Amp for measurement affects speaker performance ?

[Crosstalk]

In many reviews I have seen that there are graphs of speakers but the associated test amps are not mentioned. Some speakers are high sensitivity and some low and some has weird impedances.

Will distortion figures of speakers change with change in amps?

For example here are two measurements of kef r11 by two different people and distotion is not Same in both.

KEF R11 quick measurements single vs bi-amp

Just wanted to share some measurements I got from my new KEF R11s. mic at 1 meter in room with all reflections active. REW measurements, MacBook to Mytek DAC+ to McIntosh MC462 vs. to Wyred4Sound (W4S) ST1000 (ICE 1000ASP) low (RCA) + MC462 HF (XLR) bi-amp. Using XLR out to the Mc and single...

www.audiosciencereview.com

Why is it so?

 

[RickSanchez]

This may not answer your question but for speaker reviews Amir uses the Behringer A500. His statement about this is here:

Complaint thread about speaker measurements

Near field or far? These ratings are for far-field, near-field will have less emphasis on on-axis and place more emphasis on the listening window. The HS7 will also have more bass than the LS50. Near field at 2ft away from tweeters as I use these as computer speakers. LS 50 doesn’t sound better...

www.audiosciencereview.com

(from the post above)

There were multiple reasons why I select this amp:
1. Klippel was adamant that a Pro amplifier be used for testing. So much so that they OEM a pro amp they sell for this purpose (I think it is QSC). Their rationale is that you need to have plenty of power on hand as to not limit speaker tests. For the purposes of them trusting any issues I may have with the system, I decided to use the Behringer.
2. I need to have balanced inputs to eliminate any chance of hum. Some of the consumer amps I have don't have this issue.
3. CEA-2035 stipulates that any amplifier with THD less than 0.5% is fine so here we are.
4. I don't listen to music with this amp in the lab. It is used to produce tones. It is doing something else than listening to music in a normal system.
5. I had it, it was sitting there eating dust. So now it has a good life.
6. It had to be something I could spare without adding more expense to this endeavour. Was not in a mood to stick my Purifi amp in there just to test speakers.

 

[Waxx]

Most speakers discussed here need an amp with a high damping factor/low Rout on the amp, but more esotheric or older speakers don't like that and need a low damping factor. Those are typical fullrange drivers or old drivers from the tube amp time.

So for modern speakers in the way discussed here a good class D or class AB amp is probally the best. Make sure the psu is up to the task and does not influence the sound (so a regulated linear or a modern overbuild smps) and the amp has very little to no distortion, even not harmonic distortion (so n tube or class A transistor amps). The Behringer that Amir uses is good as he is using it way under it's power specs (don't push that amp to it's limits, and it's quiet good).

 

[pjug]

I imagine there are all kinds of issues trying to compare distortion of anechoic (soundstage) vs in room measurements so I'll leave that alone.

Comparing the REW measurements MC462 vs W4S by overlaying, there does not seem to be much difference.

 

[Crosstalk]

So irrespective of amps if the speaker has a distortion, it will show up with any amp below clipping the same way. Or did I get it wrong?

Say a 30w amp is producing 100db on a speaker with low impedance at some points without clipping and 200w produces the same without clipping, they both should show same distotion graph ?

 

[mrmojo2022]

In many reviews I have seen that there are graphs of speakers but the associated test amps are not mentioned. Some speakers are high sensitivity and some low and some has weird impedances.

Will distortion figures of speakers change with change in amps?

For example here are two measurements of kef r11 by two different people and distotion is not Same in both.

KEF R11 quick measurements single vs bi-amp

Just wanted to share some measurements I got from my new KEF R11s. mic at 1 meter in room with all reflections active. REW measurements, MacBook to Mytek DAC+ to McIntosh MC462 vs. to Wyred4Sound (W4S) ST1000 (ICE 1000ASP) low (RCA) + MC462 HF (XLR) bi-amp. Using XLR out to the Mc and single...

www.audiosciencereview.com

Why is it so?

Speakers are an order of magnitude higher distortion than any decent amp so probably variations are insignificant, but obviously the amp needs to be well within its max power output range.
However you should pay attention to frequency response. if we are now looking for speaker responses to fall into a +-1.5dB window, then it matters. amps do have frequency response deviations which are compounded by difficult speaker loads and long cable lengths. DACs, mic pre amps and ADCs also have an impact, but admittedly probably insignificant. So if you are serious about accuracy you need to account for all of these variables and perform an "end to end" calibration of the entire chain. you shouldn't just assume an amp is OK. this means measuring a sweep at the speaker terminals, specifically the speaker you are testing

 

[mtg90]

Speakers are an order of magnitude higher distortion than any decent amp so probably variations are insignificant, but obviously the amp needs to be well within its max power output range.
However you should pay attention to frequency response. if we are now looking for speaker responses to fall into a +-1.5dB window, then it matters. amps do have frequency response deviations which are compounded by difficult speaker loads and long cable lengths. DACs, mic pre amps and ADCs also have an impact, but admittedly probably insignificant. So if you are serious about accuracy you need to account for all of these variables and perform an "end to end" calibration of the entire chain. you shouldn't just assume an amp is OK. this means measuring a sweep at the speaker terminals, specifically the speaker you are testing

I agree with this.

I picked up an Aiyima a07 after seeing it's review by Amir and when I went to measure a speaker design I had been working with it on I saw a large unexpected rise in the treble level. I measured the amplifier output at the speaker input terminals and saw it caused a 2+dB peak in the top end above 10k, so back in the box it went.

I tested the A500 I normally use the same way and it was for the most part flat, just a fraction of a dB down at 20k which is good enough for me.

 

[mrmojo2022]

I agree with this.

I picked up an Aiyima a07 after seeing it's review by Amir and when I went to measure a speaker design I had been working with it on I saw a large unexpected rise in the treble level. I measured the amplifier output at the speaker input terminals and saw it caused a 2+dB peak in the top end above 10k, so back in the box it went.

I tested the A500 I normally use the same way and it was for the most part flat, just a fraction of a dB down at 20k which is good enough for me.

Yes I've seen this sort of thing with several amps. some are simply not load invariant.

TBH (all due respect to Amir) the amp reviews are pretty limited in scope and the information they provide about real world behaviour.

 

[Holmz]

Amps clip in the time domain. So looking at the frequency domain becomes overwhelmed by the gfact that most of the time there is no clipping.
You pretty much need to align both sample and difference them to get some idea about clipping, HD, or crossover distortion.

 

[mrmojo2022]

Amps clip in the time domain. So looking at the frequency domain becomes overwhelmed by the gfact that most of the time there is no clipping.
You pretty much need to align both sample and difference them to get some idea about clipping, HD, or crossover distortion.

You might need to explain this a little more clearly.

Clipping HD and crossover distortion is quite visible in the frequency domain. looking in the time domain, such as using subtraction can give you a different perspective.

 

[Holmz]

You might need to explain this a little more clearly.

Clipping HD and crossover distortion is quite visible in the frequency domain. looking in the time domain, such as using subtraction can give you a different perspective.

The plot is post #4 is like “guess the song from the frequency response”.

But I agree with a tone, then one should see things easier.

This may be worthwhile (?):