@Subsonic1050 The 15CXN88 bass mid should be wired in reverse polarity compared to the subs and the tweeters. None of the polarity questions should affect your top octave. I don't know why your driver drops 20dB over that range.
You should check whether you get the same response on both left and right sides.
It's easy enough to fix this with DSP EQ (at least for now). If I were you, I would play with the DSP EQ to try to get as close as you can to this:
I would try setting a 6dB high pass filter at 15KHz on the tweeters, which should mostly level the response. Add shelf filters to do any leveling that remains, and like I did, add a -10dB Low shelf filter that reduces output below 1kHz.
You should pay attention to the time delay settings I used between the drivers in the XML file, and try something similar. I arrived at those settings by carefully watching impulse and step response until the tweeter and bass mid aligned.
It should be fairly easy to flatten the woofer's response by adding a handful of shelf, notch and peak filters without having to resort to any standard Low Pass filters. Don't worry about anything below 50Hz. Tweak the levels of LF and HF so they match.
Then let us know how that looks.
You should check whether you get the same response on both left and right sides.
It's easy enough to fix this with DSP EQ (at least for now). If I were you, I would play with the DSP EQ to try to get as close as you can to this:
I would try setting a 6dB high pass filter at 15KHz on the tweeters, which should mostly level the response. Add shelf filters to do any leveling that remains, and like I did, add a -10dB Low shelf filter that reduces output below 1kHz.
You should pay attention to the time delay settings I used between the drivers in the XML file, and try something similar. I arrived at those settings by carefully watching impulse and step response until the tweeter and bass mid aligned.
It should be fairly easy to flatten the woofer's response by adding a handful of shelf, notch and peak filters without having to resort to any standard Low Pass filters. Don't worry about anything below 50Hz. Tweak the levels of LF and HF so they match.
Then let us know how that looks.
Last edited:
Good news! So I went to go do some sweeps separating for the Left High, Left Low, Right High, and Right Low. Everything was wired correctly. I even tried reversing the polarity of the midrange driver and also the front facing tweeter to see if that made anything better - it didn't. However, I noticed that on the high frequencies, the L and R channels were reversed! I checked my cabling and that was all correct. Turns out that I had reversed them in the MiniDSP software. I really hate how they have that set up, it's incredibly easy to do that. Anyway, that certainly explains why things sounded so "wrong" and why the imaging was so poor. It sounded so bizarre that I was having a hard time figuring out what was happening. This is not the explanation for the use of Perry's filters - that WAS extremely off - no question about it. The only thing I can think of is that the 14cxn88 must really be that different from the 15cxn88 that Perry's filters aren't even a starting point for me in this configuration.
I set everything up and ran Dirac again - much better this time. The sound was cleaned up and sounded significantly more cohesive (kind of to be expected when the 2 halves of the frequency range are reversed L&R). However it still sounded a little lackluster to me. So I played around with speaker positioning and toe in. I was limited as to exactly where I can place them because my NX-Treme's can't move, but I messed around with a wider speaker placement and a pretty significant amount of toe-in. That did it! They, for the first time, sound very good.
There's still a lot of room for improvement, I think. First, the drivers only have about 10 hours on them, and I won't even consider giving any real listening impressions until they have over 100 hours on them. Second, I think the DSP can be tweaked more. I have some very basic crossovers set up on both frequency ranges, some time delay settings, and the inversion on the high frequency range - but I'm not using any of the PEQ settings yet. I'm just using Dirac to clean all of that up. Third, I could use better bass integration with my subs. There's not a great way to do that with the SHD - there aren't enough analog outs. Lastly, and this is a big one - I feel like I can definitely hear the limitations of the MiniDSP DAC. It sounds GOOD, but it's just leaving a bit too much on the table I think. I'm going to try integrating another DAC. With the SHD I think I can only use 1 external DAC, so I'll have to pick which frequency range to cover. I'm hoping that will also free up 2 of the analog outs on the MiniDSP for the subs but I'm not sure how all of that will work. I may need to get a Flex 8 for the additional outputs.
I'm leaving tomorrow morning for 3 weeks. As soon as I get back I'm going to be having a listening weekend with a few buddies. We get together 3 or 4 times a year to just listen to music for a weekend, compare gear, etc. It will be a good time to get some real listening time with these speakers. I'll report back after all of that.
Thank you Perry for all of your help. I'm not at a point where I can give a listening impressions report yet, but I can say that these speakers are at a minimum very good. They may be excellent after all of the above mentioned things get situated. Your willingness to share your designs, help troubleshoot these issues and quick responses are sincerely appreciated.
I set everything up and ran Dirac again - much better this time. The sound was cleaned up and sounded significantly more cohesive (kind of to be expected when the 2 halves of the frequency range are reversed L&R). However it still sounded a little lackluster to me. So I played around with speaker positioning and toe in. I was limited as to exactly where I can place them because my NX-Treme's can't move, but I messed around with a wider speaker placement and a pretty significant amount of toe-in. That did it! They, for the first time, sound very good.
There's still a lot of room for improvement, I think. First, the drivers only have about 10 hours on them, and I won't even consider giving any real listening impressions until they have over 100 hours on them. Second, I think the DSP can be tweaked more. I have some very basic crossovers set up on both frequency ranges, some time delay settings, and the inversion on the high frequency range - but I'm not using any of the PEQ settings yet. I'm just using Dirac to clean all of that up. Third, I could use better bass integration with my subs. There's not a great way to do that with the SHD - there aren't enough analog outs. Lastly, and this is a big one - I feel like I can definitely hear the limitations of the MiniDSP DAC. It sounds GOOD, but it's just leaving a bit too much on the table I think. I'm going to try integrating another DAC. With the SHD I think I can only use 1 external DAC, so I'll have to pick which frequency range to cover. I'm hoping that will also free up 2 of the analog outs on the MiniDSP for the subs but I'm not sure how all of that will work. I may need to get a Flex 8 for the additional outputs.
I'm leaving tomorrow morning for 3 weeks. As soon as I get back I'm going to be having a listening weekend with a few buddies. We get together 3 or 4 times a year to just listen to music for a weekend, compare gear, etc. It will be a good time to get some real listening time with these speakers. I'll report back after all of that.
Thank you Perry for all of your help. I'm not at a point where I can give a listening impressions report yet, but I can say that these speakers are at a minimum very good. They may be excellent after all of the above mentioned things get situated. Your willingness to share your designs, help troubleshoot these issues and quick responses are sincerely appreciated.
Hi Perry,
We exchanged few PMs regarding the Bitches Brew version with FaitalPro 15HX500 coaxials. Still have some private issues that prevent me from starting the project but would like to check which speaker drivers would fit. I was thinking on using FaitalPro 18HP1030 or SB Audience NERO-18SW1900D drivers for woofers. Besides being 18” instead of 15” is there any reason not to use them?
Thanks,
We exchanged few PMs regarding the Bitches Brew version with FaitalPro 15HX500 coaxials. Still have some private issues that prevent me from starting the project but would like to check which speaker drivers would fit. I was thinking on using FaitalPro 18HP1030 or SB Audience NERO-18SW1900D drivers for woofers. Besides being 18” instead of 15” is there any reason not to use them?
Thanks,
I really like your pragmatic approaches, but must admit I'm not a fan of your 'philosophy'.
Cause after you sparingly apply a little FIR (a nuke, I presume) on your IIR DSP (bow and arrow, I again presume),
seems to me you've brought a small nuke along in case the bow doesn't kill the game. Lol
And it seems both are being used, all because multi-channel FIR isn't deemed readily available...
Which I must say I've found to be a more elegant, and much simpler solution in the long run.....can't see anything 'cheating' about it.
fwiw
These designs look excellent,
I’m curious though, does any configuration lend itself to a smaller room? My room is 13x12x7.5 which is open on the long end. Pulling them out from the back wall is really not workable. I currently listen to a pair of active 2 way speakers that are powered by Hypex FA122 plate amps which have onboard dsp. I built a pair of Dick Olsher inspired BassZilla platinum speakers many years ago, which are passive open baffle but required a larger space to fully enjoy. Anyway, great work here.
Thanks
I’m curious though, does any configuration lend itself to a smaller room? My room is 13x12x7.5 which is open on the long end. Pulling them out from the back wall is really not workable. I currently listen to a pair of active 2 way speakers that are powered by Hypex FA122 plate amps which have onboard dsp. I built a pair of Dick Olsher inspired BassZilla platinum speakers many years ago, which are passive open baffle but required a larger space to fully enjoy. Anyway, great work here.
Thanks
Given the quality of gear you're used to, getting the Flex 8 is likely a necessary step.
I'm glad you're starting to get more normal sound now! Out of curiosity, are you still running yours with naked sides, or have you put wings on like the original BB?
I got my live edge dipols more or less together. Only change to the original is the woofer. I got a the SB Audience Bianco 18W450 pretty cheap so i am not using the Kappa 18LF. Today i hooked up one channel with my MINIDSP Flex and the adopted filters from the 2x4HD and so far they do sound pretty promising. I am actually postively surprised. I will run some in room measurements over the comming days to see how it actually looks like and complete the wiring of the second speaker.
Not the expert here, but OB/dipole designs rely on the 6 ms reflection delay which requires being 3 ft or more from the back wall.
A design that might be better for you, but that I haven't myself, is the GR Research Studio monitor, which are designed to be ok closer to a back wall. The tweeter is OB and has an angled surface behind it that bounces the sound more upwards, helping get the time delay. The midwoofer is sealed and blends well with sealed subs.
I do admit, calling IIR DSP a "bow and arrow" is a bit of a stretch.I really like your pragmatic approaches, but must admit I'm not a fan of your 'philosophy'.
Cause after you sparingly apply a little FIR (a nuke, I presume) on your IIR DSP (bow and arrow, I again presume),
seems to me you've brought a small nuke along in case the bow doesn't kill the game. Lol
And it seems both are being used, all because multi-channel FIR isn't deemed readily available...
Which I must say I've found to be a more elegant, and much simpler solution in the long run.....can't see anything 'cheating' about it.
fwiw
And as I said, they do sound fantastic without FIR.
The point being that I greatly prefer to use no more brute force than absolutely necessary (especially steep crossover slopes) and I've found shelf, peak and notch filters much more phase friendly than HP or LP filters.
My room is 8.5H x 12 x 14 feet. The Bitches Brews are arguably too big for it but they sound great anyway.
The BIANCO-18SW450 looks like an excellent choice, can't wait to hear your report when you get your system working.
Some quick measurements in room with REW .
This is the orignal Filter response from the 2x4HD file you provided :
This is the response with updated EQ settings only. I did not go yet through the full circle with REW and Rephase. Only tried to update the EQ to fit the Harman curve.
Against the Harman curve
Apologies for the axis scaling, i am dealing with REW the first time. Before it was all passiv for me.
This is the orignal Filter response from the 2x4HD file you provided :
This is the response with updated EQ settings only. I did not go yet through the full circle with REW and Rephase. Only tried to update the EQ to fit the Harman curve.
Against the Harman curve
Apologies for the axis scaling, i am dealing with REW the first time. Before it was all passiv for me.
It seems this thread has drifted towards the Bitches Brew, which is great, however, I'm planning a "Live Edge Birch Dipole 3-way" build that Perry posted as the winner of the Parts Express Open unlimited, utilizing the upgraded Be tweeters & miniDSP Flex 8.
My schematic interpretation skills aren't that sharp, so I spent time visualizing the wiring of this design and have attached a visual of what I think is the correct wiring. I'm wondering if Perry or anyone familiar with the updated design can "proof" my interpretation??? I'm getting close to starting this project & do not want to cause any damage by screwing up the wiring in the tweeter section in particular - given the cost of these drivers. Thanks in advance for your expertise. I'm super excited to get this project started soon!
My schematic interpretation skills aren't that sharp, so I spent time visualizing the wiring of this design and have attached a visual of what I think is the correct wiring. I'm wondering if Perry or anyone familiar with the updated design can "proof" my interpretation??? I'm getting close to starting this project & do not want to cause any damage by screwing up the wiring in the tweeter section in particular - given the cost of these drivers. Thanks in advance for your expertise. I'm super excited to get this project started soon!
Attachments
Last edited:
Really glad to see you doing this and looking forward to your results.
With all due respect, I think the diagram here is harder to read than the original schematic, it's more complex and it has some problems. See attachment.
As long as the 20uF capacitor is between the tweeter circuit and the amp, and as long as you're not sending high power signals, you're not in danger of damaging the tweeters.
With all due respect, I think the diagram here is harder to read than the original schematic, it's more complex and it has some problems. See attachment.
As long as the 20uF capacitor is between the tweeter circuit and the amp, and as long as you're not sending high power signals, you're not in danger of damaging the tweeters.
Attachments
All,
It's time to rev up a new project and I'm selling the original Live Edge Dipoles shown in this thread. Details at https://www.diyaudio.com/community/...-edge-active-speakers-dipole-with-dsp.410128/
It's time to rev up a new project and I'm selling the original Live Edge Dipoles shown in this thread. Details at https://www.diyaudio.com/community/...-edge-active-speakers-dipole-with-dsp.410128/