Edifier MR-4 review (by Erin)

[sweetchaos]

Thanks to @hardisj for another review.

This time it's the Edifier MR-4 (same model that Amir reviewed previously):

Obviously, loudness limited. Use nearfield only.

EDIT 2022-01-11:
Video review is up

Discuss!

 

[changer]

Erin’s HF is again showing more rise than in amirm’s measurement.

 

[abdo123]

Erin’s HF is again showing more rise than in amirm’s measurement.

Not really, it's the 200Hz region that is different. Everything else is kinda spot on.

 

[changer]

So I should have written VHF. It is the same frequency region that came under scrutiny in the review of the March audio speaker.

 

[hardisj]

Erin’s HF is again showing more rise than in amirm’s measurement.

Potential reasons:

• I'm using a different mic (yes, with calibration file)
• I have a modified boom that puts my mic at the end of a tube; Amir uses a towel wrapped around the mic holder at the end of the boom. I created a video on my modified mic boom so this is public info
• Exact positioning (if Amir is off even a little bit that could lead to differences in the VHF)
• Settings (I assume Amir used the same settings I did)
• Unit-to-Unit variation (it's a tweeter, it's beaming at this point). And it's in a $130 pair of powered monitors. So...

I'm not saying my setup is better so let's nip that in the bud. I'm saying those are the differences and could be contributors. These will always be potential differences anytime you compare anyone's results.

The issue with the March was resolved on the manufacturer's end. This was stated by the manufacturer himself and shared by me in the March thread.

In other cases Amir and my measurements have been right on. Other times they aren't. FWIW, even the CTA-2034 spec states that differences of ±1.5dB are perfectly within reason. I'm not sweating this stuff.


Edit: For proof, see my post below.

 

[changer]

I did remember what mtg90 had to say:

Erin's review for the March Audio Sointuva

For me it would be very interesting to send this speakers to the more subjectivist prominent reviewers. The outcome would be so interesting. What they hear, how they do descripe. Please do this March Audio !

www.audiosciencereview.com

and was wondering if Alan was actually able to track down that the difference was caused by his own gear and verify the measurement error was happening there.

 

[pierre]

You can see the 2 Edifier measurements on top of each other. Note that's not the same speaker too, same model yes.

 

[hardisj]

I did remember what mtg90 had to say:

Erin's review for the March Audio Sointuva

For me it would be very interesting to send this speakers to the more subjectivist prominent reviewers. The outcome would be so interesting. What they hear, how they do descripe. Please do this March Audio !

www.audiosciencereview.com

and was wondering if Alan was actually able to track down that the difference was caused by his own gear and verify the measurement error was happening there.

Yes. Alan tracked it down and I shared what he said in that thread.

Erin's review for the March Audio Sointuva

@hardisj why in your reviews there is no in-room response published any more? Time. I used to do that in every review at first but after I proved to myself that the PIR indeed matched (which I discussed in this video) then I got more lax. I still do from time to time. But I have to...

www.audiosciencereview.com

I quote:

What Erin's final test data revealed was that whilst it is very consistent with our own, there was a difference above about 7kHz with Erin's data showing 1 to 2dB higher amplitudes.

So, as some members have suggested, the design intent was to keep a smooth sound power and PIR. Our data indicated a small HF lift, but this certainly remained well within the +3dB limit band. None of our subjective test subjects found this to be excessively bright.

So working on the basis that Erin's data is absolutely accurate and there was something wrong with our testing regime until proven otherwise, I have been thoroughly examining test set up, calibrations and methods. Now let's be clear, *ALL* measurement systems have tolerances and accuracies, including the Klippel. No one should expect two disparate systems to produce absolutely identical results. The examination of our system has revealed a couple of reasons (one was very interesting in fact ) for about a 1 dB difference. We will be performing some further measurements tomorrow and I am expecting this to come very close to Erin's data, although I still suspect a fraction lower.

Meanwhile, the CTA-2034 literally says:

Appendix A. Data Acquisition in Non-Anechoic and Non-Open Field Spaces The following techniques are useful in assessing the frequency response of a loudspeaker in the absence of an anechoic chamber. Many of them work well. All measurements have errors, including measurements in anechoic chambers and outdoors, so it is important to validate your own techniques by comparing the results to a measurement made in a free field environment. A good way to do this is to submit one of your speakers to a recognized testing laboratory. An agreement of ± 1.5 dB is considered to be good.

It seems when a review is posted people focus on the tiniest of nits to drill on and discussion of the performance, design intent/application and practicality of the speaker goes out the window when that happens. Amir and I have a delta of 1dB. And we are talking about a resolution of 1/20-th octave. Have you folks ever taken the time to look at raw driver measurements and seen how much they can vary?! Seriously, if you haven't, you really should. I'll link some below just for fun. I've been doing this for a decade. Even the better brands have some level of deviation as you see below. And, yet, here we are, fussing over 1dB of difference on a set of $130 powered speakers that are likely not held to the highest of QC standards to begin with.

QC Variability in some high-end drivers:

Figure 6: BlieSMa T25S-6 two-sample SPL comparison
(SOURCE)

Figure 16: Purifi PTT6.5X midwoofer two-sample SPL comparison.
(SOURCE)

 

[RickS]

Agree, this speaker has a house response curve and likely uses a passive crossover with plenty of potential variance in components too.

Even with a better speaker, measurements are not likely to match exactly (even if using supposedly the same Klippel system). As @pierre mentioned, it is the same model of speaker, not the same speaker either.

All that is reasonable to expect is to compare Amir's measurements relative to his other measurements (and the same for Erin).

 

[changer]

It seems when a review is posted people focus on the tiniest of nits to drill on and discussion of the performance, design intent/application and practicality of the speaker goes out the window when that happens.

You must not be critical. We appreciate your work, yet it lies in the nature of presenting data that people will not only judge the result, but take close note if it was produced without error. I took note when mtg90 left his impression that the rise in VHF with your rig might have a pattern. If this is happening, it could interest you as well.

 

[hardisj]

You must not be critical.

Dude. Are you for real? Did you completely miss the part where I stated that I modified the boom on a $100k machine to improve the accuracy and made a video about it (where Klippel responded)?

I test and re-test everything. I perform ground plane measurements for the majority of large speakers to make sure things make sense.

Hey, here’s the GP check I did for the Elac center channel because the results looked wonky. Turns out the NFS results were fine.

“Not critical”,… man, you couldn’t be further off. Show me anyone who goes through more than I do to verify their data. Show me anyone who is as transparent with their process. You’re being ridiculous with that statement. I’m all for checking results and I go above and beyond what anyone would deem as necessary. Implying I’m not is showing your ignorance of everything I do to make sure I provide the most accurate data I can. Now, tell me what processes you go through when you publish data.

Edit: more transparency:

 

[hardisj]

Now, can we move on with the discussion of the actual data?

 

[changer]

I do not question you do your work most rigorously, I was commenting on the way you replied to my post. Take it or leave it.

 

[hardisj]

I do not question you do your work most rigorously, I was commenting the way you replied to my post. Take it or leave it.

And you can see I am critical and through. Well above what would be expected.

I provided a response outlining why there may be differences and showing proof of where March said he resolved the error in his system and then showed the CTA standard of tolerance.

Then you replied with “you must not be critical”. That’s taking my response and ignoring it.

Take my results or leave them. I don’t need to beat this dead horse anymore.

 

[TK750]

And you can see I am critical and through. Well above what would be expected.

I provided a response outlining why there may be differences and showing proof of where March said he resolved the error in his system and then showed the CTA standard of tolerance.

Then you replied with “you must not be critical”. That’s taking my response and ignoring it.

Take my results or leave them. I don’t need to beat this dead horse anymore.

I think maybe something has been lost in translation here? I get the impression what changer meant was more along the lines of 'don't take it to heart' or 'do not consider it a criticism of yourself'? Perhaps I'm misinterpreting though!

 

[changer]

If you leave this gifs people will react to it, what do you think? Although I can empathize that you got sensitive after the previous discussion, this is not a reason to get all excited. There was an observation by a third member, here was another result that has a tendency, I made a note. That is it, easy.

It was really nothing serious.

 

[Sancus]

Do people really expect the tweeters in two cheap speakers to measure the same? Because I hate to break it to anybody, but there's no such thing as precise quality control in a $130 monitor. Even ones that cost twice as much have significant variation sometimes.

 

[changer]

The possibility of sample variation is evident, this was not about it.

 

[hardisj]

Although I can empathize that you got sensitive after the previous discussion, this is not a reason to get all excited. There was an observation by a third member, here was another result that has a tendency, I made a note. That is it, easy.

Yes, but it was after I gave you rationale for why there might be differences that you said I must not be critical. The answer is not that I'm not critical. It's that I'm experienced and understand that a small delta between two datasets taken at completely different places is of no real consequence. Take into account that differences exist from time to time and you're essentially left with mtg90's measurement being different than mine (because I already discussed the other difference of March's). No offense to him but I'll trust my own measurements any day of the week until I'm able to verify how his (or anyone else's) were conducted. That's the nature of someone who has done due diligence and publicly provided the transparency I have. I don't think it should be taken as ego or anything of that nature. I'd hope/want a reviewer to be confident in his methods and the more transparent they are the more likely I am to believe the measurements. But, again, to bring this all back... we are talking about 1dB difference here and the CTA-2034 says even ±1.5dB is acceptable. My issue about these kind of things extends past your specific post; it's pretty rampant in these discussions where small nuances get drawn out and overanalyzed when the story the data is telling is more important than such differences.

If you meant no ill-intent then I apologize for reading too much into it. However, the chronology of replies I discussed above is why I took your comment to be critical of me and my process.

Water under the bridge at this point.

 

[abdo123]

1 dB difference is so small that you don't even need sample differences or component tolerances. it could be simply a temperature difference even.