements

[amirm]

I really want to know true distortion numbers for the KEF product range. The entire reference line, including the Blades, Blade Two, Reference 1/3/5 all quote the exact same thing, 0.5% distortion at 40hz at 90dB. It really makes zero sense for the single bass driver bookshelf to measure the same as the Reference 5 with its 3 extra drivers. And Imagine trying to play movies with these speakers which can go much lower then 40hz! It’ll sound so compressed and crap, because THX reference levels calls for 105dB peaks.

All of these measurements are performed with microphones which have their own distortions. After all, a microphone is a speaker in reverse. Microphones are usually rated to have 3% THD up to their max SPL. So that 0.5% distortion may very well be dominated by microphone itself.

There are many problems with accurate distortion measurements of speakers and this is one of them.

Other problems are room modes impacting the detection levels of harmonics as they are exaggerated or reduced by room modes. And no, you don't escape this in anechoic chamber either because they are not anechoic down to 20 Hz.

 

[spacevector]

@amirm How loud does it get inside your home and outside your garage when the test is ongoing? Enough to bother your family and neighbours?

 

[amirm]

@amirm How loud does it get inside your home and outside your garage when the test is ongoing? Enough to bother your family and neighbours?

As I just noted, the garage where the rig is, is quite well insulated both from outside and rest of the house. We can hear the test tones if we turn off the TV. Otherwise, the TV sound drowns them out. It is good that a bit leaks out so I know when the test is done.

We live in rural area and the neighbor is too far away to hear such tones. Now if I performed max SPL tests with subs and such, that may be another matter.

 

[Thomas_A]

The gating threshold can be whatever you want. The higher you set it, the more effort on the part of computational system to compute the correct soundfield. In my measurements, the error in computational portion is acceptable (below -20 dB) up to many multiples of this. Here is the analysis for the KEF LS50:
View attachment 47915

So gating can be moved to above 10 kHz and take advantage of exclusion of reflections. I will see if I can run a computation both ways and see the difference.

I am not clear on your first question. If you mean error in field identification, then that answer is in the above graph. The error is 40 dB lower than the speaker response so it should be lose in the noise.

The first question. would include sums of errors including precision of the microphone calibration. If microphones are calibrated to +/- 0.2 dB, that would be a specification.

The question about gating is related to the ups and downs >5 kHz which looks quite similar on most speakers tested and may indicate reflections from the microphone attachment. Since the microphone itself is moved including the attachment, I am not sure if the system can differentiate those reflections (as opposed to reflections from fixed boundaries in the room etc).

 

[ctrl]

All of these measurements are performed with microphones which have their own distortions. After all, a microphone is a speaker in reverse. Microphones are usually rated to have 3% THD up to their max SPL. So that 0.5% distortion may very well be dominated by microphone itself.

When using a good measuring microphone, one can expect less than 0.09% THD at sound pressures up to about 110dB.

But as Amir said, as always, there are many other factors that make measuring distortion difficult.

 

[spacevector]

As I just noted, the garage where the rig is, is quite well insulated both from outside and rest of the house. We can hear the test tones if we turn off the TV. Otherwise, the TV sound drowns them out. It is good that a bit leaks out so I know when the test is done.

We live in rural area and the neighbor is too far away to hear such tones. Now if I performed max SPL tests with subs and such, that may be another matter.

So it may not be too bad for you to run the linearity/compression/power-handling/max-SPL tests that a few members so far really want to see?

 

[restorer-john]

Maybe I should test then if they use real wood, that it is real wood. Or that if the veneer comes from renewable resources, I go chase that too.

I am not the verification and QC lab for manufacturers. If this is such an important thing, why don't you set up shop to do SPL testing and members send you speakers? You just need a microphone and REW. What are you waiting for?

Same tactics Amir. They simply don't work. You are the one setting up shop as a speaker reviewer and this thread is about complaints, measurements and suggestions.

So let's get this straight. You have zero intention of testing power handling capabilities or dynamic compression of loudspeakers and you have zero intention of testing the performance of the amplifiers in powered loudspeakers in the loudspeakers you review?

Simple yes or no answers will suffice.

 

[pma]

The amp used must be tested with the speaker - by measuring directly at speaker terminals, behind the speaker cable. Balanced interfaces enable this method and AP input schemes should make it easy. My measuring loop is about 20m long and still, with balanced interface, there is no problem with induced hum/buzz.

 

[amirm]

So let's get this straight. You have zero intention of testing power handling capabilities or dynamic compression of loudspeakers and you have zero intention of testing the performance of the amplifiers in powered loudspeakers in the loudspeakers you review?

Where the heck did you read that? I take offense to your comments saying reviews are useless right now because they don't have all the graphs you used to see in magazines in 1970s. Research fully backs my position on what is important to measure and that is precisely what I am doing now. I am spending quite a bit of time offline trying to determine other things to test. I don't want to throw out graphs and then deal with demands to prove their correctness.

To wit, Ray asked for distortion measurements. No sooner than I put them in the KEF measurements that a ton of questions have risen already on how they were measured, if they are correct, etc., etc.

My goal here was to solve the biggest problem in speaker measurements: creating data that correlates well with listening preference. Hardly any manufacturer provides this data including companies like Harman which have this data. I now provide this data. You ignore that and then spit in my face with your commentary about lack of SPL testing, right now and right this minute?

This is why I don't jump the moment you demand. As it is, it feels like a few of you are so spoiled with your demands on my time, resources and knowledge of what matters in acoustic measurements.

Every day I measure a new speaker and wonderful set of data pops out in the form of spinorama information. It is fantastic journey of discovery. Your desire for me to replicate a bunch of graphs to sell magazines in the past is not my priority. Doesn't mean I won't do it but it is not a priority relative to measuring more speakers to verify the usefulness of the spinorama measurements.

You think my current measurements don't advance the needle infinitely over where we are now in the industry/press, then move along. I am not here for you. I am here for people who believe what acoustic research has taught us is important after nearly 40 years.

 

[amirm]

So to be absolutely clear: my mission is to produce CEA 2034 spinorama measurements for speakers. I am laser focused on that. And the system I have is best in class to produce that information efficiently so that I can test a ton of products as I have done with electronics measurements.

Anything else added to that which dilutes that mission is of low interest for me. Rude remarks, demands, etc. only diminish my interest there, not the other way around.

I was clear about this from day one that I discussed this project with the membership. For months now. I want to predict how well you may or may not like a speaker using measurements.

The system I have is NOT optimized for other types of measurements. It makes no sense to use this $100K system to make SPL measurements and such at a single point. All the time wasted on that could be used to measure more speakers in its automated manner. Any of you can buy a microphone and point it at a speaker and measure SPL handling and such. You don't need this system for that.

If you are on a different mission than me, I suggest getting off the train and following other work on this site.

 

[edechamps]

It makes no sense to use this $100K system to make SPL measurements and such at a single point. All the time wasted on that could be used to measure more speakers in its automated manner.

I can understand that. It's just a bit of shame because, when making a purchase decision for a loudspeaker, as far as performance is concerned, typically there are two main questions: (1) does it sound good enough and (2) is it loud enough. If your reviews only answer (1) then we really only have half the information that's relevant for choosing a speaker. Yes, it's the most important half - but still. However I do get the point that one doesn't really need an anechoic chamber to do SPL measurements, so I guess this is a gap others could easily fill.

 

[jtwrace]

@amirm can you also do the same for the electronics side of reviews?

 

[Trouble Maker]

@amirm
Glad to hear you won’t become an audiophile martyr due to frostbite testing speakers in your garage.

For the heat, many people I know just repurpose old/used (gas) heaters from their whole house HVAC. One can even use the a/c part of it. Then the system is completely separate from the house. The reason most do this separate system is to use it on demand if they go to work in the garage, or keep a warmer but not really livable temperature in the winter so their cars aren't freezing when they get in them. There is usually little reason to keep a garage at the same temperature as your house most of the time.

Since you already have radiant heat, if you find yourself wanting to add a/c you could add a small mini split. I'm planning to redo the HVAC in our house and plan to go with a multi split system due to large temperature imbalances across our house, with hidden air handlers (I personally don't like the on wall ones). I'm also planning to at least save the heater portion of our current system to use in the garage when we get around to building a new, bigger one. The current one is ~25 years old, but is still kicking so why not reuse it!

 

[AudioJester]

Maybe I am an odd one out here, but how loud a speaker performs does not really interest me. In fact, I would go the other way and like to see how a speaker performs in typical domestic applications, with much lower than reference levels. My anecdotal experience suggests the frequency response can vary a lot at lower volumes with certain amps/speakers.
I have a typical multi-purpose area for listening, which I share with others. Max SPL is just not relevant to me.

Keep up the great work Amir, I can assure you that myself and others truly appreciate your efforts.

 

[MZKM]

I can understand that. It's just a bit of shame because, when making a purchase decision for a loudspeaker, as far as performance is concerned, typically there are two main questions: (1) does it sound good enough and (2) is it loud enough. If your reviews only answer (1) then we really only have half the information that's relevant for choosing a speaker. Yes, it's the most important half - but still. However I do get the point that one doesn't really need an anechoic chamber to do SPL measurements, so I guess this is a gap others could easily fill.

If speaker companies could themselves be transparent, then we wouldn’t need this conversation. I’ve seen ones stating:

Recommended Amplifier: 1000W​

It is just silly, most speakers can’t handle >100W without heavily distorting.

Although I praise Ascend Acoustics for their measurements and how good their top of the line models perform, their CBM-170 bookshelf states this:

Maximum Continuous Power:*: 200 watts​

Maximum Short Term Peak Power*: 400 watts​

*: Unclipped peaks​

One day a user told me they don’t even handle 10W without distorting, I obviously didn’t believe him, yet he linked me THD measurements at 10W (+10dB from sensitivity), and the distortion in the woofer was very high (luckily it has high sensitivity such that 10W is loud).

 

[restorer-john]

Anything else added to that which dilutes that mission is of low interest for me. Rude remarks, demands, etc. only diminish my interest there, not the other way around.

Please don't intimate or suggest I make rude remarks. I do not, and you know it. So do the members.

Again, this is a complaint thread you instigated, but it is abundantly clear your mind was already made up, complaints and suggestions are not what you actually want to hear, especially valid ones.

In all my time here, I have never "spat in your face" as you suggest. In fact the opposite, it's always been about furthering the usefulness of your reviews through constructive suggestions or constructive criticism.

I believe you cannot "test" or "review", let-alone rate loudspeakers, without determining their capabilities in terms of level (SPL), power compression behaviour and power handling- especially with the myriad miniature speakers people are buying (and you are reviewing) as main speakers these days. Many are simply not fit for the task and your "laser focus" on CEA 2034 spinorama measurements although extremely cool, tells them nothing about those aforementioned attributes.

 

[MZKM]

Please don't intimate or suggest I make rude remarks. I do not, and you know it. So do the members.

Again, this is a complaint thread you instigated, but it is abundantly clear your mind was already made up, complaints and suggestions are not what you actually want to hear, especially valid ones.

In all my time here, I have never "spat in your face" as you suggest. In fact the opposite, it's always been about furthering the usefulness of your reviews through constructive suggestions or constructive criticism.

I believe you cannot "test" or "review", let-alone rate loudspeakers, without determining their capabilities in terms of level (SPL), power compression behaviour and power handling- especially with the myriad miniature speakers people are buying (and you are reviewing) as main speakers these days. Many are simply not fit for the task and your "laser focus" on CEA 2034 spinorama measurements although extremely cool, tells them nothing about those aforementioned attributes.

I think one thing is that unless you just want to risk blowing a speaker by feeding it say 200W to measure THD, you’d have to increase wattage until you see a sharp rise, which may or may not take a long time with the Klippel (don’t know if it can do continuous THD measuring).
The other thing is, how would you standardize it, showing the max SPL until it reaches a threshold, or simply do a single measurement at say 100dB and report that? You also have to discuss audibility then, as you have people complaining about 5% THD at 50Hz.

I also suggested linearity/compression as SoundStage does this, but then it was reminded to me that most speakers measured show basically no compression, and the ones that do are usually <2dB in the low bass or upper treble; so it‘s not really worthwhile to set up that test for data that will be mostly similar.

You mention in regards to these small speakers people are now buying; well they usually are for near-field, and thus their max SPL @1m does not need to be as high as a speaker meant for a living room, as a ~12ft listening distance reduces SPL by ~8dB, aka ~1/6 the wattage.

 

[amirm]

I also suggested linearity/compression as SoundStage does this, but then it was reminded to me that most speakers measured show basically no compression, and the ones that do are usually <2dB in the low bass or upper treble; so it‘s not really worthwhile to set up that test for data that will be mostly similar.

I performed a stepped frequency response of the JBL 305P versus level and the curves visually were identical which supports what you state. I had to wear ear plugs for the highest levels and this is a modest bookshelf speaker.

 

[LTig]

Please don't intimate or suggest I make rude remarks. I do not, and you know it. So do the members.

You have a lot of valid points which I understand and often agree with. But I also note that the tone of your postings when you disagree is not friendly any more. It started with class D amps and continues here and does not help to keep discussions on a good level.

I thought long before I clicked the Post reply button since I value you high but I think it should be said.

 

[restorer-john]

I think one thing is that unless you just want to risk blowing a speaker by feeding it say 200W to measure THD, you’d have to increase wattage until you see a sharp rise, which may or may not take a long time with the Klippel

Low cycle count tone-bursts. The duty cycle is such that the drivers are not stressed and damage is unlikely.