How to read measurements: 2 speakers in 2 rooms
2 different speakers are evaluated in 2 different rooms, to show how measurements can be interpreted to give an indication of sound quality.
Some general conclusions:
- Better directivity reduces negative effects of room acoustics
- Proper acoustic treatment is required to achieve the best result regardless of speaker
- Speakers will still sound different in a good room
- Measurements can show if there is potential for improvement with better room acoustic treatment
- It is difficult to predict exactly how a system sounds from measurements
Loudspeaker directivity has a huge effect on sound presentation, and proper room acoustic treatment is mandatory for any speaker system to achieve top level performance.
Method
Listen to music on the 4 different systems, giving a subjective evaluation, with rating failed or passed.
A fail means there is no longer a believable rendering of instruments and recording venue. This will always be a subjective evaluation, there is no measurable and definitive limit for this failed or passed rating.
Each system is measured, and then analyzed.
NOTE - I did not do all the listening and measurements for this text. This data I already had. I did a short session, using the music listed, for the F205 (speaker A) in both rooms.
What is "good enough" - examining 2 speakers in 2 rooms
Can we see from a measurement whether the performance is good enough, that the systems performs at a level where further improvement will be more of taste and a different presentation, rather than better.
Comparing 2 speakers in 2 room can give some insight. In Room2, with good acoustic treatment, 2 speakers both provide a good presentation, with holographic rendering of images in a 3D-landscape. In the media room, one speaker is borderline acceptable, while the other fails.
The frequency response as usually presented, does not give any answers, as expected:
With 1/1 smoothing, we can see that there is no significant tilt on the frequency response, for any of the speakers, in any room. We can also see the fr differences caused by room-placement is much larger than the difference between the speakers. Green lines are from media room, red lines from Room2:
And adding one with 1/12 smoothing, to better visualize frequency response differences in detail:
If a simple frequency response measurement is sufficient to describe differences in sound between speakers, then those 2 speakers should sound the same, in the same room. But they do not.
So, what do we have here. 2 speakers, 2 rooms. We need a description. Starting with the rooms:
Media room:
Small media room with some basic acoustic treatment, listening distance 2.5m, distance between speakers 2.5m. Room has insufficient absorption in upper bass- low mid frequency range. No acoustic treatment on front wall close to speakers. (Description in the Media Room thread).
Room 2:
Small 2-ch listening room with acoustic treatment. Works quite well, good decay profile from 100hz and up. (Description in the Room2 thread).
Then we have the 2 speakers. We call them A and B.
Speaker A:
This is the F205. It has controlled radiation across its usable frequency range from 120Hz and up.
Speaker B:
The F105 is even smaller, and also has controlled directivity, BUT - the pattern is wider and shaped more like a traditional hifi speaker, high frequencies are closer to a direct radiating dome.
Both speakers have good descriptions on the product pages, with links to articles with further text.
Listening impressions
Music
Art Of Noise - album In Visible Silence.
Dire Straits - Six Blade Knife, from album Dire Straits, original Japan pressing not molested by the remaster idiots.
Hadouk Trio - Shamanimal, from Shamanimal album.
Flashbulb - Island on an Endless Plane.
Black Uhuru - album Chill Out.
Sera Una Noche - Tanguito, from Sera Una Noche album.
Speaker A Room2 - PASSED
The reference. In the sense that it may not necessarily be the best on all aspects, just that it serves a a common reference point to which all the other combinations are evaluated.
It is also better. There is a clear view into the recording, where instruments appear as physical objects, and the room information from the recording is preserved very well.
Tonality tends toward bright.
Speaker B Room2 - PASSED
This is fun. Small instruments placed accurately in a 3D landscape, like you are viewing everything from slightly above.
Dynamics and transient reproduction is not where this system is at its best, and it can not play very loud.
When listening far off-center, the presentation collapses into the nearest speaker. This does not happen with Speaker A.
This speaker does not sound equal to Speaker A. Still, given its limitations in ability to play loud, it is more of a different experience, with different presentation. Softer, smaller.
Unfortunately, this is not an easy solution for nice sound using very small speakers, because the room acoustic treatment required is so extensive. You can not achieve this by just placing these speakers in a living room with no acoustic treatment.
Speaker A Media Room - PASSED
It works, sort of. How good seems to depend a lot on the content. Acoustic instruments with acoustic room sound from the recording suffers most, rock and dub and electronica works quite well.
Everything is presented a bit closer, and a little wider. Clarity suffers, like there is a fog coming from far away, covering objects as they move deeper into the soundstage.
Six Blade Knife is not far from the presentation in Room2, nice presence of the vocal. But Room2 is better, the guitar is like there is a guitar present in the room, while here it is more like guitar sound, without the solid virtual image of the instrument.
Shamanimal looses clarity in lower midrange, like it is out of focus.
Sera Una Noche is where the shortcomings of the acoustic environment really shows. The magic illusion of being transported down to Argentina is simply not there.
Speaker B Media Room - FAIL
This is a fail. The presentation falls apart, it is now just like listening to two speakers.
Summary
Speaker A - Room2: PASSED
Speaker B - Room2: PASSED
Speaker A - Media Room: PASSED
Speaker B - Media Room: FAIL
Now we can examine the measurements, to see if there are differences that correlates with the listening.
Measurements show an interesting difference between the speakers:
- Speaker A has better early decay attenuation, but increased late decay energy, compared to speaker B.
Speaker A left/first, in Media room, speaker B right/last, in media room:
We see the lines for speaker A is a little more spaced out early in time, and then compress later in time, compared to speaker B.
And Room2 with better acoustic treatment actually show slower decay, compared to the media room, which is more dead.
This can be seen on the decay profile, speaker A, Room2 left, Media room right/last:
Early decay is faster in Room2, but late decay is slow. This shows up in the calculated RT60 value, which is 0.326s for Room2 and 0.163s for Media room. Then we realize the RT60 number is useless for the purpose of describing room acoustic properties of a small room.
Radiation pattern can be shaped to improve early reflection attenuation and at the same time increase late reverb from the room.
The differences we see in decay profile between speaker A and speaker B are caused by differences in radiation pattern. This leads to differences in sound signature, and can explain why speaker A barely manages to pass the listening test in the media room, while speaker B fails.
Spectrogram with 60dB scale and 200ms range, in Room2, show how speaker A has similar late energy compared to speaker B, and early decay is better for speaker A.
Spectrogram, 60dB scale, 200ms time range, Room2, speaker A left/first, speaker B right/last:
What about the room, and how can measurements show if there is potential for huge improvements.
Obviously, looking at the frequency response in a simple way does not give any answers. But if we look at the decay profile, we see large differences between the treated rooms and a typical non-treated space. The treated rooms have a much faster decay profile, and decay is smoother across the whole frequency range. Also, early reflection level usually is much better.
Look at this room, with F205 speakers, a small room, treated properly:
Now compare to this non-treated space, same speakers, and room correction has been implemented to improve tonality of the decay profile:
Now we have graphs that show a very visible difference, and this also correlates to the sound experienced in those rooms.