To anyone contemplating bypassing their speaker's passive crossover and upgrading to a bi-amp/DSP setup, I say just do it (you will love it).
My ESLs are homebuilt hybrids with segmented wire stators and open baffle woofers, but the setup and tuning steps will be similar for any hybrid ESL, and a DSP crossover provides almost infinite tuning capability, all in real time, from your listening position, and its not hard to do.
I use a DBX Driverack Venu 360 DSP/Crossover and I love it (PA2 version is essentially the same at 1/2 the price but it doesn't have a digital input).
Crossover frequency and slope:
Given ML's higher diaphragm tension and its use of the diaphragm's drum-head resonance to enhance the mid-bass output, I'm not sure what the ideal crossover point and slope would be.
So for a ML, I would initially set the crossover to the same frequency set by the original passive crossover, using a symmetric 24db Linkwitz-Riley filter, and go from there. In any case, I would not go lower than 250 Hz. A 48db slope would be cleaner and allow driving the panel harder, but it's a bit more difficult to dial in the tuning.
The DBX PA2 or Venu 360 setup Wizards include a menu of built-in level presets for many popular amp and speaker brands. When prompted by the Wizard to select a brand, I ignore the brand presets and select the "custom" option. The "custom" option requires me to set the levels manually, and I will explain how to do that below.
Time alignments:
Time alignment/phasing is less critical with subs because bass wavelengths are so long that you can have quite a bit of distance difference before it would be audible. I use a pair of subs, one beside each main speaker, so I don't need to time align my subs.
If your room layout allows it, and you are using a single sub, I think it's best to place it between the mains, at the same distance from your listening position.
The closer you can get to same-distance, the more accurate the time-alignments (phasing) will be. If your room layout doesn't permit equi-distant placement of the sub(s) you can use the DBX's digital delay to time-align the subs to the mains and/or the mid-bass woofer to the panels.
The DBX has digital delay but it must be manually calculated and applied, as the DBX can't do it automatically (my only beef with the DBX).
I time-align my speakers' mid-bass woofers to the panels because the woofers' voice coils are physically located about 4" behind the ESL diaphragms (equates to about 0.3 milliseconds).
There are two methods to time-align the woofers and panels:
1. Measure the difference between the diaphragm and woofer voice coils and calculate the delay time based on the speed of sound (1126 ft/sec).
2. [more accurate] Using the DBX's signal generator and RTA screen on the iPad control app, and with the microphone place about 2 feet away at a hight between the panel and woofer;
a) Feed in a sine wave test tone matching the crossover frequency, first to the
woofer (with the panel channel muted) and if needed, adjust its volume level.
b) Now mute the woofer channel, feed the test tone to the panel and adjust its
volume to the same level.
c) Unmute both the woofer and panel channels and observe their combined
volume level on the RTA screen.
d) Access the digital delay screen and incrementally apply time delay on the
panel to achieve the maximum volume level on the RTA screen (i.e. max
constructive interference). The panel and woofer are now exactly aligned.
Setting the woofer/panel channel gains:
1. Place the mic at your listening position.
2. Go to the RTA screen, signal generator, and feed in the same sine wave test tone matching the crossover frequency.
3. On the Crossover screen, adjust the band gains to bring their respective volumes to about the same SPL on the RTA screen.
4. Having adjusted the woofer & panel band/channel gains at the crossover frequency, now change the test tone to PINK NOISE and observe the frequency response across all octaves on the RTA screen.
On the Crossover Screen;
Re-adjust the band/channel gains to bring High/Low bands (or High/Mid/Low bands if setup as 3-way) to bring all bands to roughly equal levels.
5. Now run the Auto EQ Wizard;
Select the desired curve (flat, high shelf, or low shelf) and follow the prompts for positioning the mic. Run three frequency sweeps with the mic placed as shown on the iPad app's display screen. After the three sweeps, the Auto EQ will automatically overlay parametric EQ's to match the response to your pre-selected curve.
If you are using a DSP which doesn't have Auto-EQ, you will have to set adjust the frequency response manually, using its parametric and/or graphic EQ's, until the response is about flat across all octaves, or to your liking.
Fine Tuning:
Any additional tuning I do by ear, using music I'm familiar with, and tweaking the response using the DBX's 31-band graphic EQ, as follows:
I start with the Erin Bode song "Holiday". Erin's clear, piercing voice will cut diamond if the tuning is even slightly hot, so this is my go-too sone for that purpose.
In my system, Erin's voice is a bit too hot at around 3kHz so I take the edge off by pulling the 3kHz band down about 2-3db. I then know that no other music I'm likely to play will sound harsh.
I follow up with Dave Brubeck's "Take Five", and if Joe Morello's hi-hat still sizzles, I know I haven't taken too much off Erin's voice.
For my system, that pretty much gets me happy.
Hope this helps!
Charlie
My ESLs are homebuilt hybrids with segmented wire stators and open baffle woofers, but the setup and tuning steps will be similar for any hybrid ESL, and a DSP crossover provides almost infinite tuning capability, all in real time, from your listening position, and its not hard to do.
I use a DBX Driverack Venu 360 DSP/Crossover and I love it (PA2 version is essentially the same at 1/2 the price but it doesn't have a digital input).
Crossover frequency and slope:
Given ML's higher diaphragm tension and its use of the diaphragm's drum-head resonance to enhance the mid-bass output, I'm not sure what the ideal crossover point and slope would be.
So for a ML, I would initially set the crossover to the same frequency set by the original passive crossover, using a symmetric 24db Linkwitz-Riley filter, and go from there. In any case, I would not go lower than 250 Hz. A 48db slope would be cleaner and allow driving the panel harder, but it's a bit more difficult to dial in the tuning.
The DBX PA2 or Venu 360 setup Wizards include a menu of built-in level presets for many popular amp and speaker brands. When prompted by the Wizard to select a brand, I ignore the brand presets and select the "custom" option. The "custom" option requires me to set the levels manually, and I will explain how to do that below.
Time alignments:
Time alignment/phasing is less critical with subs because bass wavelengths are so long that you can have quite a bit of distance difference before it would be audible. I use a pair of subs, one beside each main speaker, so I don't need to time align my subs.
If your room layout allows it, and you are using a single sub, I think it's best to place it between the mains, at the same distance from your listening position.
The closer you can get to same-distance, the more accurate the time-alignments (phasing) will be. If your room layout doesn't permit equi-distant placement of the sub(s) you can use the DBX's digital delay to time-align the subs to the mains and/or the mid-bass woofer to the panels.
The DBX has digital delay but it must be manually calculated and applied, as the DBX can't do it automatically (my only beef with the DBX).
I time-align my speakers' mid-bass woofers to the panels because the woofers' voice coils are physically located about 4" behind the ESL diaphragms (equates to about 0.3 milliseconds).
There are two methods to time-align the woofers and panels:
1. Measure the difference between the diaphragm and woofer voice coils and calculate the delay time based on the speed of sound (1126 ft/sec).
2. [more accurate] Using the DBX's signal generator and RTA screen on the iPad control app, and with the microphone place about 2 feet away at a hight between the panel and woofer;
a) Feed in a sine wave test tone matching the crossover frequency, first to the
woofer (with the panel channel muted) and if needed, adjust its volume level.
b) Now mute the woofer channel, feed the test tone to the panel and adjust its
volume to the same level.
c) Unmute both the woofer and panel channels and observe their combined
volume level on the RTA screen.
d) Access the digital delay screen and incrementally apply time delay on the
panel to achieve the maximum volume level on the RTA screen (i.e. max
constructive interference). The panel and woofer are now exactly aligned.
Setting the woofer/panel channel gains:
1. Place the mic at your listening position.
2. Go to the RTA screen, signal generator, and feed in the same sine wave test tone matching the crossover frequency.
3. On the Crossover screen, adjust the band gains to bring their respective volumes to about the same SPL on the RTA screen.
4. Having adjusted the woofer & panel band/channel gains at the crossover frequency, now change the test tone to PINK NOISE and observe the frequency response across all octaves on the RTA screen.
On the Crossover Screen;
Re-adjust the band/channel gains to bring High/Low bands (or High/Mid/Low bands if setup as 3-way) to bring all bands to roughly equal levels.
5. Now run the Auto EQ Wizard;
Select the desired curve (flat, high shelf, or low shelf) and follow the prompts for positioning the mic. Run three frequency sweeps with the mic placed as shown on the iPad app's display screen. After the three sweeps, the Auto EQ will automatically overlay parametric EQ's to match the response to your pre-selected curve.
If you are using a DSP which doesn't have Auto-EQ, you will have to set adjust the frequency response manually, using its parametric and/or graphic EQ's, until the response is about flat across all octaves, or to your liking.
Fine Tuning:
Any additional tuning I do by ear, using music I'm familiar with, and tweaking the response using the DBX's 31-band graphic EQ, as follows:
I start with the Erin Bode song "Holiday". Erin's clear, piercing voice will cut diamond if the tuning is even slightly hot, so this is my go-too sone for that purpose.
In my system, Erin's voice is a bit too hot at around 3kHz so I take the edge off by pulling the 3kHz band down about 2-3db. I then know that no other music I'm likely to play will sound harsh.
I follow up with Dave Brubeck's "Take Five", and if Joe Morello's hi-hat still sizzles, I know I haven't taken too much off Erin's voice.
For my system, that pretty much gets me happy.
Hope this helps!
Charlie
Last edited:
