# Selecting optimum frequencies for tone-testing

#### presscot

I’m bi-amping my speakers—Braun LS200. Each individual speaker has a bi-amp switch that will bypass woofers’ passive crossover. So, the active crossover for woofer-to-midrange is required.

To date, I’ve finished establishing the system. Next part is to perform the level matching between woofer’s amplifier and mid & high amplifier. I set the crossover point at 450Hz.

I know the process of level matching should be done by frequency sweep. However, as my amplifiers have been equipped VU meters. I’m thinking to use them instead of using the traditional frequency sweep method. And I’ve already done it. I randomly picked two test tones/ frequencies locating below and above 450Hz. Then, fed them to the amplifiers and adjusted level on each amplifier. After that playing musics to check for tonal balance. Repeat the process until achieve sweet spot. Finally, I ended up with 100Hz and 2kHz combination.

Although I’ve successfully obtained satisfying result. I still have a question. If I won’t use random frequency selection, what is the true method for selecting the optimum test tones/ frequencies?

Last edited:

#### GM

No clue per se, but to my way of thinking we would want a balanced sound power on each side of the XO, so for 450 Hz, I'd use at least 2 octaves on each side, so 450/2^2 = 112.5 Hz, 450*2^2 = 1800 Hz, so based on your derived solutions, looks like a reasonable way to calculate it.

1 user

#### jan.didden

Paid Member
Its an interesting question.
Because of room influence and speaker position, it may be that at a certain frequency they sound equal loud, while at other frequencies one or the other may be louder or softer. You can't tell without some measurement.
A good way is to do a swept test in the room at the listening position and looking at the curves for each speaker separately.
The best method probably is to do a room correction with REW or a similar program.
That guarantees equal and flat loudness over the audio frequency band.

Jan

Last edited:
1 users

#### mark100

Hi,
Sine waves don't work for setting acoustic levels indoors. They vary too much by location / by frequency, due to combing from reflections.
Also, amp output is not what you want to level. Acoustic output is the goal.

Best technique I know for leveling acoustic outputs, is use pink noise as signal,
(that will be bandpassed by the active crossover) and measure/match acoustic SPL's.

3 users

#### chinoy

Im in a similar situation. Had similar questions.
1. Whats the best song / music that has all the frequencies covered i.e. it has 30 Hz sections to 20 Khz signals in the music.
2. Why is it that if I match all the drivers to the same SPL. Using a calibrated mic. I find that some drivers like the CD just over power everything. I have to lower the CB output by a good 8 DB to get it into sync with the 15 inch woofer and 8 inch full range.

#### jan.didden

Paid Member
Hi,
Sine waves don't work for setting acoustic levels indoors. They vary too much by location / by frequency, due to combing from reflections.
Also, amp output is not what you want to level. Acoustic output is the goal.
Programs like REW use stepped/swept sine waves with processing and they can be used perfectly for this purpose.
I agree that what you want is acoustic levelling, not amp output signal level.
Again, programs like REW (there are many others, free or low cost) do this almost perfectly.
I use Uli Brüggeman's Acourate which is very good but is pay-for.

jan

1 users

#### rayma

Try both white noise and pink noise. You may find that one works better for you than the other.

1 user

#### mark100

Programs like REW use stepped/swept sine waves with processing and they can be used perfectly for this purpose.
Sure. Sine sweeps, log chirps, et al, work great for measuring response.

But OP stated he was trying not to use sweeps for setting levels, and trying some steady state sine waves. Steady state sines don't work for reasons previously given.

Simple fairly accurate levels settings, like OP appears to be trying to do..... can be easily done by downloading any of the myriad sources of pink noise, and using a SPL phone app (i suggest unweighted Z-weight).
Doesn't need REW etc (although highly advised, if for nothing more than the SPL meter)

1 user

#### jan.didden

Paid Member
Actually, if you think about it, the whole idea of levelling sound level is moot unless you have a flat acoustic frequency response.
If you have not, then levels can be very different at different frquencies do to room modes and such.
Levelling at one frequency will un-level at others.
Then again, if you do a room correction to get a flat response, you automagically have a levelled system.

Jan

1 users

#### mark100

If you have not, then levels can be very different at different frquencies do to room modes and such.
Levelling at one frequency will un-level at others.
Then again, if you do a room correction to get a flat response, you automagically have a levelled system.

Hi Jan, I've found even if you do have flat response from room correction, steady sine waves do not produce steady SPL levels.
A single sine is a 'no smoothing' signal. Out corrected room response curves are inevitably too smoothed to level a single frequency. (or they damn well should be
Plus single sine reflections and modes vary by the most minute mic movements.

Which also raises the issue that room correction is to a spot...or very narrow region at best.

Actually, if you think about it, the whole idea of levelling sound level is moot unless you have a flat acoustic frequency response.
Well, that where the idea of using pink noise comes in...to get an average integral of the response over the frequency range of interest.

REW's LEQ-Z SPL meter is great for this, when used with its noise generator.
(That measurement method is how driver sensitivity should be determined and specified imo/ime.)

1 user