Soundeasy, yes. Mini dsp or sigma studio, no. @dcibel helped me set it up. I use Soundeasy to capture the impulse files and then convert them to FRD's using VituixCAD. VituixCAD has an option to export any driver icon output as a wav file. So, I can effectively export my passive crossover for each driver icon by saving each passive crossover section as an individual wav file (cool function).
Then I load these individual driver icon wav files into the Equalizer APO sofware package, which, in turn, convolutes theses individual wav files and sends them to my usb connected 7.1 channel external sound card. I'm using a Soundblaster X3 that I picked up at Besy Buy. The Soundblaster X3 then creates analog outputs for each passive crossover driver icon and sends the signal to my power amplifiers, which, in turn, send the crossover output to the indvidual drivers. Then I can listen to my passive crossover without buying parts. It is a really cool setup and I had it working for a short while. Then I started getting problems with the Equalzier APO software freezing up (or locking up). Had to use windows taskmgr to shut it down and try again. I was using Equalizer APO version 1.3 and I just noticed that they are now up to Equalizer APO version 1.4. Maybe the new version will be more stable on my computer, but I have not tried it yet.
Finished the bases, except for bondo, sanding, and painting. Progress has been slow, as I have been spending most of my time working on the crossover. Moving drivers around, re-measuring, tweaking xover, then re-measuring again, etc. I'm using this project as a personal educational tool to improve my understanding of how driving spacing and crossover frequency affect both the vertical and horizontal polar response patterns.
Probably not. Based on @rjj45 's comment in the other thread, a CBT's ". . . front curve is always a portion of a perfect circle." I traced my curves from a large, elliptically shaped serving plate. So it is more of an egg shape, so to speak.
@4thtry said:
Probably not. Based on @rjj45 's comment in the other thread, a CBT's ". . . front curve is always a portion of a perfect circle." I traced my curves from a large, elliptically shaped serving plate. So it is more of an egg shape, so to speak.
You still have nice curves for your stands, even if they are not arcs of a circle.
But Chahly - Stahkist don't want speakers that look good, Stahkist wants speakers that sound good!
Progress report. Bases attached. Lightly stuffed each cab with denim. Wiring is 16ga OFC from PE. I still need to mount the drivers in one channel. Then it is on to convoluting the xover sections into wave files and hooking them up for stereo 2 channel playback using EQ APO. If I can get EQ APO working, my plan is to tweak the xover in Vcad, re-convolute, then listen. Then tweak some more, re-convolute, and listen again. I still have plenty of time before Indy to order parts and finish the project passively.
The "head in a vise" situation should not be too bad based on my modeling so far. Maybe the imaging will be a little bit more precise when all the mid and tweeter driver acoustic centers are properly aligned. Maybe not. But the tonal balance should be roughly the same when changing from a seated to standing position. The alternative would be an MMTMM "non-aligned" type situation with a slightly less focused image in all listening positions. I'll be testing this to see what happens.
Got the software bugs fixed and have been listening to the pair in stereo, with subs, using Equalizer APO connected to my Soundblaster X3 7.1 channel sound card and three stereo power amplifiers. Listening to crossover variation #8. Sounding much better than my first attempt using the Bozhen CQ76B tweeters. When I tried to use them previously, as a two way with the Esoteric 7" woofers, they sounded harsh and bright. Not this time. Very smooth and liquid sounding now. The harsh sound was probably due to power response bulging in the 3-5kHz region caused by a woofer/tweeter directivity mismatch. The much smaller 3" BMR mids have a very wide and smooth horizontal polar response characteristic, very similar to the Bozhen tweeter.
Also, there appears to be no problem with "head in the vise" type imaging. The tonal balance does not change from a seated to standing position either. Vertical lobe appears to be fairly wide and imaging locks in well at a listening distance of about 9 feet. I did a series of measurements and verified that the acoustic centers of all five drivers ( tweeter and the four 3" BMR mids) converge at a listening distance of 9 feet directly on-axis with the tweeter.
Today, I got the "bright idea" to attempt an on-design axis validation measurement with OmniMic and discovered that the model was not matching my actual measurements in the 500 to 800Hz region. For whatever reason the tweeter's output was conflicting with the four midrange driver's outputs and creating a huge 5 to 10dB suck out in the 500 to 800Hz region that did not show up in the model. I was playing OmniMic's test track #2 whoop whoop file using the laptops CD player and windows media player. The whoop whoop test track was then convoluted and split out to my three amplifiers using Equalizer APO. At first I thought that I must have the drivers hooked up out of phase, or something stupid like that. But no, everything was correct. The suck out was consistently showing up in the 500 to 800Hz region regardless of microphone distance (1/2 to 2.5 meters), speaker height (40 to 55" above floor). But then the huge suck out would completely disappear if I measured the drivers alone and then summed the individual FRD files in OmniMic.
My theory is that I am attempting to do something that simply cannot be done. The Equalizer APO software package is probably introducing a processing time delay that is somehow messing up OmniMic's ability to properly synchronize the whoop whoop test tracks with the software package. So I guess I am going to have to perform JR's LMT (listen, measure, tweak) process without a validation measurement. For now, I will simply have to assume that what I am currently hearing should be somewhat similar to what the actual passive xover will sound like later on.
@Steve_Lee said:
Can you run OmniMic on a separate PC from the one playing/processing the digital audio, 4thtry?
(this is what I do to reduce latency/processor overhead).
Thanks for the suggestion. I have another PC that I could use for this, but it will take me some time to move the PC over to my test area and set things up. I also have a stand alone CD player that I could feed into the analog audio inputs of the laptop that is running the Equalizer APO software package.
Comments
mini dsp, sigma studio, soundeasy?
Soundeasy, yes. Mini dsp or sigma studio, no. @dcibel helped me set it up. I use Soundeasy to capture the impulse files and then convert them to FRD's using VituixCAD. VituixCAD has an option to export any driver icon output as a wav file. So, I can effectively export my passive crossover for each driver icon by saving each passive crossover section as an individual wav file (cool function).
Then I load these individual driver icon wav files into the Equalizer APO sofware package, which, in turn, convolutes theses individual wav files and sends them to my usb connected 7.1 channel external sound card. I'm using a Soundblaster X3 that I picked up at Besy Buy. The Soundblaster X3 then creates analog outputs for each passive crossover driver icon and sends the signal to my power amplifiers, which, in turn, send the crossover output to the indvidual drivers. Then I can listen to my passive crossover without buying parts. It is a really cool setup and I had it working for a short while. Then I started getting problems with the Equalzier APO software freezing up (or locking up). Had to use windows taskmgr to shut it down and try again. I was using Equalizer APO version 1.3 and I just noticed that they are now up to Equalizer APO version 1.4. Maybe the new version will be more stable on my computer, but I have not tried it yet.
LSPcad has a similar function.
Decided to build a pair of "curvy" bases instead of a simple straight post.
Finished the bases, except for bondo, sanding, and painting. Progress has been slow, as I have been spending most of my time working on the crossover. Moving drivers around, re-measuring, tweaking xover, then re-measuring again, etc. I'm using this project as a personal educational tool to improve my understanding of how driving spacing and crossover frequency affect both the vertical and horizontal polar response patterns.
Wonder if these bases would compliment Ugly Woofer's new build . . .
Probably not. Based on @rjj45 's comment in the other thread, a CBT's ". . . front curve is always a portion of a perfect circle." I traced my curves from a large, elliptically shaped serving plate. So it is more of an egg shape, so to speak.
You still have nice curves for your stands, even if they are not arcs of a circle.
Curves ya
Progress report. Bases attached. Lightly stuffed each cab with denim. Wiring is 16ga OFC from PE. I still need to mount the drivers in one channel. Then it is on to convoluting the xover sections into wave files and hooking them up for stereo 2 channel playback using EQ APO. If I can get EQ APO working, my plan is to tweak the xover in Vcad, re-convolute, then listen. Then tweak some more, re-convolute, and listen again. I still have plenty of time before Indy to order parts and finish the project passively.
I didn't understand how the bases were going to work from the prior pics, but that is quite cool!
As cool looking as it is, seems like that would really emphasize the "head in a vise" sweet spot.
He must have a very large desk . . . (I like them too . . .).
What I was thinking.
Maybe not with those Tectonic drivers. After the comments tho, I'm imagining these as giant headphones, lol
The "head in a vise" situation should not be too bad based on my modeling so far. Maybe the imaging will be a little bit more precise when all the mid and tweeter driver acoustic centers are properly aligned. Maybe not. But the tonal balance should be roughly the same when changing from a seated to standing position. The alternative would be an MMTMM "non-aligned" type situation with a slightly less focused image in all listening positions. I'll be testing this to see what happens.
These are floor standers. They will sit on top a pair of subwoofers.
Ahhh . . . makes perfect sense, now.
Looking forward to your XO development and success stories . . .
Me too
Got the software bugs fixed and have been listening to the pair in stereo, with subs, using Equalizer APO connected to my Soundblaster X3 7.1 channel sound card and three stereo power amplifiers. Listening to crossover variation #8. Sounding much better than my first attempt using the Bozhen CQ76B tweeters. When I tried to use them previously, as a two way with the Esoteric 7" woofers, they sounded harsh and bright. Not this time. Very smooth and liquid sounding now. The harsh sound was probably due to power response bulging in the 3-5kHz region caused by a woofer/tweeter directivity mismatch. The much smaller 3" BMR mids have a very wide and smooth horizontal polar response characteristic, very similar to the Bozhen tweeter.
Also, there appears to be no problem with "head in the vise" type imaging. The tonal balance does not change from a seated to standing position either. Vertical lobe appears to be fairly wide and imaging locks in well at a listening distance of about 9 feet. I did a series of measurements and verified that the acoustic centers of all five drivers ( tweeter and the four 3" BMR mids) converge at a listening distance of 9 feet directly on-axis with the tweeter.
Today, I got the "bright idea" to attempt an on-design axis validation measurement with OmniMic and discovered that the model was not matching my actual measurements in the 500 to 800Hz region. For whatever reason the tweeter's output was conflicting with the four midrange driver's outputs and creating a huge 5 to 10dB suck out in the 500 to 800Hz region that did not show up in the model. I was playing OmniMic's test track #2 whoop whoop file using the laptops CD player and windows media player. The whoop whoop test track was then convoluted and split out to my three amplifiers using Equalizer APO. At first I thought that I must have the drivers hooked up out of phase, or something stupid like that. But no, everything was correct. The suck out was consistently showing up in the 500 to 800Hz region regardless of microphone distance (1/2 to 2.5 meters), speaker height (40 to 55" above floor). But then the huge suck out would completely disappear if I measured the drivers alone and then summed the individual FRD files in OmniMic.
My theory is that I am attempting to do something that simply cannot be done. The Equalizer APO software package is probably introducing a processing time delay that is somehow messing up OmniMic's ability to properly synchronize the whoop whoop test tracks with the software package. So I guess I am going to have to perform JR's LMT (listen, measure, tweak) process without a validation measurement. For now, I will simply have to assume that what I am currently hearing should be somewhat similar to what the actual passive xover will sound like later on.
Can you run OmniMic on a separate PC from the one playing/processing the digital audio, 4thtry?
(this is what I do to reduce latency/processor overhead).
Thanks for the suggestion. I have another PC that I could use for this, but it will take me some time to move the PC over to my test area and set things up. I also have a stand alone CD player that I could feed into the analog audio inputs of the laptop that is running the Equalizer APO software package.
Crossover parts placement in progress. I have room for a 3.5x11" masonite board behind the tweeter, so it will be a tight fit.