• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Help with REW EQ for desk setup - measurements attached

Gershy13

Member
Joined
Apr 18, 2022
Messages
71
Likes
6
Hi All,

So i've been trying to tune the sound at my desk setup for a while. I recently got a UMIK-1 and have taken some measurements with REW, however i am not sure what i should be correcting to, and what to ignore.

I will share some screenshots and the full mdat file if anyone would like to take a look and give me suggestions. I took some nearfield measurements as well just to see what the speakers were doing.

All measurements were taken with the MMM on a Umik-1. Speakers are Tannoy Mercury 7.2 (ports plugged to be sealed), Sub is an old 12" sealed car subwoofer. My DSP is an ADAU1701 from Wondom. I've attached a picture of what it looks like in the room, i know the setup is not ideal, but i am trying to get the best out of what i am working with. Room is roughly 2.5m X 2.5m X 3m. MLP is about 1m from each speaker.
My questions are: What should i cross the speakers to the sub at, if at all. What should i set my REW target to be, should i correct full frequency response.

Any help and suggestions would be appreciated. I have plans in the future to build DIY Speakers and subs, and thought this would be a good time to learn room correction.

Green: Sub, Blue: L, Orange: R
1735423838659.png
IMG_20241014_172937.jpg
 

Attachments

  • dec UMIK full set.zip
    149 KB · Views: 24
  • 1735424248276.png
    1735424248276.png
    219 KB · Views: 34
There is something really screwy with your measurements. No time domain information is available - can't see phase, impulse response, group delay, waterfall, or anything else. All I can see is amplitude response. There must be a serious error with your measurement. I am far from a REW expert, is someone able to tell him what he did wrong?
 
There is something really screwy with your measurements. No time domain information is available - can't see phase, impulse response, group delay, waterfall, or anything else. All I can see is amplitude response. There must be a serious error with your measurement. I am far from a REW expert, is someone able to tell him what he did wrong?
They are RTA measurements using the MMM method. Not sweeps.
 
That explains it.
 
Well, i'll give you a partial response then. It would be nice to see the distortion of your main speakers to help determine your XO point. At the moment, I would recommend about 120Hz because that's where your speakers fall off a cliff - probably because you have plugged the ports. 120Hz is very high for a sub XO, so it would also be nice to see its distortion products. If you XO a sub that high, you will be able to hear the location of the sub. Most people cross over the sub between 50-80Hz, and even then I tend not to like 80Hz. Your sub should ideally ONLY be making omnidirectional low bass frequencies. But at 120Hz, it's now doing midbass and that should squarely be the territory of your woofers.

Of course, I am speaking from an "ideal" point of view. It is not my intention to rubbish your speakers, we all have to make do with what we have. If you must do the XO at 120Hz, then do that for now until something better comes along.

I would recommend a standard sweep of your speakers with ports plugged AND unplugged for both L and R speakers, and another standard sweep of the sub alone.
 
Well, i'll give you a partial response then. It would be nice to see the distortion of your main speakers to help determine your XO point. At the moment, I would recommend about 120Hz because that's where your speakers fall off a cliff - probably because you have plugged the ports. 120Hz is very high for a sub XO, so it would also be nice to see its distortion products. If you XO a sub that high, you will be able to hear the location of the sub. Most people cross over the sub between 50-80Hz, and even then I tend not to like 80Hz. Your sub should ideally ONLY be making omnidirectional low bass frequencies. But at 120Hz, it's now doing midbass and that should squarely be the territory of your woofers.

Of course, I am speaking from an "ideal" point of view. It is not my intention to rubbish your speakers, we all have to make do with what we have. If you must do the XO at 120Hz, then do that for now until something better comes along.

I would recommend a standard sweep of your speakers with ports plugged AND unplugged for both L and R speakers, and another standard sweep of the sub alone.
Sure, I can do sweeps and post them too. Would this be NF or from the MLP?

If you look at the nearfield measurements, the speakers definitely play lower than 120 even when plugged, I think it's just the location and placement that makes it drop off a lot at my seat. Also they are rear porter bookshelf speakers so maybe they were designed to be placed against a wall. They are 6" drivers and I'm sure can take a decent amount of power if I were to EQ them to add in low end (or drop the mid area).

The reason I plugged the ports is because I thought that it would be easier to integrate a sealed sub as well as sealed speakers. I also heard that it is safer/ less harmful for the speaker driver when EQing (especially lower frequencies) if the box is sealed.
 
I think you need to get your speaker positions closer to the 'ideal' equilateral layout, with the tweeters at ear height (or pointing at your ears) before attempting to measure and correct with REW.

Unfortunately, it's really hard to get the speakers positioned correctly with a 3 monitor layout. I think your best option would be to place your speakers on stands (or mounted to the wall) behind your left and right monitors with them angled down so that the tweeters point at your ears when seated.

I found some nice diagrams on Adam's website...

1735468614439.png


1735468630204.png


 
I think you need to get your speaker positions closer to the 'ideal' equilateral layout, with the tweeters at ear height (or pointing at your ears) before attempting to measure and correct with REW.

Unfortunately, it's really hard to get the speakers positioned correctly with a 3 monitor layout. I think your best option would be to place your speakers on stands (or mounted to the wall) behind your left and right monitors with them angled down so that the tweeters point at your ears when seated.

I found some nice diagrams on Adam's website...

View attachment 417149

View attachment 417150

Would this make that much of a difference in the frequency response? I know the placement isn't ideal, but I thought this would affect the top end more than the mids/bass. Unfortunately I can't mount them to the wall, and I don't think there is enough space for stands on the desk.

They are close to ear level where I sit (simply because they are so big), maybe a little bit off, would it help if i put some foam or something underneath to tilt them up? The tweeters are pretty much on axis as I intentionally turned them to be tilted towards me.
 
and I don't think there is enough space for stands on the desk.
Would these fit?
 
Would these fit?
Interesting, I didn't think of desk clamp stands. That would work, but not sure if the platform that holds the speakers would hit the wall (the desk is pushed up against the wall). Unless there are any similar stands that have the flat part offset from the pole so it can be flat against a wall (I assume this would cause balancing issues tho)
 
Interesting, I didn't think of desk clamp stands. That would work, but not sure if the platform that holds the speakers would hit the wall (the desk is pushed up against the wall). Unless there are any similar stands that have the flat part offset from the pole so it can be flat against a wall (I assume this would cause balancing issues tho)

I modified a pair of monitor + laptop arms for my iLoud MTMs. They were on offer for £20 each when I bought them, so £47 in total with the screws I need for the bottom of the MTMs...

IMG_20240519_134053360 (Medium).jpg
IMG_20240519_134037155 (Medium).jpg

IMG_20241229_104712723 (Medium).jpg
 
Please take a quick look at this thread and make sure the measurements have been taken properly.

The nearfield responses of your bookshelf speakers are probably almost impossible to interpret if care has not been taken to window out reflections / or the speakers not placed so that it is possible to window out reflections in the first place. This is why I always look at the ETC to see whether there are any early reflections.

I would recommend that you leave the high freqs alone and only DSP the bass. For this you need a measurement from the listening position. Looking at your picture, it appears that you are using your speakers almost like headphones anyway - you are sitting so close to them.
 
Please take a quick look at this thread and make sure the measurements have been taken properly.

The nearfield responses of your bookshelf speakers are probably almost impossible to interpret if care has not been taken to window out reflections / or the speakers not placed so that it is possible to window out reflections in the first place. This is why I always look at the ETC to see whether there are any early reflections.

I would recommend that you leave the high freqs alone and only DSP the bass. For this you need a measurement from the listening position. Looking at your picture, it appears that you are using your speakers almost like headphones anyway - you are sitting so close to them.
I have followed the thread as best as I could when doing these sweeps, it should be in a similar chair location to the MMM measurements for comparison. I placed the umik on a mic arm and it was just a little in front of my nose (I was sitting in the chair during measurements). I used the 90 degree umik cal file, so i pointed the mic up.

For the sub measurements i muted the left channel of the speakers in the dsp and sent the output to the sub from the left channel of rew. the right speaker was used as the timing reference for everything.

I cannot attach the zipped file here as it is too big? 8mb! So here is a google drive link. https://drive.google.com/file/d/1m053JUiMDeZdmEAe3R0HBMaQIUQAEwSW/view?usp=drive_link
 
OK step 1. Is it a good measurement? First, we look at the energy-time curve to see if the measurement is contaminated by reflections. You can look at the rest of the measurements yourself. I am using "L Sealed -4dBFS" as an example:

1735645462489.png

Short answer: not really. It is contaminated by very early reflections which I have marked 1, 2, 3 for you. Reflection no. 1 arrives 0.16ms after the main impulse. Using the calculation d = (t/1000 * c) (t = time in ms, c = speed of sound) the reflection travelled an extra 55mm compared to the main signal. I am guessing table bounce, chair reflection, etc. This also gives us the reflection-free window. One period is 0.16ms, therefore all freqs below 6250Hz (1000/0.16) are irretrievably contaminated by reflections.

With DSP, the usual strategy is to:

- correct the anechoic response above the Schroder frequency. Or leave it alone.
- correct the in-room response below the Schroder frequency.

In short: this measurement can only be used to look at bass frequencies, and not for high frequencies. If you want to correct the high frequencies, you need to make a better effort at measurements without early reflections.

Next question: do we have a good signal-noise ratio?

1735646093973.png


Here is the same measurement, but with waterfall view. Note I have extended the window to 1000ms so the noise floor can clearly be seen. This is an "adequate" measurement - the signal is about 40dB above the noise floor, with two "fingers" at 119Hz and 201Hz. These are not room modes because they remain constant in amplitude and do not decay. My guess is computer fan (a 7200rpm fan will produce noise at 120Hz), air conditioning, etc.

1735647515510.png


And finally, we have distortion (this one is with the ports sealed). We can see that bass distortion is pretty high, but not so high that I would consider it problematic.

So, even before answering your question, we know the limitations of your measurement. You can use this measurement to look at bass, but not high frequencies. We also need to bear in mind that your measurement is contaminated by noise at 119Hz and 201Hz.

If you excuse me, my next response will be delayed because I need to do a few sims with your measurement.
 
I exported the measurements to Acourate so that I can do the sims.

1735648763670.png


Here we have left ported (red), right ported (green), and subwoofer (brown). I have normalised the volume of all the curves and overlaid them so we can see things more clearly.

First, look at the left highlight which I marked 1. We see that the subwoofer and the left speaker have the same room mode which is not affecting the right speaker. If you are using linear-phase FIR as basis of your DSP, you would be able to invert the phase of either the sub or the speaker in this specific frequency band to remove the cancellation caused by the room mode.

The second highlight marked 2 is also a room mode, in this case it appears to affect all 3 speakers. Again, inverting the phase of 1 or 2 of your speakers in this specific band will get rid of the cancellation. However, the problem is that it is rather high up the freq range, which means you have to cross your sub over > 100Hz if you want to tackle this with DSP. I do not like high XO freqs for subs - the reason is that bass becomes audibly directional above 80Hz, and if your sub distorts, the harmonic distortion products are even higher (e.g. the first harmonic of 80Hz is 160Hz, which is easily audible). I personally favour lower XO's as much as possible.

There are things your measurement does not tell me, for e.g. how loud you listen to your music and whether that volume is loud enough to produce distortion in your sub which would make its location audible.

In your case, my suggestion would be a sub XO at 80Hz which will tackle (1) but not (2). But you can also produce another set of filters with an XO at 120Hz.

My suggestion would be to do this:

- decide if you want to do speaker correction. For this, you need an anechoic measurement of your speaker down to a cut-off frequency. Below this, you will correct the speaker with the room. I suggest you aim for 400Hz which will roughly be the end of your transition zone. To get down to 400Hz, you can do the reverse calculation from my previous post. 400Hz has a period of 2.5ms. In 2.5ms, sound travels 0.86m. This means you need a minimum clearance of 0.86m from any reflective surface including floor and ceiling. If you can not achieve this in-room, take your speakers outside.

Note, for simplicity I am only discussing doing a correction of the axial freq response. It is more correct to measure the listening window and correct that.

- For subwoofer correction, the first step is to time align the subwoofer to the tweeter. Then perform all the amplitude corrections by doing a minimum phase inversion. There will be dips left over, these can be addressed by inverting the phase at specific frequency bands as discussed.

Good luck.
 
I exported the measurements to Acourate so that I can do the sims.

View attachment 417606

Here we have left ported (red), right ported (green), and subwoofer (brown). I have normalised the volume of all the curves and overlaid them so we can see things more clearly.

First, look at the left highlight which I marked 1. We see that the subwoofer and the left speaker have the same room mode which is not affecting the right speaker. If you are using linear-phase FIR as basis of your DSP, you would be able to invert the phase of either the sub or the speaker in this specific frequency band to remove the cancellation caused by the room mode.

The second highlight marked 2 is also a room mode, in this case it appears to affect all 3 speakers. Again, inverting the phase of 1 or 2 of your speakers in this specific band will get rid of the cancellation. However, the problem is that it is rather high up the freq range, which means you have to cross your sub over > 100Hz if you want to tackle this with DSP. I do not like high XO freqs for subs - the reason is that bass becomes audibly directional above 80Hz, and if your sub distorts, the harmonic distortion products are even higher (e.g. the first harmonic of 80Hz is 160Hz, which is easily audible). I personally favour lower XO's as much as possible.

There are things your measurement does not tell me, for e.g. how loud you listen to your music and whether that volume is loud enough to produce distortion in your sub which would make its location audible.

In your case, my suggestion would be a sub XO at 80Hz which will tackle (1) but not (2). But you can also produce another set of filters with an XO at 120Hz.

My suggestion would be to do this:

- decide if you want to do speaker correction. For this, you need an anechoic measurement of your speaker down to a cut-off frequency. Below this, you will correct the speaker with the room. I suggest you aim for 400Hz which will roughly be the end of your transition zone. To get down to 400Hz, you can do the reverse calculation from my previous post. 400Hz has a period of 2.5ms. In 2.5ms, sound travels 0.86m. This means you need a minimum clearance of 0.86m from any reflective surface including floor and ceiling. If you can not achieve this in-room, take your speakers outside.

Note, for simplicity I am only discussing doing a correction of the axial freq response. It is more correct to measure the listening window and correct that.

- For subwoofer correction, the first step is to time align the subwoofer to the tweeter. Then perform all the amplitude corrections by doing a minimum phase inversion. There will be dips left over, these can be addressed by inverting the phase at specific frequency bands as discussed.

Good luck.
Thank you for your very in depth responses! I have learned quite a bit from this. I believe i can do FIR filters on the adau1701.
I think for now i will skip the high frequency correction, but from what i understand i can take my speakers outside and measure them to get a better picture of what their natural frequency response is, and then correct it gently (wide filters/shelves only) if necessary?

My questions are:
1. If i have the subwoofer crossover at 80hz, I assume the speakers would also be set to roll off at 80hz, or should i leave them full range?
2. How would i go about creating the correction using the in room measurements, can i use PEQ from REW or is it better to do something with FIR filters (if i can on my DSP), if so, what would i use for that and how would I do it.
3. How would i time align the subwoofer to the tweeter.
4. What is the Schroder frequency of my room (is it 400hz?), or how would i find out where i should correct below.
5. Should i use the speakers sealed or ported.
6. Is the bad measurement and lack of bass response from the speakers because of the speakers or the room/placement. If i were to get better speakers, would i be limited by the location/room.

Thank you for your help, it is greatly appreciated. My goal is to learn about room acoustics and speaker correction, while also making my setup even with its limitations sound the best it can in its current setup. (it does not need to be perfect as i know placement and speaker quality is not ideal, but I would like to make the best out of what I have, until i upgrade.)
 
Thank you for your very in depth responses! I have learned quite a bit from this. I believe i can do FIR filters on the adau1701.
I think for now i will skip the high frequency correction, but from what i understand i can take my speakers outside and measure them to get a better picture of what their natural frequency response is, and then correct it gently (wide filters/shelves only) if necessary?

I will put it this way: for high frequencies, DSP should not be used for room correction. The reason is because DSP corrects for one extremely specific area (where the mic was placed) and is not a representation of reality. However, DSP can be used for speaker correction - provided the measurement is taken correctly. If you have high confidence in your measurement, and you are able to window out reflections, you can correct to your heart's content. I have showed you how to look at your measurement to see if you have early reflections.

1. If i have the subwoofer crossover at 80hz, I assume the speakers would also be set to roll off at 80hz, or should i leave them full range?

You can do anything you like. Here are some arguments either way:

- high pass speakers at 80Hz, low pass subs at 80Hz. Advantages: removes bass frequencies from main speakers. Reduces distortion. Increases apparent amplifier power and headroom. Low XO point for sub removes directional higher freq reproduction from the sub, making it harder to hear the sub as a separate speaker and making it sound more "unified".
- speakers full range, low pass subs at 80Hz. Having a substantial overlap between main speakers and subs increases the number of bass sources in the room, effectively "adding subwoofers". This will make it easier to achieve a more even bass field over a wider listening area. You can use MSO (Multi-Sub Optimizer) to tune out the dips in the freq response. If you have only one sub, it will be difficult to tune out a dip.

Which solution is best for you? The former is preferred if you have multiple subwoofers. But if you only have one sub, the latter might be a better option. I don't know. You need to experiment.

2. How would i go about creating the correction using the in room measurements, can i use PEQ from REW or is it better to do something with FIR filters (if i can on my DSP), if so, what would i use for that and how would I do it.

You need software capable of designing correction filters. There are a few on the market. The difference between them is how much automation they offer (more automation: easier to use) vs. how much control you get (more control means more to learn!). You can use REW + RePhase, but this has the least automation, and TBH it's not easy for beginners. But it's free. Acourate gives you slightly more automation and has the benefit of a few user guides (including one guide I am putting finishing touches on which I will release for free). Audiolense is also popular on ASR. Then there's Dirac, Focus Fidelity, and probably a few others.

Bear in mind that Acourate does not let you use USB mics. Audiolense does, but there are many reports on AL's forums of problematic measurements with USB mics.

There are very clever things you can do with subs, e.g. compensate for excess phase with DSP, adjusting the phase of frequency bands between speakers and subs to avoid cancellation, etc. Explaining the concept is difficult, let alone implementing it manually. But I will say this: when you get the bloody thing to work, and you see a nice flat bass response, you feel like a king. And maybe a bit smug ;)

3. How would i time align the subwoofer to the tweeter.

Every software has its own procedure, including some software that automatically time aligns the sub for you. In general: mic at listening position, then the subwoofer is swept with a timing chirp from the tweeter. The software "knows" that the timing discrepancy should be a known quantity, e.g. 0ms or whatever arbitrary delay the software uses. By measuring the distance between the subwoofer impulse and the tweeter impulse and comparing it to the known quantity, the delay is derived.

4. What is the Schroder frequency of my room (is it 400hz?), or how would i find out where i should correct below.

The calculation is:

1735669741309.png


Your T30 is about 300ms. I had a quick look at your measurements.

5. Should i use the speakers sealed or ported.

You can try both. In general, I don't like ported speakers because it introduces an out-of-phase rear wave which can not be independently manipulated with DSP. I noticed that when your ports are plugged, you lose about 5dB below 80Hz. Whether this is important depends on where you decide to do the XO. If you decide to high pass speakers + low pass sub, I would choose to seal the speakers.

6. Is the bad measurement and lack of bass response from the speakers because of the speakers or the room/placement. If i were to get better speakers, would i be limited by the location/room.

As long as your speakers produce bass, the frequency response is dictated by the room. ALL speakers, if placed in the same position in your room, will have the exact same pattern of room modes that you observed. It is a fact of life, all of us have to deal with it.

Thank you for your help, it is greatly appreciated. My goal is to learn about room acoustics and speaker correction, while also making my setup even with its limitations sound the best it can in its current setup. (it does not need to be perfect as i know placement and speaker quality is not ideal, but I would like to make the best out of what I have, until i upgrade.)

Buy yourself a copy of Toole and a copy of Mitch Barnett's book. My criticism of Mitch's book is that it is too Acourate-specific and does not include enough information about how to take measurements, or enough info about room acoustics. For room acoustics, you need Toole's book. Note that a new version of Toole's book is about to be released, probably in the next few months. So I would hold off on your purchase for now. As for how to take measurements, read this by the late Jeff Bagby.
 
I will put it this way: for high frequencies, DSP should not be used for room correction. The reason is because DSP corrects for one extremely specific area (where the mic was placed) and is not a representation of reality. However, DSP can be used for speaker correction - provided the measurement is taken correctly. If you have high confidence in your measurement, and you are able to window out reflections, you can correct to your heart's content. I have showed you how to look at your measurement to see if you have early reflections.



You can do anything you like. Here are some arguments either way:

- high pass speakers at 80Hz, low pass subs at 80Hz. Advantages: removes bass frequencies from main speakers. Reduces distortion. Increases apparent amplifier power and headroom. Low XO point for sub removes directional higher freq reproduction from the sub, making it harder to hear the sub as a separate speaker and making it sound more "unified".
- speakers full range, low pass subs at 80Hz. Having a substantial overlap between main speakers and subs increases the number of bass sources in the room, effectively "adding subwoofers". This will make it easier to achieve a more even bass field over a wider listening area. You can use MSO (Multi-Sub Optimizer) to tune out the dips in the freq response. If you have only one sub, it will be difficult to tune out a dip.

Which solution is best for you? The former is preferred if you have multiple subwoofers. But if you only have one sub, the latter might be a better option. I don't know. You need to experiment.



You need software capable of designing correction filters. There are a few on the market. The difference between them is how much automation they offer (more automation: easier to use) vs. how much control you get (more control means more to learn!). You can use REW + RePhase, but this has the least automation, and TBH it's not easy for beginners. But it's free. Acourate gives you slightly more automation and has the benefit of a few user guides (including one guide I am putting finishing touches on which I will release for free). Audiolense is also popular on ASR. Then there's Dirac, Focus Fidelity, and probably a few others.

Bear in mind that Acourate does not let you use USB mics. Audiolense does, but there are many reports on AL's forums of problematic measurements with USB mics.

There are very clever things you can do with subs, e.g. compensate for excess phase with DSP, adjusting the phase of frequency bands between speakers and subs to avoid cancellation, etc. Explaining the concept is difficult, let alone implementing it manually. But I will say this: when you get the bloody thing to work, and you see a nice flat bass response, you feel like a king. And maybe a bit smug ;)



Every software has its own procedure, including some software that automatically time aligns the sub for you. In general: mic at listening position, then the subwoofer is swept with a timing chirp from the tweeter. The software "knows" that the timing discrepancy should be a known quantity, e.g. 0ms or whatever arbitrary delay the software uses. By measuring the distance between the subwoofer impulse and the tweeter impulse and comparing it to the known quantity, the delay is derived.



The calculation is:

View attachment 417676

Your T30 is about 300ms. I had a quick look at your measurements.



You can try both. In general, I don't like ported speakers because it introduces an out-of-phase rear wave which can not be independently manipulated with DSP. I noticed that when your ports are plugged, you lose about 5dB below 80Hz. Whether this is important depends on where you decide to do the XO. If you decide to high pass speakers + low pass sub, I would choose to seal the speakers.



As long as your speakers produce bass, the frequency response is dictated by the room. ALL speakers, if placed in the same position in your room, will have the exact same pattern of room modes that you observed. It is a fact of life, all of us have to deal with it.



Buy yourself a copy of Toole and a copy of Mitch Barnett's book. My criticism of Mitch's book is that it is too Acourate-specific and does not include enough information about how to take measurements, or enough info about room acoustics. For room acoustics, you need Toole's book. Note that a new version of Toole's book is about to be released, probably in the next few months. So I would hold off on your purchase for now. As for how to take measurements, read this by the late Jeff Bagby.
Thank you for this very detailed information. I think i will just experiment with rephase and different outputs and see what sounds best to my ears.
 
I exported the measurements to Acourate so that I can do the sims.

View attachment 417606

Here we have left ported (red), right ported (green), and subwoofer (brown). I have normalised the volume of all the curves and overlaid them so we can see things more clearly.

First, look at the left highlight which I marked 1. We see that the subwoofer and the left speaker have the same room mode which is not affecting the right speaker. If you are using linear-phase FIR as basis of your DSP, you would be able to invert the phase of either the sub or the speaker in this specific frequency band to remove the cancellation caused by the room mode.

The second highlight marked 2 is also a room mode, in this case it appears to affect all 3 speakers. Again, inverting the phase of 1 or 2 of your speakers in this specific band will get rid of the cancellation. However, the problem is that it is rather high up the freq range, which means you have to cross your sub over > 100Hz if you want to tackle this with DSP. I do not like high XO freqs for subs - the reason is that bass becomes audibly directional above 80Hz, and if your sub distorts, the harmonic distortion products are even higher (e.g. the first harmonic of 80Hz is 160Hz, which is easily audible). I personally favour lower XO's as much as possible.

There are things your measurement does not tell me, for e.g. how loud you listen to your music and whether that volume is loud enough to produce distortion in your sub which would make its location audible.

In your case, my suggestion would be a sub XO at 80Hz which will tackle (1) but not (2). But you can also produce another set of filters with an XO at 120Hz.

My suggestion would be to do this:

- decide if you want to do speaker correction. For this, you need an anechoic measurement of your speaker down to a cut-off frequency. Below this, you will correct the speaker with the room. I suggest you aim for 400Hz which will roughly be the end of your transition zone. To get down to 400Hz, you can do the reverse calculation from my previous post. 400Hz has a period of 2.5ms. In 2.5ms, sound travels 0.86m. This means you need a minimum clearance of 0.86m from any reflective surface including floor and ceiling. If you can not achieve this in-room, take your speakers outside.

Note, for simplicity I am only discussing doing a correction of the axial freq response. It is more correct to measure the listening window and correct that.

- For subwoofer correction, the first step is to time align the subwoofer to the tweeter. Then perform all the amplitude corrections by doing a minimum phase inversion. There will be dips left over, these can be addressed by inverting the phase at specific frequency bands as discussed.

Good luck.
I plan to just use IIR filters for now and normal PEQ, until i can get a more capable DSP such as the minidsp 2x4HD which supports FIR filters better.
My adau1701 supports all pass filters, can i use an all pass filter to correct these issues in the screenshot?
 
I plan to just use IIR filters for now and normal PEQ, until i can get a more capable DSP such as the minidsp 2x4HD which supports FIR filters better.
My adau1701 supports all pass filters, can i use an all pass filter to correct these issues in the screenshot?

You need a reverse all pass filter to correct an excess phase peak. Reverse AP filters can only be obtained with linear phase FIR filters. A MiniDSP 2x4HD does have FIR filters, but these are severely limited in tap count - only 1024 taps. Low tap count means limited resolution of correction. You would need about 65536 taps as a minimum.
 
Back
Top Bottom