Studio monitors response curves measured today with REW (input please)...

Yep. Note that the wall may be producing boundary effect, a low frequency boost caused by the speaker being close to a reflective surface. The switch on the back labeled "Room Compensation" counteracts the effect.
 
Yep. Note that the wall may be producing boundary effect, a low frequency boost caused by the speaker being close to a reflective surface. The switch on the back labeled "Room Compensation" counteracts the effect.
I wondered if that might be a possibility, even though the wall was at 45 degrees to the speaker faces, and underneath the speaker. Maybe draping a load of blankets over the wall, under the speaker would be worthwhile. Thanks also for your earlier suggestion re isolation, and trying the measurement with mic pointing at the tweeter.

My measurement microphone is omnidirectional, which I can see would be appropriate for measuring room resonances. However, when measuring speakers, outdoors, I'm wondering if a unidirectional mic would be more appropriate, as it would presumably pick up less reflected sound.
 
Last edited:
The effect of blankets won't extend low enough to make much of a difference. It would take specific materials to absorb LF.
 
You now have three perfectly sensible traces. I do think you are over thinking the entire thing, losing perspective. If you turn down the blue speaker a tiny bit it tracks perfectly, the green trace suggests the HF unit has a difference in HF response. Its a gradual tapering off, then a sudden spike. Being blunt, you are seeing things you cannot hear. I’d bet that that big spike at the end you cannot hear. When you listen to music on them that you know well, what do they sound like? The bass end starts to tail off, like all small speaker designs at the bottom. Hear it is 100Hz where it starts. If you want to mix music with lots at the bottom, then you have to guess the level of bass in your mix. I must be careful in my video studio. I often mix there and when i go home, i hear too much bass. My car hears too much bass on the journey home. Depending on the type of bass, as in synth bass, piano bass, double bass bass orchestral bass, or electric bass, i can pretty well adjust my choice, and get it mostly right, but bass at my audio studio is more accurate somehow - probably just bigger drivers and older design I guess.

Nothing you have, bar that measurement mic has a flat frequency response. Flat is boring. As you have discovered, small differences in your test produced very different and alarming results. Ask yourself the question ‘did the speakers sound as bad as the results, or was it a shock?’ Those little peaks and troughs did not alarm you. You didn’t think my god I've lost 30dB of level, or i have a huge peak at X or a total lack of signal at Y. Your ears and brain told you a sort of compensated result. The flawed tests have surely proven that none of your worries are the speakers. You can compensate. Forget trying to EQ it out. If you remove a tiny peak at 4.23KHz by taking 2.6dB out, can you hear it? Of course not, but flat line chasing is pointless.

1. Do you like the sound?
2. Do your mixes sound good on as many systems as you have access to?

If it is yes to both, move on. If it is no, borrow a new set and repeat.

Forget what your eyes tell you. All it has done is produce predetermination. You see a fault, and are fixing something probably that is just ‘character’. I have listened to a tape of the Beatles at Abbey Rd, on the studio speakers and original amp they used in the early 70s. It sounded absolutely terrible. Tinny, weedy, lacking every quality word we now take for granted. They would fail your test spectacularly.
 
You now have three perfectly sensible traces. I do think you are over thinking the entire thing, losing perspective. If you turn down the blue speaker a tiny bit it tracks perfectly, the green trace suggests the HF unit has a difference in HF response. Its a gradual tapering off, then a sudden spike. Being blunt, you are seeing things you cannot hear. I’d bet that that big spike at the end you cannot hear. When you listen to music on them that you know well, what do they sound like? The bass end starts to tail off, like all small speaker designs at the bottom. Hear it is 100Hz where it starts. If you want to mix music with lots at the bottom, then you have to guess the level of bass in your mix. I must be careful in my video studio. I often mix there and when i go home, i hear too much bass. My car hears too much bass on the journey home. Depending on the type of bass, as in synth bass, piano bass, double bass bass orchestral bass, or electric bass, i can pretty well adjust my choice, and get it mostly right, but bass at my audio studio is more accurate somehow - probably just bigger drivers and older design I guess.

Nothing you have, bar that measurement mic has a flat frequency response. Flat is boring. As you have discovered, small differences in your test produced very different and alarming results. Ask yourself the question ‘did the speakers sound as bad as the results, or was it a shock?’ Those little peaks and troughs did not alarm you. You didn’t think my god I've lost 30dB of level, or i have a huge peak at X or a total lack of signal at Y. Your ears and brain told you a sort of compensated result. The flawed tests have surely proven that none of your worries are the speakers. You can compensate. Forget trying to EQ it out. If you remove a tiny peak at 4.23KHz by taking 2.6dB out, can you hear it? Of course not, but flat line chasing is pointless.

1. Do you like the sound?
2. Do your mixes sound good on as many systems as you have access to?

If it is yes to both, move on. If it is no, borrow a new set and repeat.

Forget what your eyes tell you. All it has done is produce predetermination. You see a fault, and are fixing something probably that is just ‘character’. I have listened to a tape of the Beatles at Abbey Rd, on the studio speakers and original amp they used in the early 70s. It sounded absolutely terrible. Tinny, weedy, lacking every quality word we now take for granted. They would fail your test spectacularly.
Thanks, Rob. Two of the curves are acceptable to me. The green one isn't, but I think I can rectify that. I don't think I'm over-thinking anything. Okay, I may have geekish tendencies! My reason for wanting to do these speaker tests was to determine how accurate the speakers are, while still within the time window for returning them, if necessary.

When I first tried them out, I immediately noticed something that sounded amiss: The high frequencies seemed lacking. Now that I see the green curve on my latest test result, I think I see why. (It's roughly 6dB lower than it should be, through much of the upper frequency range.) I'm hoping to fix that by swapping two of the tweeters. If not, then the electronics may be at fault - in which case, I'll just use the two most accurate speakers out of the three.

Personally, I don't see flat as boring; I see it as exciting! :-) Flat may sound boring to the average hi-fi consumer, but for me, when mixing and mastering, I want as flat as I can comfortably afford, provided the bass extension and clarity is adequate for me (which, in the case of these speakers, it is).

For me this experiment was very worthwhile, as it taught me a lot (not least, thanks to helpful input from yourself and others). Most importantly, it convinced me that the speakers are suitably accurate, apart from the issue with the green tweeter, which I can now address.

This project is not totally done and dusted yet, though. The next step will be to run tests when they are in situ, in my studio. Before that, I need to install wall brackets, to raise them up off my desk, so that the desk becomes a bit less of a reflector and resonator.
 
Last edited:
If you want to scare yourself, repeat the tests in different places in the room.

I think I learned that (from doing this in theatres) interpretation is key - on it's own testing frequency response means very little. Pretty much now at a new venue I just play a track I know and listen. Very rarely do I need to dig out he measurement mic - most times I just think, bottoms a bit weak, or top is a bit harsh - and just do some basic output EQ. Chasing dBs on a screen showing horror stories I can usually live without.
 
If you want to scare yourself, repeat the tests in different places in the room.

I think I learned that (from doing this in theatres) interpretation is key - on it's own testing frequency response means very little. Pretty much now at a new venue I just play a track I know and listen. Very rarely do I need to dig out he measurement mic - most times I just think, bottoms a bit weak, or top is a bit harsh - and just do some basic output EQ. Chasing dBs on a screen showing horror stories I can usually live without.
I know what you mean. You surely have more experience than I do, so I respect your approach, if it works for you. I'm fully prepared for the horror show when I get the speakers in place and do more tests. It will be very revealing to compare the test curves from outdoors to the test curves done in the studio. It might even persuade me to move the studio to a better room. But probably, not, as they seemed fine when I tested them in the studio (apart from that treble deficiency I noticed). It remains to be seen how well my mixes translate, when I start mixing with them.
 
Last edited:
Back
Top