On a quest to improve my B&W DM640's

[Sixto]

Hello, my name is Sixto, and I live in Minneapolis, MN. and the first thing I want to say is Thank you John Mulcahy for making this tool accessible and supporting this forum for the audio community!.

11.20.2023 - If you are reading this thread for the first time, the real work starts after post #19…. that is when I realized my USB interface was faulty.

11.10.2023 - (Resuming original post). I am at the beginning of a quest to improve my in-home audio experience using the equipment and tools I currently own.
I am somewhat familiar with REW (i'm guessing beginner level...used it on-and-off over the past few years to measure and compare a number of speakers I own).
My provisional goal is to bypass (Remove?) the existing (original) crossover in my speakers and use REW, possibly Rephase and Minidsp to get to an Active Hybrid, Biamped setup with better fidelity for stereo music playback. I already have tons of questions but will pace myself and try to document the process, generating appropriate questions for each step along the way.

The attached files illustrate my testing setup, and the first successful attempt at a baseline measurement. Microphone and Tweeter aligned at ear-height, and 36" apart. The first measurement (attached below) is a full-range sweep from 0 to 24KHz no-smoothing. It is not nearly as good as I had hoped, so my work is clearly cut-out for me.

My next steps are to bypass the factory crossover and measure the twin 8" woofers (8ohms stacked and wired in parallel) as a pair. I have also blocked the rear port with a roll of compressed foam because the bass coming from the port was too boomy in my Livingroom.
I will also separately measure the Midrange and the Tweeter drivers to get confirmation of how they perform, though I'm hoping to be able to join the two of them as a pair with a passive XO so I can use a pair of two-channel amps to Biamp the Left and Right speakers.

Once I get those measurements, I plan to do some tweaking (such as aligning step-response, removing echoes and reflections as much as possible, adjusting phase, smoothing or flattening the FR curve to match a desired house-curve, generate some EQ/Phase filters, import them to MiniDsp, tweak gain as needed to match SPL between the two main sets of drivers and measure again (probably repeating this several times) to confirm results. This approach is based on guides and recommendations that I've read and watched in various forums and videos)

I have the following hardware and software:
B&W DM640 speakers (aka the DUT"s)
Sony Fullrange speaker used for timing reference.
NAD 2600A amplifier (150 wpc)
Focusrite Scarlet 2i2 USB Interface
Dayton EMM-6 microphone with calibration file
Laptop computer with REW Rephase, and Minidsp console.

I guess my first question is: does this sound like a good plan? Even if my speaker choice is not ideal, I'm hoping this will be a rewarding learning experience that I can apply to any future system. I also haven't seen many step-by-step examples of this process, so I hope this thread can serve as a guide for future DIYers on a similar quest. I look forward to all helpful suggestions from forum participants!

 

[steve138]

Hi Sixto. Steve (a Brit) here by the Black Forest in Germany. There's something not right with your sweep.
Did you set the output of your amp to give about 85dB at 1m? I always measure at Industry standard of 1W at 1m. To set the output on my PC and pre-amp I connect a voltmeter across the speaker plugs terminals while they're connected to the speaker and find output levels that show around 2.83 V. Using Pink noise from REW the output moves around a bit so I mean to try it with a 1kHz tone next time.
Then after running each sweep response go to Impulse and IR Windows and set a gated time window to remove the effect of nearest room reflections. I typically have Left Width at 0.03 and Right width at 4.6ms for full range sweeps of the whole speaker and tweeter driver measurements and 0.06 and 4.0ms for justwoofer measurements but you should adjust so you only have the impulse without the first reflections on the Impulse chart.

If not happy with your B&W's (and I can understand that). I think the best thing you can do being in the US is send one to Danny Richie at GR-Research in Iowa Park, Texas as he does loads of speaker upgrades and fixes. Check out the YouTube cannel. If you view the channel in the browser on yr PC you can search just within the channel for B&W and see all the existing upgrades / fixes he's done. Sadly there's a lot of recurring issues in B&W responses. See also the many many Klipsch fixes which are way worse.
From Danny you'll get a proper flat response, good off axis consistency, better quality crossover parts (way better than the standard from factory cheese), and we'll get another video. What he usually does in these cases is turn them into a 3 and a half way, leaving one Bass woofer just playing bottom end, rolling it into the upper woofer which then crosses to the Mid. At least that's what happens with the many Klipsch models that use two large woofers that were originally playing the same signal.... causing big suckouts in the response due to phase issues and comb filtering. If the cabinets are a bit too resonant, line them with GR-Research No-Rez to stop the cabinet resonance and standing waves.
Obviously you'd need to pay for shipping - but he does the testing, crossover design for free and then you just pay for the new crossover parts kit, as he earns his time back by selling the upgrade to other owners of the same model. And he'll do a sho-huff better joib than you could on your own. Just write them an e-mail.
Danny also does a lot of Tech Talks that give background on various aspects of speaker performance and crossover design. I learned a lot and have since redesigned and built crossovers for 3 pairs of Linn Tukans as wall mounted surrounds in my home Cinema room and 2 pairs of Linn AV 5110's as Atmos Height speakers. Good luck

 

[Sixto]

Then after running each sweep response go to Impulse and IR Windows and set a gated time window to remove the effect of nearest room reflections. I typically have Left Width at 0.03 and Right width at 4.6ms for full range sweeps of the whole speaker and tweeter driver measurements and 0.06 and 4.0ms for just woofer measurements but you should adjust so you only have the impulse without the first reflections on the Impulse chart.

Hi Steve, thanks for the response. I was just about to ask the forum about how best to gate the individual driver measurements?

1. My mic is 36" from the speaker, 48" from floor, 60" from ceiling, and 40" from nearest wall, so a RT window between 4 and 5ms sounds right to reduce reflections. (havent done that yet)
2. Sweeps are as follows: Woof 0-3K hz; Mid 60-10K; Tweet 1k-24K
3. Since I'm trying to time-align 3 individual drivers, What Window type should I use? (Rect, Hann, Tukey, Blackman, etc?)
4. Since the mike and speaker are in the same spot for all 3 drivers, can I use the same window type and L/R widths for all 3 sweeps?
5. Should I select the Frequency-Dependent window checkbox to narrow the window in the higher frequencies? and how many cycles should the window width be?
6. Finally, it looks like after a sweep has been made, the window for that sweep can be altered. So once I settle on the right settings, can I apply them to the existing sweeps, or do I need to make new sweeps with the right IR Window settings?

I'm attaching PDFs and measurements for the woofer pair, midrange and tweeter.

I also read this in the help file: "If you will be comparing measurements from several speakers, or comparing a series of measurements from a speaker, make sure they are measured with the same Sweep Level." I had to boost the tweeter Sweep level from -8 dbfs to -4 dbfs to get a high enough SPL for the mic. Should I re-do my midrange and woofer sweeps at that level, or can I offset (reduce) the SPL by X db for the Tweeter sweep?

Lastly, thanks for your recommendation, my quest is probably as much (maybe even more) about learning the process than improving the results. So, I appreciate your thinking, but sending the speakers to a professional is not something I'm interested in at the moment.

 

[Sixto]

Hi Sixto. Steve (a Brit) here by the Black Forest in Germany. There's something not right with your sweep.
Did you set the output of your amp to give about 85dB at 1m? I always measure at Industry standard of 1W at 1m. To set the output on my PC and pre-amp I connect a voltmeter across the speaker plugs terminals while they're connected to the speaker and find output levels that show around 2.83 V. Using Pink noise from REW the output moves around a bit so I mean to try it with a 1kHz tone next time.

Yes, my speakers are at 75 DB measured with a meter at the mike when I check the levels prior to doing a sweep, but I believe that since I'm using an EMM-6 connected to the Focusrite Scarlett 2I2, the sweep level I get is around -30 db... and that's setting up the USB preamp interface settings and mic gain as high as I'm comfortable with. The sound coming out of the speakers during the sweeps is ear-splitting.

Thanks - Sixto.

 

[steve138]

your time windows aren't right. see the values I gave that i use and look at the Impulse tab in REW you should see how the time gate is framing the sweep. Left side value should be around 0.06 and right side value around 0.4 ms. when you do this those charts you show above won't be all spiky but will appear as a cleaner trace which you can smoot with 1/3rd smoothing. as you apply these changes you can apply to all sweeps in the open set or adjust individually - you do not need to run more sweeps unless you ran them wrong to start with. I have my settings at -12dBFS n REW, PC output at 61 and my Linn Kairn preamp at 83 into a Linn Chakra 4100 power amp. that gives me around 2.83V at the speaker plugs and 94dB SPL at the mic at 1 m on tweeter axis.
to the right side of the Impulse chart it needs to reach the horizontal axis by the first reflection. 36" is around 1 metre so you shouldn't be far off but the images of your sweeps are still way off expectations. On tweeter axis you should be getting smthg that looks fairly horizontal on a full range sweep. I don't see that yet. try to get as far from side walls as you can and any surround objects. make sure you're behind the mic or spkr when running a sweep.Once you have a decent time window you can use 1/3rd smoothing on the curves. The gated method will only give you a response down to 100 - 200 Hz at 1 m. Maybe possible to get to lower frequencies with mic closer to drivers but I haven't tried that as I did all mine at 1m. I ran my drivers separately and made separate sweeps and exported the freq responses as .frd files and ran the Impedance swees using a Dayton Audio DATS v3 and exported those as .zma files and pulled those into XSim to model the crossover responses, with varying degrees of success. just to give me a starting point which I thin refined by adjusting parts values clipping together with crocodile clip leads outside the cabinet.

I understand wanting to learn as that's what I'm doing but I'm only working on small bookshelf speakers for Home Theater surrounds. for more serious speakers for main stereo system listening - I'd want a more perfect result, and no matter how much I tried to learn I would not compete with someone who does it for a liing, both upgrading speakers and designing their own - which he sells as DIYkits. Check out the videos you'll learn a lot from them. see the Tech talks on crossovers and the one called Fixing your DIY projects. and any of the B&W and Klipsch upgrades - especially the towers.
Please post updates after fixing your sweeps and time gating.

 

[steve138]

oh right I'm using a UMIK-1 directly into the PC, I bought a used Focusrite and a new EMM-6 but have not used them yet.

 

[steve138]

This is what I'm getting with 1W at 1m for Woofer and Tweeter and full crossover response on one of my Linn Tukans. Time gate as I described with 1/3rd smoothing

 

[sam_adams]

I was just about to ask the forum about how best to gate the individual driver measurements?

This is your un-gated, raw measurement of the tweeter:

You really don't need to set the Left window width to more than the default of 125ms.

The below is the same tweeter measurement with 3ms gating applied to remove all reflections greater than 3.375ft away. Technically, 1ms is 1.125 feet or 13.5 inches, so 3ms is 40.5 inches or just over 1m (39.4 inches). Note the changes in the Left and Right windows widths as-well-as the frequency resolution and span in samples.:

The discussion on ASR regarding how to make quasi-anechoic measurements with REW may be technical, but it does contain quite a bit of useful info.

 

[steve138]

see here for crossover diagram and values:

Upgrading my B&W's

I am upgrading my DM 640's and because they are "old" some are saying that they are dated. What would you say is the speaker systems weakest link? Bass driver, Mid, tweeter? I am going to upgrade the XO parts. I have my idea of where to make some mods and I wanted to see if your opinions were...

www.diyaudio.com

Refurbish crossover B&W DM640

Hi there, I have my DM640's for a long long time by now. A few years back I replaced the ferrofluid ad put them on spikes. Other than that, they are as they came from the factory. They are almost 30 years old by now and I thought it to be a good time to refurbish the crossovers. The new parts...

www.diyaudio.com

Tweeter has 7mf (C4) and 16mf (C5) caps and a 0.15mH inductor in shunt (ie a third order filter)

 

[Sixto]

Thanks for all the responses! I've been digesting the feedback, reading the REW Help file, plus more reference materials, and updating my sweeps.

Re-did measurements with higher (and consistent) SPL calibration and Measurement Level set at -6 DBFS for all drivers. After exploring the effects of various window types and timing, I settled on the Blackman-Harris 7 window with 3ms Left Window and 5ms Right Window for the Tweeter and Midrange with an increase in time to the Right Window for the Woofers at 62ms to keep frequency resolution down to 16.13hz. (I did not check the Frequency Dependent window box because that gave me a less smooth SPL trace above 10K hz. I also Used Psychoacoustic smoothing on all 3 traces.

I still have about 20db difference in sensitivity between woofer and tweeter, but I expect that will be mostly fixed after I combine and EQ the measurements. Plus I can always tweak SPL level with the MiniDSP later.

At this point, I'm favoring a DSP crossover between Woofers and Mid between 200 and 300 hz., and also adding an analog XO between the Mid and Tweet at between 1.5 and 2.5k. This will allow me to Bi-amp the woofers and mid/tweet with a 2-channel amp on each side.

Next steps are to explore Alignment Tool and Trace Arithmetic buttons in the Controls wheel to get a single trace that combines all three individual drivers prior to using EQ to flatten the curve closer to a Harman house curve. This may take some time.

I'm cheering myself onwards, while my wife is wondering HOW LONG am I going to keep making these loud noises! (Goal is to have something done by Christmas)

Gratefully yours, Sixto

 

[Sixto]

see here for crossover diagram and values:

Upgrading my B&W's

I am upgrading my DM 640's and because they are "old" some are saying that they are dated. What would you say is the speaker systems weakest link? Bass driver, Mid, tweeter? I am going to upgrade the XO parts. I have my idea of where to make some mods and I wanted to see if your opinions were...

www.diyaudio.com

Refurbish crossover B&W DM640

Hi there, I have my DM640's for a long long time by now. A few years back I replaced the ferrofluid ad put them on spikes. Other than that, they are as they came from the factory. They are almost 30 years old by now and I thought it to be a good time to refurbish the crossovers. The new parts...

www.diyaudio.com

Tweeter has 7mf (C4) and 16mf (C5) caps and a 0.15mH inductor in shunt (ie a third order filter)

Thanks Steve! - I had forgotten I had posted on one of those forums when I first got the speakers (2017) I found 2 replacement woofer drivers and replaced the ferrofluid on the tweeter based on the recommendations from members then. I do have one of the crossovers out now, and will see if I can test components and determine where the mid-to-tweet XO is at, and might just reuse that or tweak it for the analog XO. One of the comments said typically B&W likes to cross high at around 3Khz. Based on my latest measurements, I think it could go lower without damaging the tweeter.

Sixto

 

[Sixto]

The below is the same tweeter measurement with 3ms gating applied to remove all reflections greater than 3.375ft away. Technically, 1ms is 1.125 feet or 13.5 inches, so 3ms is 40.5 inches or just over 1m (39.4 inches). Note the changes in the Left and Right windows widths as-well-as the frequency resolution and span in samples.:

The discussion on ASR regarding how to make quasi-anechoic measurements with REW may be technical, but it does contain quite a bit of useful info.

Hi Sam, Thanks for the feedback and link. I was able to test various windows and timing based on what you showed. I ended up re-doing the sweeps for other (SPL Level) reasons, but I like the ability to change my mind on IR windows on any given sweep, and being able to get back to the original trace if needed!.

Sixto

 

[sam_adams]

The response of the tweeter seems seriously deficient. Given their age, there is a good likelihood that the ferrofluid is dried out. You should remove them from the speaker and do an impedance check on them to see if that is the case. Good HiFi's Speaker Repair Shop site has a how-to on replacing the ferrofluid in those B&W tweeters. In the how-to they mention using tape to clean the dried fluid from the gap. What works better is coffee filter paper with some 99% Isopropyl alcohol on it to remove any dried residue. You must also clean any residue from the voice coil of the tweeter itself using some alcohol and cotton swabs. For both the gap and the VC, keep wiping until the paper and swab are clean.

 

[Sixto]

The response of the tweeter seems seriously deficient.

Hi Sam, I agree with you, and what's really disappointing is that I actually had the Ferrofluid replaced just before I started this journey. In fact, these sweeps are the first time I've used the repaired tweeters since. The shop that did the work is a very reputable speaker repair shop in the Twin Cities, so I had high hopes. Maybe they need to break-in?

I'm still thinking I will go through the process with the drivers I have, even if it requires reducing the power to the woofers and tweeters to match... Watts are cheap these days. (but so are aftermarket tweeters) Anyways, looking at it from another angle, maybe I have overly sensitive mids and woofs!

I'm posting the DATS sweeps I did when I had the drivers installed in their own (right size for DATS) sealed box. This confirmed that I can almost go down to 1500Hz in the crossover between the Mid and the Tweet (Fs is 1544).

Question:

• Should I fabricate (and run sweeps with) the passive XO installed between the Mid and Tweet before moving forward with alignment and EQ, or if I can proceed to do EQ based on the assumption that I'll get a smooth transition between those two drivers?

Thanks! - Sixto

 

[Sixto]

While I wait for a response to the question in my previous post, I thought I'd experiment with the Alignment Tool button on the ALL SPL Controls Tab. My goal was to combine the individual Woofer and Midrange measurements so I can generate filters to export to MiniDSP for the digital crossover.

Attached is a clip of how I set-up the Alignment Tool:

1. Reduce the gain of the Woofers measurement by 5.2 db. to get better spl alignment to Mid.
2. Placed my cursor at 251 Hz, where I plan to set the MiniDSP crossover from Woofer to Mid.
3. Select "Level Phase at cursor" button to level the 1st trace (woofers) Phase.
4. Select "Align Phase at cursor" button to align the 2nd trace (midrange) Phase to the woofers. (This generated a .5 ms delay for the midrange - I'm assuming based on the acoustical timing references shared by both measurements)
5. Select "Aligned Sum" to create a new measurement that is a summed copy of the selected drivers's measurement with the alignment gain, delay and polarity selections applied. (Paraphrasing the HELP file)

And Voila! I believe the new "Aligned Sum" measurement will be my basis for going to the next (EQ) stage... and attempting to flatten this curve to more closely correspond with a Target Curve...

Question - am I on the right track so far?
Thanks - Sixto

 

[steve138]

I would not use psychoacoustic smoothing as it will give you an inaccurate impression. Apparently 1/3rd octave is industry standard so that's what I use And I didn't try different types of filters like Blackman-Harris I stuck with whatever the default was (Tukey 0.25). I need to read up on the differences and what may or may not be relevant. but Balackman-Harris doesn't get me any lower in resolution frequency. I also didnt find explanations of them in Help.
On the Impulse tab your Y axis should be in %.

I don't know if your measurements reaching down to 16 Hz will be valid, I suspect not. You would only really get close to that outdoors with no reflection points in the vicinity and wit the mic further away. Even Danny Richie only measures down to 100 wuth a more Professional measuring system. To get lower Fequency measures you need to do near field measurements in front of the woofer (up to about 11mm from the dust cap, plus the entre of the Port, if ported, and then re-scale the measurements and splice to the far field at 1m. see following method:

How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

Quasi-anechoic measurements are a way to capture the 'real' frequency response of a speaker without the influence of a room. They are called 'quasi-anechoic' because they have the potential to get you really close to the results one might achieve with a non-reflective environment like an...

www.audiosciencereview.com

scroll down to part 4) - Near Field Bass measurements. I've done this and it works.

While I wait for a response to the question in my previous post, I thought I'd experiment with the Alignment Tool button on the ALL SPL Controls Tab. My goal was to combine the individual Woofer and Midrange measurements so I can generate filters to export to MiniDSP for the digital crossover.

Attached is a clip of how I set-up the Alignment Tool:

1. Reduce the gain of the Woofers measurement by 5.2 db. to get better spl alignment to Mid.
2. Placed my cursor at 251 Hz, where I plan to set the MiniDSP crossover from Woofer to Mid.
3. Select "Level Phase at cursor" button to level the 1st trace (woofers) Phase.
4. Select "Align Phase at cursor" button to align the 2nd trace (midrange) Phase to the woofers. (This generated a .5 ms delay for the midrange - I'm assuming based on the acoustical timing references shared by both measurements)
5. Select "Aligned Sum" to create a new measurement that is a summed copy of the selected drivers's measurement with the alignment gain, delay and polarity selections applied. (Paraphrasing the HELP file)

And Voila! I believe the new "Aligned Sum" measurement will be my basis for going to the next (EQ) stage... and attempting to flatten this curve to more closely correspond with a Target Curve...

Question - am I on the right track so far?
Thanks - Sixto

i dont understand why you're not getting a flat frequency response. Yes the in room response at the main listening position should have smooth linear downward tilt, but on axis at 1 m (Quasi anechoic) you should be getting a flat response.

 

[Sixto]

I don't know if your measurements reaching down to 16 Hz will be valid, I suspect not…

I agree, I’m just looking to get wide baseline measurements (to whatever accuracy I get) that I can trim-off later with a high-pass filter.

scroll down to part 4) - Near Field Bass measurements. I've done this and it works.

Than you! Sam had also pointed this thread out in post #8. Your link helped me find it. I’m in the process of going through this thread now, particularly interested in how to best “splice” two measurements, and also to go back and see how my woofers look when I do near field measurements, instead of 3’ away for each of the drivers.

i dont understand why you're not getting a flat frequency response. Yes the in room response at the main listening position should have smooth linear downward tilt, but on axis at 1 m (Quasi anechoic) you should be getting a flat response.

I think there may be a number of possibilities to track this down:

1. My drivers may be old and tired, (speakers were made in 1991) specifically, the mid and tweeter seem to be falling off quite rapidly. (as Sam pointed out) In this case, I may just need to get replacement drivers, but before I do that I want to go through the whole process as proof of concept.
2. There may be something wrong with how my microphone handles higher frequencies. (I will test another speaker using the same setup and see if this is the case)
3. Maybe there is some incorrect setting deep in REW or Windows that I’m overlooking. (I’ll continue researching this)

Thanks, Sixto

 

[Sixto]

I don't know where the problem is, but there's clearly something strange going-on with my REW Measurement setup...
And it's not my Tweeters! (Yay?)
Similar drooping high-frequency responses from 2 other totally different speakers means it's something with my equipment, or measurement process that's messed up, not what I'm measuring!

I'm attaching a JPG screen capture of measurements taken from 3 different speakers, the B&W 3-ways, a Definitive Technology 1000 2-way and a Sony Fullrange single driver speaker from a Home Theater system.

• They all drop about 20 db between 1K and 10KHz.
• They all peak above 105 db between 100Hz and 400Hz.

It's just too weird... So, I tried removing my calibration files for both the microphone and the USB interface - No significant change.

In the next few days, I'll try borrowing a different Amp, Mic, and USB interface and see if any of it makes a difference.

If anybody has an idea of what may be wrong with my setup, PLEASE HELP!!!

Thank you! - Sixto

 

[Sixto]

Its the USB Interface that's the cause of the odd SPL measurements.

Tried a different amp - no difference.
Tried a different interface, an older Tascam US-322 and even though the interface calibration file is far from flat, I was able to get more consistent SPL levels across the range of frequencies I’m using to test each driver.

Here are the 3 speakers I posted yesterday plus a 4th (Pioneer 5x9" 5-way car audio in a sealed box) for good measure.... the SPL graphs are much flatter. I will go back and do a nearfield measurement for the Woofers later, The attached measurement is my new baseline for the DM640 with the stock Crossover.

I'm just glad I was able to trace the issue. For some reason, the Phantom Power button is not working on my Focusrite 2I2, even though it stays lit red. The input level LED's on the mic dials would flash green/red when I plug in the Mic or push the Phantom Power button, but I could not get them to stay green or red when checking levels, no matter how many dials or buttons I pushed. The 3rd file attached below shows a comparison of the calibration files for the 2 interface cards... The Tascam has a huge LF dropoff, but the Scarlett really messed me up.

Time to get a NEW USB interface and do all the measurements over again.

I can't thank Steve138 and Sam_Adams ENOUGH for pointing out there was something wrong with my measurements and sticking with me even though I was in denial. Eventually their comments got through my thick skull.
Sixto

 

[Sixto]

While I wait for an answer from Focusrite (repair the broken unit or buy new) I figured I'd go ahead and re-measure the individual driver responses without any XO using the Tascam USB interface. I also did the woofer as a single driver, at close range, though I have not yet corrected it for baffle-step. Much better measurements, and I'm still thinking I'll cross over between 200-250Hz for the low range and 2000 to 2500 for the Mid-tweeter XO.

I will also measure just the HF input using the standard XO, and if that looks good enough, I may just keep it in place for the Mid/Tweet Passive XO only, and use the MiniDSP for the LF XO and PEQ for a flatter response. It looks pretty flat to me already! It looks to me that all 3 drivers are within +or- 3db between 20hz and 20,000hz.

 

[steve138]

I use a UMIK-1 plugged direct into my laptop. look on HIFiShark.com if you can find a used UMIK-1 at a good price. It's literally plug and play, or plug and record

 

[steve138]

I looked at your last mdat file and applied the time gating. your first reflections are very close at only 1.33ms which is less that two feet away. You need to get the speaker and mic as far away from large objects and room boundaries (walls, ceiling, floor) as possible. I have my bookshelf test speaker in a room with 2.55m concrete floor and ceiling, on a stand so the tweeter is half way between floor and ceiling with about 5 feet to the wall behind the speaker which has wall and floor cupboards in my workshop and about 3m to each side wall and 5m in front. the mic is on a stand with tip at 1m from tweeter, on tweeter axis, with 2.83V at speaker terminals, measuring around 85dB on Pink Noise.

this image above is your impulse response with right marker at the onset of first reflections. which gives a very rough response: for a multiway speaker you can try measuring on tweeter axis but further away like 54"

Whereas this is the Impulse I get on one of my Linn Tukans with the new crossover I developed and built:

and the associated response:

 

[Sixto]

I use a UMIK-1 plugged direct into my laptop. look on HIFiShark.com if you can find a used UMIK-1 at a good price. It's literally plug and play, or plug and record

Thanks! I will keep my eye out for a deal

I looked at your last mdat file and applied the time gating. your first reflections are very close at only 1.33ms which is less that two feet away. You need to get the speaker and mic as far away from large objects and room boundaries (walls, ceiling, floor) as possible. I have my bookshelf test speaker in a room with 2.55m concrete floor and ceiling, on a stand so the tweeter is half way between floor and ceiling with about 5 feet to the wall behind the speaker which has wall and floor cupboards in my workshop and about 3m to each side wall and 5m in front. the mic is on a stand with tip at 1m from tweeter, on tweeter axis, with 2.83V at speaker terminals, measuring around 85dB on Pink Noise.

Steve, thanks for taking the time to provide such a thoughtful response. While I wait to connect with Focusrite, I’ve been reading tons of threads on this forum, and running various measurement scenarios to see what works better using the Tascam USB interface. Learning a lot along the way.

So far all measurements have been done in my living room with the speaker and furniture in their actual position, on the assumption that I will get closer to the real performance of the whole system. Is it better to measure and equalize based on quasi-anechoic conditions? And if so, why? The reason I ask is that I could move the REW setup to a larger, empty room for measurement purposes… but I worry that if I do that, when I move the speakers back to the living room, all the corrections will be “off” somehow.

Thanks, Sixto

 

[skid00]

" Is it better to measure and equalize based on quasi-anechoic conditions?"

It is crucial. Your ear/brain notices the freq response from the direct radiation of the speaker, disregarding (for the most part) the reflections of all objects around the speaker. Your measurement mic doesn't. Gating isn't a perfect solution for this. Most experienced audio folks don't EQ above bass frequencies *at all*.

 

[Sixto]

Took a while, after several calls to product support (these guys were great) but we finally came to the conclusion that my USB interface has a hardware problem we can't fix with Phone Support, so I am going through their trade-in process and hope to get a brand-new Scarlett V4 soon so I can resume the process. (My old V2 unit was at least 5 years old, so I hope the new 3-year warranty will do the trick if anything else crops up.) Sixto