acceptable 250Hz-20Khz range deviation?

**Drudge** · **07-10-09, 04:31 PM**

When it comes to speakers matching timbre with each other and using REW to look at the frequency plots, what is an acceptable deviation from 250Hz-20KHz between each speaker?

Not necessarily to a perfectly flat response,but to where the speakers seem to blend with each other without changing tone when sounds pan from one speaker to the other.

**Wayne A. Pflughaupt** · **07-10-09, 08:04 PM**

That's kind of hard to answer. If you have matching speakers all around, then they will pretty much sound timbre-matched, even with response deviations due to placement. Those response deviations can be corrected with equalization, using REW.

If your speakers are not matching, you can't accomplish timbre matching via equalization.

Regards,
Wayne

**Drudge** · **07-11-09, 01:40 AM**

I have 3 matching Mackie 824's up front which deviate about 1.5dB from one another in that range.In the rear I have 2 624's that deviate about 3.5dB from the mains.I'm wondering if the rears are deviated a little too much from the front.

They all follow the same general trend,but the rears are just a bit higher in the vocal range than the fronts.

Thanks for the response Wayne.

**Wayne A. Pflughaupt** · **07-11-09, 12:04 PM**

Quote:

In the rear I have 2 624's that deviate about 3.5dB from the mains.

That shouldn't be enough to lose sleep over, but if it's audible and it bothers you, you might want to try some equalization. Many receivers these days allow for separate front/back EQ, if you don't want to go the outboard route.

Regards,
Wayne

**hifisponge** · **07-19-09, 03:08 AM**

Quote:

Wayne A. Pflughaupt wrote:

That's kind of hard to answer. If you have matching speakers all around, then they will pretty much sound timbre-matched, even with response deviations due to placement. Those response deviations can be corrected with equalization, using REW.

If your speakers are not matching, you can't accomplish timbre matching via equalization.

Regards,
Wayne

Interesting. I had always thought that timbre was largely dictated by the frequency response of the speaker. If it isn't, then what affects timbre? The distortion characteristics of the speaker?

**Andysu** · **07-19-09, 08:01 AM**

If you had the one of matching fronts and moved a few inches over at time while performing each REW and looking at each graph (while getting out of the room so that you don’t interfere with the frequency sweep).

Then yes they will not mirror 100%. Like Wayne said a few db isn’t really enough to panic over that few tiny db is pittance.

But if you what to address the tiny little deviations go for it perfection

Timber! The meaning to me is like each of the LCR and one is mismatched the tone should sound like…Left HISS/ Centre hiss/ Right HISS.

The tone mismatches poorly with one speaker not just the HF but with bass/mid the signature is many db’s off.

If you feed the pink noise into an audio mixer LCRS and (balanced all the tones so that they each display on single loudspeaker for monitoring). It would sound like continuous PINK NOISE.

Adjust one channel by a few db it will sound off when it reaches that channel. Same if you decreased/increased LF/HF with EQ when it reaches that channel it will sound like its not a part of the community of tones.

If you had an all-channel PINK NOISE signal going into the mix with only single speaker to monitor each tone you add will double up in SPL db when read on an (SPL db metre).

If you altered one channel polarity/phase/EQ/level up or down/ it will display. The tone would most certainly make phasy dull pink noise whooshyness when switching polarity (ON/OFF/ON/OFF).

**JohnM** · **07-19-09, 02:17 PM**

Quote:

hifisponge wrote:

Interesting. I had always thought that timbre was largely dictated by the frequency response of the speaker. If it isn't, then what affects timbre? The distortion characteristics of the speaker?

A speaker does not have a single "frequency response". The response measured will vary with the distance from the speaker (depending on the number, separation and types of the drive units) and the angles relative to the speaker's main axis. Speakers with very similar on-axis responses can have very different directivities and hence very different off-axis responses, which will make them sound different in a room.

**Wayne A. Pflughaupt** · **07-19-09, 07:06 PM**

Quote:

hifisponge wrote:

Interesting. I had always thought that timbre was largely dictated by the frequency response of the speaker. If it isn't, then what affects timbre?

Sure, frequency response is the primary factor of timbre. I suppose if every speaker had perfectly flat frequency response they would all sound the same. But no speaker has perfect response.

Even if a speaker can generate a graph looks perfectly flat, the graph is only going to show the resolution of the measurement platform - 1/24, 1/60-octave or whatever. That means the measurement will only have (for example) 24 or 60 samples per octave. There can be (and will be) plenty of deviations from flat at frequencies between the samples.

That's why it's hopeless for an equalizer used on one speaker to duplicate the timbre of another: Even if you could find one with a gazillion filters capable of 1/60-octave correction, it still wouldn't be enough.

Bottom line, the best way to timbre-match speakers is to use matching speakers.

Regards,
Wayne

**hifisponge** · **07-19-09, 10:49 PM**

Quote:

JohnM wrote:

A speaker does not have a single "frequency response". The response measured will vary with the distance from the speaker (depending on the number, separation and types of the drive units) and the angles relative to the speaker's main axis. Speakers with very similar on-axis responses can have very different directivities and hence very different off-axis responses, which will make them sound different in a room.

Yes, I guess I oversimplified what frequency response means.

The polar response would have a significant effect on the sound no wouldn't it.

Thanks much,

- Tim

**Drudge** · **07-22-09, 03:43 AM**

I used REW's to make a "High Shelving" filter starting at 2Khz were the measurements deviate most noticeable from the fronts and adjusted the gain down -1.5dB on only the rears response.

According to REW's modeled response that puts all five speakers within about 1.5db of each other in that range.

So,I put the DEQ2496 in GEQ mode and switched it to HS mode which they call a paragraphic eq which is a graphic eq emulating a parametric at 1/3 octave bandwidth.I then adjusted the gain to -1.5dB on both channels.

I listened to some classical music and the over all soundfield sounded a little smoother and slightly more enveloping,but I'll have to wait till the weekend to test it on a movie and see if it is really an improvement or not.

I seem to notice the sonic mismatch on movies that pan the vocals more than anything.I notice it on some movies and not on others so maybe it's more a source material problem.

**Drudge** · **07-27-09, 03:01 PM**

I watched The GOLDEN COMPASS yesterday to test the front to rear panning and even though the movie stinks it has great sound.

Overall the tonal balance front to back seems to have improved and has a slightly smoother sounding transition than before.

So far that EQ adjustment has made a noticeable improvement.I just need to test it out on more material to make sure,but so far so good.

Advertisement