Member
Joined 2009
Paid Member
Well Stig owns a speaker driver dealership so he gets to play with all the toys he wants !
I was going to build his InConcert but for me it would be $4,000 a pair! so I decided to delay that idea. I find it satisfying looking for how to build things without spending a fortune. Loyaties? for me I'd find it boring to build the same thing over and over. I started with building an Onken, I've built a slot ported bass reflex, a push-push sealed sub-woofer and I've got plans for an ML-TL around some EL-70's cooking up plus I do really like the look of some of the BLH's I've seen and plan to build one of those sometime. But it's always possible to take it further, with expensive drivers digital XO's fancy cables and big boxes and well - if you have the gear it can be fun all by itself.
Last edited:
toys
In the mid 80s I had a little car audio shop and in between the deck+2 or deck+eq+4 installs I got to play with cool drivers from Just Speakers. Not a dealer but I got to play with excellent stuff. Especially Scanspeak. Of course a car cabin is a colored box itself, then there's road and drivetrain noise, imaging issues, etc...but it was fun to look at the driver updates and say "lets try these."
In the mid 80s I had a little car audio shop and in between the deck+2 or deck+eq+4 installs I got to play with cool drivers from Just Speakers. Not a dealer but I got to play with excellent stuff. Especially Scanspeak. Of course a car cabin is a colored box itself, then there's road and drivetrain noise, imaging issues, etc...but it was fun to look at the driver updates and say "lets try these."
Member
Joined 2009
Paid Member
sounds like a blast !
well, my employer was in the speaker business back then, I even owned a pair of their speakers (although I was a student at the time with no notion that I'd be working for this company today). I no longer have the speakers, the surrounds melted with age and I donated them to charity.
well, my employer was in the speaker business back then, I even owned a pair of their speakers (although I was a student at the time with no notion that I'd be working for this company today). I no longer have the speakers, the surrounds melted with age and I donated them to charity.
Member
Joined 2009
Paid Member
??? Yes, it can work!
Hi Lolo, perhaps I wasn't clear enough in what I was proposing and have actually implemented on my own OB systems.
Precisely because to try and get the high amount of LF boost in the DCX will create digital clipping, we add maybe +10 or +20 dB of analog gain between the dcx (or digital eq) and the woofer amplifier section. No extra analog gain on the mids or highs as only the woofers need it.
Yes, at 40-60Hz the woofers will see a lot of gain... example: with +20dB analog gain and +1dB eq boost, the sum is now +21dB.
At 80 to 120Hz, you could shape the equalization to create less boost. e.g. if at 120Hz you have +20dB analog gain, and you reduce the level on the eq to -15dB, the overall boost that the woofer sees now sums to only +5dB at 120Hz
At 250Hz, you might end up creating a -25dB cut on the eq, so that after the +20dB section, you end up with a net reduction of -5dB
NOTE: The actual filter slopes do not change!! we are only shifting a lot of the boost to an analog section that follows the digital eq, and this allows the digital eq to operate within its proper range and stay out of clipping, or conversely, to be too low in level to be useful.
Here's why: The reason for this is that in some cases you can get extraordinarily musical tonality and dynamic impact!
On my entry level system, I use a single 15" woofer with the eq similar to what I previously described, except I'm only adding +6dB analog boost via a 1:2 transformer prior to the amp. I had initially feared it would not perform properly, but in practice it works extremely well, suitable for a wide range of music. It can play very loud, and with excellent tone and clarity.
That being said, there is definitely an advantage in using two, three or more 15" woofers, we get a lot more headroom, and with much less need for heavy eq in the bass. I am developing some larger systems that offer that option. They really perform at amazingly good levels, and can really produce prodigious SPL with incredible dynamics.
Also, like many who have worked with OB, I don't see that it is practical to try and cover the low bass below 40Hz with OB. For that I use servo subs in multiples and in arrays to reduce the room modes. That being said, with my OB's tuned to give -3dB at 38Hz in the room, on 95% of the music I listen to, I don't need the subs very much at all. Exception: Beethoven's 9th, Dvorak and several other full scale symphonic work. For this type of music, I like it better with the servo-subs for sure!
As a general aside, I find it very interesting to see how many people in this forum are not willing to experiment before drawing conclusions. Like many here, I initially believed OB with heavy digital eq would not work well at all. It was after I tried it and refined the eq process that I finally became convinced it can be an excellent solution.
Theory is great, when it is correct. Simulations are great when they are correct. Even opinions can be great when they are correct.
Experimentation is still de rigueur to prove or disprove a postulate.
And, since I actually have conducted the experiments, very successfully, perhaps I can help others who would like to conduct them as well. Or at least help them develop a better idea of the trade-offs.
Hi Lolo, perhaps I wasn't clear enough in what I was proposing and have actually implemented on my own OB systems.
Precisely because to try and get the high amount of LF boost in the DCX will create digital clipping, we add maybe +10 or +20 dB of analog gain between the dcx (or digital eq) and the woofer amplifier section. No extra analog gain on the mids or highs as only the woofers need it.
Yes, at 40-60Hz the woofers will see a lot of gain... example: with +20dB analog gain and +1dB eq boost, the sum is now +21dB.
At 80 to 120Hz, you could shape the equalization to create less boost. e.g. if at 120Hz you have +20dB analog gain, and you reduce the level on the eq to -15dB, the overall boost that the woofer sees now sums to only +5dB at 120Hz
At 250Hz, you might end up creating a -25dB cut on the eq, so that after the +20dB section, you end up with a net reduction of -5dB
NOTE: The actual filter slopes do not change!! we are only shifting a lot of the boost to an analog section that follows the digital eq, and this allows the digital eq to operate within its proper range and stay out of clipping, or conversely, to be too low in level to be useful.
Here's why: The reason for this is that in some cases you can get extraordinarily musical tonality and dynamic impact!
On my entry level system, I use a single 15" woofer with the eq similar to what I previously described, except I'm only adding +6dB analog boost via a 1:2 transformer prior to the amp. I had initially feared it would not perform properly, but in practice it works extremely well, suitable for a wide range of music. It can play very loud, and with excellent tone and clarity.
That being said, there is definitely an advantage in using two, three or more 15" woofers, we get a lot more headroom, and with much less need for heavy eq in the bass. I am developing some larger systems that offer that option. They really perform at amazingly good levels, and can really produce prodigious SPL with incredible dynamics.
Also, like many who have worked with OB, I don't see that it is practical to try and cover the low bass below 40Hz with OB. For that I use servo subs in multiples and in arrays to reduce the room modes. That being said, with my OB's tuned to give -3dB at 38Hz in the room, on 95% of the music I listen to, I don't need the subs very much at all. Exception: Beethoven's 9th, Dvorak and several other full scale symphonic work. For this type of music, I like it better with the servo-subs for sure!
As a general aside, I find it very interesting to see how many people in this forum are not willing to experiment before drawing conclusions. Like many here, I initially believed OB with heavy digital eq would not work well at all. It was after I tried it and refined the eq process that I finally became convinced it can be an excellent solution.
Theory is great, when it is correct. Simulations are great when they are correct. Even opinions can be great when they are correct.
Experimentation is still de rigueur to prove or disprove a postulate.
And, since I actually have conducted the experiments, very successfully, perhaps I can help others who would like to conduct them as well. Or at least help them develop a better idea of the trade-offs.
Last edited:
You simply have to reduce the DCX's input to counterbalance any boost you apply later on. Then there will be no clipping.
It may not be as bad as you think! Parts express sells a dandy 15" woofer, the GRS 15PF-8, usually on sale for under $30USD. Put 2 of 'em on a simple OB and you've got the makings of a pretty good bass section. Use 4 if you really want to get loud and clean.
Part #: * 292-415
As for loyalties, I had them too. Now I just want to make what sounds best, and gives the biggest bang for the buck.
Last edited:
Very true.
However, as an example, if we need a 36dB swing between max boost and max cut to get the bass flat (before implementing a low-pass filter), and assuming we don't want the bass going above +0dB, then we will lose as much as 6 bits of resolution, right in the mid-bass.
+20dB analog eq boost after the dcx allows us to reduce the amount of resolution that is lost by about three bits in this example.
How audible is the difference? Nothing dramatic, but I am pretty sure I hear a gain in transparency and inner detail.
Not sure what frequencies you're talking about but going from 24 bits to 18 bits at low frequencies is simply insignificant. And, how do you make up for lost bits???
By the way, if you need to EQ 36dB then the dynamic capabilities of your speakers are probably not very good.
Dipoles at very low frequencies are inefficient (which results in reduced dynamic capabilities or high amount of distortion or both) and they can't pressurize the room below the lowest mode.
A better concept probably would be multiple subs below 80Hz and mains with a dipole midrange. This might help with modal interaction at frequencies around the Schröder frequency if there's really a significant benefit of using dipoles instead of monopoles. And this is a big if. I've yet to see reliable data proving that dipoles around the Schröder frequency produce less modal problems than monopoles.
losing 6 bits of 24 is insignificant? I would strongly disagree with that Markus. If it were insignificant then 16bit audio would still be the standard. Im not going to work out bits per dB, but i could. I know i can hear the change from 16 bit to 24.
Well... 16 bits is the standard.... all CD's and virtually all streaming formats and downloadable music files are 16 bits.....
Regardless of whether 16 bits is enough for full bandwidth audio, I would point out that fletcher Munson curves mean that our hearing has far less dynamic range at bass frequencies in the first place. We don't actually need 96dB of dynamic range down at bass frequencies because the minimum discernible SPL is far far higher than it is at higher frequencies, thus less bits (relative to full scale) are required to represent that.
In a properly dithered digital system the symptoms of not enough bits is simply a higher noise floor (anyone who still doesn't believe that really needs to read and understand digital sampling theory) and with a flat noise floor it is always noise at the high frequency end that you will hear first.
If the digital EQ is only driving the woofer and handling the sloping bass section of the spectrum as long as level matching is correct and clipping is avoided (both digital and analogue) I really can't see any issues with "lack of bits".
On the other hand if the same digital EQ output is also providing the EQ for the rest of the spectrum (with a lot of digital attenuation) then there MIGHT be a SNR issue, at least in theory. Whether it was audible is another matter, and probably depends more on the analogue stage design of the digital EQ.
In a properly dithered digital system the symptoms of not enough bits is simply a higher noise floor (anyone who still doesn't believe that really needs to read and understand digital sampling theory) and with a flat noise floor it is always noise at the high frequency end that you will hear first.
If the digital EQ is only driving the woofer and handling the sloping bass section of the spectrum as long as level matching is correct and clipping is avoided (both digital and analogue) I really can't see any issues with "lack of bits".
On the other hand if the same digital EQ output is also providing the EQ for the rest of the spectrum (with a lot of digital attenuation) then there MIGHT be a SNR issue, at least in theory. Whether it was audible is another matter, and probably depends more on the analogue stage design of the digital EQ.
Last edited:
Jack, thanks for explaining, I did not exactly understand the first time, it makes sense now.
well, I don't think the bits loss is a problem, but this deserves some thoughts.. SPL wise, sure, dipoles will never compete in the low bass range, but as you ask, what about modes and distorsion? dipoles placed at pressure points vs monopole at velocity points, who wins?
well, I don't think the bits loss is a problem, but this deserves some thoughts.. SPL wise, sure, dipoles will never compete in the low bass range, but as you ask, what about modes and distorsion? dipoles placed at pressure points vs monopole at velocity points, who wins?
Problem is that velocity points vary with frequency, they move around the room so to speak. Pressure points are always at the room boundaries.
Who places dipoles at pressure points ? The only pressure points in a room at all frequencies are room boundaries.
Who places monopoles at velocity points ? There is nowhere in a room that is a velocity point at all frequencies.
The reality is all speakers (except maybe subwoofers with restricted bandwidth) are placed in locations that are pressure points at some frequencies and velocity points at others...
As for distortion, a dipole is always at a massive disadvantage, because it takes so much more volume displacement (thus more excursion) for a given SPL.
Room modes ? Best dealt with by distributed sources of bass...
I have placed my dipole H-baffles very close to the walls, so I am in the pressure zone for frequencies below approx 100 Hz, and below there I have no problems with standing wave modes, which I certainly did with monopole woofers. But above 100 Hz, modes starts to show up on frequency response measurements.
Sure! It's always a compromise, but you can deal with the worst offenders at least, for one given listening point..
it's not easy to blend a sub with a dipole, maybe not impossible, but..
Not sure I understand this right but are you saying that the dipole helped below 100Hz but not above?