White noise vs Pink Noise vs Sine sweep for speaker compensatory EQing

alpha1

Active Member
Joined
Apr 30, 2009
Messages
588
Points
43
Location
βχανδυπ
All right, so the reason why this thread:
1) I have Quad 11L2 as well as Boston Acoustic A26 speakers
2) Initially when installed, I remarked to myself how the Quad sounded much louder than the BA, but soon got over it since both the speakers are in different rooms with different playing volume
3) Over a period of time I also felt that BA sounded too soft (in the higher mid-range) compared to Quads (very evident in overdriven guitars, saxophone and violin)
4) Did a REW log sine sweep analysis (https://www.roomeqwizard.com/) and saw the evident differences between the response of both the speakers (Quad is flatter across the spectrum, whereas BA dips considerably between 500 Hz and 7000 Hz)
5) Also did white noise and pink noise analysis to SEE the marked difference between both the speakers.
6) The aim is to use the Pulse Audio EQ on my computer to boost the mid-range of the signal going in to the Boston speakers to bring it as close to flat
7) The funny thing is that all three methods produce a different results (that is: which frequency and how much boost/cut)
8) Now my question is: shouldn't the compensation results be same irrespective of which method you choose (we would aim for a flat curve in case of sine wave or white noise, whereas we would want a -3dB per octave reduction in case of pink noise) since the source, the speakers, the mic remain the same
 
All right, so the reason why this thread:
1) I have Quad 11L2 as well as Boston Acoustic A26 speakers
2) Initially when installed, I remarked to myself how the Quad sounded much louder than the BA, but soon got over it since both the speakers are in different rooms with different playing volume
3) Over a period of time I also felt that BA sounded too soft (in the higher mid-range) compared to Quads (very evident in overdriven guitars, saxophone and violin)
4) Did a REW log sine sweep analysis (https://www.roomeqwizard.com/) and saw the evident differences between the response of both the speakers (Quad is flatter across the spectrum, whereas BA dips considerably between 500 Hz and 7000 Hz)
5) Also did white noise and pink noise analysis to SEE the marked difference between both the speakers.
6) The aim is to use the Pulse Audio EQ on my computer to boost the mid-range of the signal going in to the Boston speakers to bring it as close to flat
7) The funny thing is that all three methods produce a different results (that is: which frequency and how much boost/cut)
8) Now my question is: shouldn't the compensation results be same irrespective of which method you choose (we would aim for a flat curve in case of sine wave or white noise, whereas we would want a -3dB per octave reduction in case of pink noise) since the source, the speakers, the mic remain the same
No, they would be same only in an anechoic chamber
Your room reverberations will have a much bigger impact on the white noise curve.

Don’t forget that the goal of room correction is more of correcting the room response, esp at lower frequencies
Speaker curve calibration is a side benefit but most commercial room correction algorithms focus more on fixing the lower spectrum and leave the upper frequencies mostly untouched
 
No, they would be same only in an anechoic chamber
Your room reverberations will have a much bigger impact on the white noise curve.

Don’t forget that the goal of room correction is more of correcting the room response, esp at lower frequencies
Speaker curve calibration is a side benefit but most commercial room correction algorithms focus more on fixing the lower spectrum and leave the upper frequencies mostly untouched
Thanks, but as I mentioned, it is not the lower frequencies that are troubling me (of course there is a comb filter effect in my room but not my concern right now and will require additional expenditure like bass traps etc) but the mid and upper midrange correction is sorely required. What should I do?
 
Thanks, but as I mentioned, it is not the lower frequencies that are troubling me (of course there is a comb filter effect in my room but not my concern right now and will require additional expenditure like bass traps etc) but the mid and upper midrange correction is sorely required. What should I do?
I don't have much experience on using a PC for PEQ so I am presuming you have that part covered and all you need is the right correction curve that you can use on said PEQ

If you are OK with spending some money on the pursuit, I would strongly recommend the Paradigm PW Link Pre-amplifier/DAC (available on amazon locally) which comes with a calibrate microphone + microprocessor that runs anthem room correction . I use 3 of these now and the end result is way more than subtle

If you are looking at doing it purely via Software then keep it simple and use REW to:

a) hook up your connections and place the Microphone on a tripod facing up at your primary listening position (Which Mic are you planning to use? You need a calibrate mic to get good results from REW - along with of course the calibration file it comes with)

b) Use the level check to set the amplifier volume to where REW reads ~75db

c) Start the Measurement - REW will use a logarithmic sine wave sweep - start with 1 measure to begin with, once you get the hang of the procedure, you can always fine tune with multiple

d) Once the measurements are done, you can go to the EQ window to
1) Select your EQ type and adjust the target curve (EQ type will be mostly generic in your case, leave the target curve untouched in your 1st try)
2) CLick Match response to target to get the full list of PEQ adjustments that you need to make


Please bear in mind that REW based equalization will probably make things worse if you do not use the right Mic for it.
Even with a good mic, I have had a mixed experience with REW vs Audyssey XT32 and ARC
 
I don't have much experience on using a PC for PEQ so I am presuming you have that part covered and all you need is the right correction curve that you can use on said PEQ

If you are OK with spending some money on the pursuit, I would strongly recommend the Paradigm PW Link Pre-amplifier/DAC (available on amazon locally) which comes with a calibrate microphone + microprocessor that runs anthem room correction . I use 3 of these now and the end result is way more than subtle

If you are looking at doing it purely via Software then keep it simple and use REW to:

a) hook up your connections and place the Microphone on a tripod facing up at your primary listening position (Which Mic are you planning to use? You need a calibrate mic to get good results from REW - along with of course the calibration file it comes with)

b) Use the level check to set the amplifier volume to where REW reads ~75db

c) Start the Measurement - REW will use a logarithmic sine wave sweep - start with 1 measure to begin with, once you get the hang of the procedure, you can always fine tune with multiple

d) Once the measurements are done, you can go to the EQ window to
1) Select your EQ type and adjust the target curve (EQ type will be mostly generic in your case, leave the target curve untouched in your 1st try)
2) CLick Match response to target to get the full list of PEQ adjustments that you need to make


Please bear in mind that REW based equalization will probably make things worse if you do not use the right Mic for it.
Even with a good mic, I have had a mixed experience with REW vs Audyssey XT32 and ARC
Thanks for the comprehensive reply.
I would choose the microphone + REW route since that would be more economical plus the microphone can be put to use for recording music :) (my old Nady microphone is bust)

Should I choose condensor type or dynamic?
 
Thanks for the comprehensive reply.
I would choose the microphone + REW route since that would be more economical plus the microphone can be put to use for recording music :) (my old Nady microphone is bust)

Should I choose condensor type or dynamic?
Room calib is typically done with a condenser mic - but at the risk of sounding repetitive, the mic needs to come with a calibration file else the exercise will be mostly futile
 
Back
Top