• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Which speakers do the measurements say I should keep?

Kuppenbender

Active Member
Forum Donor
Joined
Aug 12, 2020
Messages
120
Likes
355
Location
UK
I really wish I knew what was doing with REW. I could then move on to the next step and generate some BruteFIR filters.

Here’s a pair of measurements from two different speakers. I know which speakers I prefer, and I know which ones I’m going to keep. However, to my (very) untrained eye, these measurements look pretty similar, even though one pair of speakers sounds significantly better than the other.
8F048C79-4763-4205-B435-32ED8E60EEB7.jpeg

I can’t really do much in terms of speaker placement or room treatment (it’s not a dedicated listening room) so I‘d also like an opinion on whether I’m on the right path with setting up REW measurements (if you can tell from the graph) for eventual room EQ settings.
I have FLACs on a music server running Daphile OS to a Soncoz sgd1 and an Audiophonics mpa-s250nc XLR (Hypex ncore 252mp).
 
OP
K

Kuppenbender

Active Member
Forum Donor
Joined
Aug 12, 2020
Messages
120
Likes
355
Location
UK
Excellent. So the measurements (probably) confirm my post-purchase rationalization, assuming the measurements were level matched.

The blue line is a pair of PMC Twenty5 21s (Amir tested the previous generation PMC Twenty 21, and wasn’t impressed). The orange line is an ex-display pair of Revel m16s I got for £380 (just over $500) a couple of days ago.
I’m keeping the Revels!

As far as ’level matching’ is concerned, if that means the db level I think one was 62db and the other 64db. I tend to listen at quieter volumes, as I live in a flat (apartment).

Thank you all.
 

DVDdoug

Major Contributor
Joined
May 27, 2021
Messages
2,916
Likes
3,831
As far as ’level matching’ is concerned, if that means the db level I think one was 62db and the other 64db. I tend to listen at quieter volumes, as I live in a flat (apartment).
It means that when you compare, you should be comparing at the same loudness. (I assume that should be done with pink noise.) Otherwise you'll generally just score the louder one as better.

Of course, it's OK to compare loudness too. But you want to isolate sensitivity/efficiency from the other characteristics.
 
OP
K

Kuppenbender

Active Member
Forum Donor
Joined
Aug 12, 2020
Messages
120
Likes
355
Location
UK
It means that when you compare, you should be comparing at the same loudness. (I assume that should be done with pink noise.) Otherwise you'll generally just score the louder one as better.

Of course, it's OK to compare loudness too. But you want to isolate sensitivity/efficiency from the other characteristics.
Thanks. I’ll tinker around a bit more, as I still have both pairs of speakers.
 

gsp1971

Senior Member
Forum Donor
Joined
May 26, 2021
Messages
471
Likes
821
Location
Europe
Both have been tested here in ASR. Check out the spinorama data, directivity, early reflections, distortion, etc.

The M16s seem a better bet.
 

GDK

Addicted to Fun and Learning
Forum Donor
Joined
Aug 8, 2020
Messages
560
Likes
1,525
Location
Toronto
I really wish I knew what was doing with REW. I could then move on to the next step and generate some BruteFIR filters.

Here’s a pair of measurements from two different speakers. I know which speakers I prefer, and I know which ones I’m going to keep. However, to my (very) untrained eye, these measurements look pretty similar, even though one pair of speakers sounds significantly better than the other.
View attachment 139471
I can’t really do much in terms of speaker placement or room treatment (it’s not a dedicated listening room) so I‘d also like an opinion on whether I’m on the right path with setting up REW measurements (if you can tell from the graph) for eventual room EQ settings.
I have FLACs on a music server running Daphile OS to a Soncoz sgd1 and an Audiophonics mpa-s250nc XLR (Hypex ncore 252mp).
The thing I find fascinating about this is how, for either speaker, the peaks and nulls are all pretty much lined up to 350Hz (where clearly the room dominates), but even past that up to 4kHz they are closely aligned.
 

Jdunk54nl

Addicted to Fun and Learning
Joined
Aug 5, 2020
Messages
968
Likes
1,048
Location
Arizona
The orange one once you superimpose a line to get a good visual reference definitely looks better to me.
Much less deviation from linear and a boost in the bass region.

View attachment 139480

Which one do you preferer, which one are you keeping?

That line is WAY too steep. You are losing nearly 20db from 20hz to 20,000hz. Usually you want about 0.6db - 1.0db per octave or about 5-10db total loss as you go from low to high frequencies.

It would help if OP would post one with 5db increments (40, 45, 50, 55, etc) scale on the y-axis.
 

Jdunk54nl

Addicted to Fun and Learning
Joined
Aug 5, 2020
Messages
968
Likes
1,048
Location
Arizona
OP, with you admitting that you don't know how to use REW, these measurements could easily be messed up and wrong info. For anyone to say based on these measurements of which one you should keep would be just guessing.

There are plenty of youtube videos, readings (including Amir's on this site) on how to use REW. Look up moving microphone method, as that will be the one you would want to use.

here is one:

This one has more on the setup of REW, it is set for car audio, but that doesn't matter when measuring. Also note, there are quite a few updates in REW eq section that helps too.
 
Last edited:

Trouble Maker

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
676
Likes
709
Location
Columbus, Ohio, US
The thing I find fascinating about this is how, for either speaker, the peaks and nulls are all pretty much lined up to 350Hz (where clearly the room dominates), but even past that up to 4kHz they are closely aligned.

It seems like orange has much less deviations from 'linear'.

1625604572850.png


Note: This was just a quick and dirty image to data trace tool grab, so the data isn't perfect, but the curves on this kind of scale should match pretty closely to the OPs without having their actual raw data.
 
OP
K

Kuppenbender

Active Member
Forum Donor
Joined
Aug 12, 2020
Messages
120
Likes
355
Location
UK
Both have been tested here in ASR. Check out the spinorama data, directivity, early reflections, distortion, etc.

The M16s seem a better bet.
I suppose I’m exhibiting the remnants of subjectivist scepticism. I’ve seen Amir’s test results, hence the purchase of the Revels. I asked for a ‘blind‘ opinion based on quick and dirty measurements for confirmation of my understanding of my measurements.
 
OP
K

Kuppenbender

Active Member
Forum Donor
Joined
Aug 12, 2020
Messages
120
Likes
355
Location
UK
OP, with you admitting that you don't know how to use REW, these measurements could easily be messed up and wrong info. For anyone to say based on these measurements of which one you should keep would be just guessing.

There are plenty of youtube videos, readings (including Amir's on this site) on how to use REW. Look up moving microphone method, as that will be the one you would want to use.

Brilliant! Thanks for those. I’ll take a look tomorrow.
 
Top Bottom