Just downloaded everything. You're right, that should be enough reading material to last for quite some time. Thanks, it looks interesting.
I just finished the first PDF from the link GM provided. Cool history lesson on how we ended up with two stereo channels as a standard for music playback in the first place, and how multi channel can be really useful for music. I'm gonna start on the second PDF from his link this evening.
-Chris
Hey guys,
Haven't posted on this lately but krivium has been tutoring me over emails the last few days. He really helped me understand how my two 10" drivers would interact with each other, off axis, vertically and horizontally.
Considering just the mid bass modules, everything looked good up to my planned crossover point. Above that crossover point, vertical lobing quickly started to be a factor. But everything looked pretty darn good below ~500Hz.
Then we considered adding in the Scanspeak 12MU, crossed over at 400Hz. Horizontal response was really nice up to about 2K but vertically things got nasty at and a little below the crossover point. And, unfortunately, we already know the mid bass modules will have vertical issues on their own if we try to experiment with raising the crossover point to see if it might help.
I suspect I could live with the vertical axis issues and especially so with the shape of my roof. And Remi did tell me there are options for reducing the vertical lobing issues, or maybe even taking advantage of them. But I think I'm going to ditch the two tens stacked vertically for something that will, hopefully, be more flexible. A single 12" driver per bass module.
Beaming will start at a little lower frequency, and I'll lose a little headroom compared to the cone area of two 10" woofers, but it seems to give me a much wider range of crossover points. I can still cross over to a Scanspeak 12MU at around 400Hz if I want to, or, with the right driver, I can run them up closer to 1000Hz and hand off to something like a SEOS 12.
So, I've been modeling a bunch of different 12" woofers the last couple evenings. The Eminence, Faital and AE woofers mentioned by LineSource all look like they could be made to work with a linkwitz transform. Especially the TD12M. The Denovo Magnum 12 looks pretty interesting but doesn't go up to 1000Hz as I would like.
I finally found the AE TD12S-4, and I think I may be in love. I know the lower sensitivity make it not pair quite as well with the SEOS when using passive crossovers, but I don't have that constraint. It looks to me like just about the perfect match for taking over at 60Hz from my subs and playing up to 1000Hz to hand off to whatever mid range solution I end up with.
WinISD shows a 3 cubic foot sealed box will give Qtc of 0.58, Fsc of 56Hz and the TFM graph shows 3DB down at 70Hz. If WinISD can be trusted, a bit of linkwitz transform can bring the F3 down to 60Hz while still not exceeding XMax at ridiculous volume levels, without even adding in a high pass filter.
The main thing I'm not sure how to interpret is being so close to the resonant frequency and impedance spike at a 60 Hz crossover point. That makes me think 60Hz maybe a little low for best results. Maybe forgetting about the linkwitz transform and crossing over at 70Hz would be a better idea.
Sorry this post got so long but do you guys think I'm on a better path here? I would start with very simple 85 liter rectangular enclosures that I expect to throw away down the road. I'll still brace them well, of course. I am considering a removable baffle though so I can move the woofers up or down based on whatever floor bounce I might get and distances between drivers I might need for adding a midrange.
Thanks,
Chris
Haven't posted on this lately but krivium has been tutoring me over emails the last few days. He really helped me understand how my two 10" drivers would interact with each other, off axis, vertically and horizontally.
Considering just the mid bass modules, everything looked good up to my planned crossover point. Above that crossover point, vertical lobing quickly started to be a factor. But everything looked pretty darn good below ~500Hz.
Then we considered adding in the Scanspeak 12MU, crossed over at 400Hz. Horizontal response was really nice up to about 2K but vertically things got nasty at and a little below the crossover point. And, unfortunately, we already know the mid bass modules will have vertical issues on their own if we try to experiment with raising the crossover point to see if it might help.
I suspect I could live with the vertical axis issues and especially so with the shape of my roof. And Remi did tell me there are options for reducing the vertical lobing issues, or maybe even taking advantage of them. But I think I'm going to ditch the two tens stacked vertically for something that will, hopefully, be more flexible. A single 12" driver per bass module.
Beaming will start at a little lower frequency, and I'll lose a little headroom compared to the cone area of two 10" woofers, but it seems to give me a much wider range of crossover points. I can still cross over to a Scanspeak 12MU at around 400Hz if I want to, or, with the right driver, I can run them up closer to 1000Hz and hand off to something like a SEOS 12.
So, I've been modeling a bunch of different 12" woofers the last couple evenings. The Eminence, Faital and AE woofers mentioned by LineSource all look like they could be made to work with a linkwitz transform. Especially the TD12M. The Denovo Magnum 12 looks pretty interesting but doesn't go up to 1000Hz as I would like.
I finally found the AE TD12S-4, and I think I may be in love. I know the lower sensitivity make it not pair quite as well with the SEOS when using passive crossovers, but I don't have that constraint. It looks to me like just about the perfect match for taking over at 60Hz from my subs and playing up to 1000Hz to hand off to whatever mid range solution I end up with.
WinISD shows a 3 cubic foot sealed box will give Qtc of 0.58, Fsc of 56Hz and the TFM graph shows 3DB down at 70Hz. If WinISD can be trusted, a bit of linkwitz transform can bring the F3 down to 60Hz while still not exceeding XMax at ridiculous volume levels, without even adding in a high pass filter.
The main thing I'm not sure how to interpret is being so close to the resonant frequency and impedance spike at a 60 Hz crossover point. That makes me think 60Hz maybe a little low for best results. Maybe forgetting about the linkwitz transform and crossing over at 70Hz would be a better idea.
Sorry this post got so long but do you guys think I'm on a better path here? I would start with very simple 85 liter rectangular enclosures that I expect to throw away down the road. I'll still brace them well, of course. I am considering a removable baffle though so I can move the woofers up or down based on whatever floor bounce I might get and distances between drivers I might need for adding a midrange.
Thanks,
Chris
Not to be repeating what Remi has said would be useful.
So he is basically explaining the function of driver diameter and the transition between hemispherical sound pattern and lobing due to frequencies of interest being of greater wavelength than the active diameter of the driver.
Beaming!
Me
Up
Scotty?
So he is basically explaining the function of driver diameter and the transition between hemispherical sound pattern and lobing due to frequencies of interest being of greater wavelength than the active diameter of the driver.
Beaming!
Me
Up
Scotty?
Not quite sure what you mean by that. Are you suggesting that I copied and pasted what Remi sent me without understanding it? That's kind of how I read it but hopefully that isn't what you mean.
-Chris
Yes.
Not sure what you mean by that (lost in translation). We looked after physical layout of drivers and about vertical stacked units and behavior in vertical directivity and horizontal directivity: comb filtering once you are off axis, in either vertically stacked, horizontal pair layout or with a single driver .
We talked about strategys about integer BSC only in the midbass driver(s) (original 2x10"), introduced Chris about different filters choice an pros and cons of them (LR and Bessel studys from Rane white papers). This led us to talk about CTC distance of differents drivers, acoustic coupling and relationship with frequency and drivers (the ubiquitous nescessity of keeping 1/4wavelength ctc at fc), what happens when you increase CTC distance in vertical layout an horizontal ones, i teased briefly about mtm configuration and why i do think vertical mirror stacked layout are beneficial from my point of view versus tmw or tmww but theorycal cons too (including link to biro technology parper about d'Appolito/mtm). I introduced Chris about bounce floor issue and how you could try to minize its effect using vertical directivity behavior too.
I explained how to read polar maps and polar plots. We drifted to mid/high possible choices with pro and cons of some exemple including waveguide (SEOS) and 12MU and fc to choose... The need to explore many layout and choice of drivers before doing choice for the final layout... and need of other advice than mine!
Probably others things too but now i'll have to check our exchange. Most of my explanation was illustrated by polar maps and plots of simulated scenario... lots of thing in little time! I may have given wrong informations (i hope not!) and or not be clear enough about what i said (but done my most to be clear enough and easily understable), so the need for others point of view than mine. And didn't want to induce my biased choice in Chris mind! 🙂
Last edited:
No Chris.
I have zero reason to to think or say nything but good things about you.
My cannon is for Remi only. 😀
The reason is self evident in what Remi posted. I don't want to type a whole bunch of the same things you have come to understand.
Nice!
Chris's head will explode some time soon if we are not careful!
Has the subject of crossover slope and directivity at or near crossover points been discussed?
Not directly, but it appeared in some plots.
I wanted Chris first have a look about pro and cons of different filters type about phase behavior, transient smearing and slope steepness and how they can help or be detrimental for others parameters first.
I simulated one of the scenario including 2*10" + 12mu filtered at 400hz using 24LR in vertical layout. Other sims were either full range either lp 4pole LR at 400hz included.
Most of them was for theorical/educational purpose but as Chris had Daytons in mind i used them as starpoint (Derivated radius from SD and used overall diameter and practical issue -gasket clearance of 2cm min- for the build for being near a practical implementation).
My point was/is to give some basis points about a whole design... Already talked about BSC and baffle edge diffraction before. As well as pro and cons about multi amp dsp filtering.
Thanks! My pleasure. 😀
Last edited:
Ok.
Looks like you are covering the information very well, and I will stay out of the way.
Send me a PM or an email if I can be of assistance.
And I will watch this thread.
Few times have I seen one set up with such a well thought out method.
Looks like you are covering the information very well, and I will stay out of the way.
Send me a PM or an email if I can be of assistance.
And I will watch this thread.
Few times have I seen one set up with such a well thought out method.
Oh no, please don't! I really not an expert and do believe that i've already given most of my knowledge to Chris... Your thoughs and opinions are a great advantage for Chris project!
Well there is really no method just discussions with Chris in private. There is so many ways to approach the problem... You just have to make a choice in what to begin with: and i do believe that it's not the brand of the drivers which are the first interesting step in the process. 🙂
Awesome. Thanks, Mark.
But don't you be shooting any cannons at Remi. He has spent a huge amount of his personal time trying to educate me. I wonder at times if he ever sleeps. I'm taking it in as fast as I can but have enough bookmarked reading material to last me a month.
In the computer industry we call it drinking from a firehose. There is so much information available that you have to choose a portion of it on which to focus at a given time. It does help a lot that I have a decent handle on the electrical side of things already. And I already knew what beaming is but I didn't quite have a grasp of why it really mattered.
I'm trying to focus in on getting these mid bass modules as "right" as I can get them for now. Then I can focus on the mid range and treble as I learn more.
Remi didn't once tell me my two 10" woofers stacked vertically were a bad idea or wouldn't work. Just the opposite actually. He tried to show me how to get the best from them in that config while slipping in a few gentle hints that a single driver doesn't have the same off-axis lobing issues. At least when looked at in isolation from the other drivers needed to make a real speaker.
I decided to drop the stacked 10" woofers and go with the single 12". I actually feel a little bad about it after Remi spent so much time teaching me about other config. But I think he agrees that it's a good decision.
-Chris
Please don't step back because I misunderstood what you meant to say earlier, Mark. After some nasty comments by someone else in the sub stuffing thread I thought I might have stepped on your toes in there.
I am blown away by the knowledge level of many of you guys on diyaudio. I would like nothing more than to get as much feedback as possible from every person who has posted in this thread.
-Chris
My offer stands.
I'll help where I can.
I'm not the only guy on this thread with audio design experience.
Not to thin of a skin either.
The other thread blowup was stupidity on my own part.
Anthony can be a generous guy at times. And he really is quite intelligent.
I'll help where I can.
I'm not the only guy on this thread with audio design experience.
Not to thin of a skin either.
The other thread blowup was stupidity on my own part.
Anthony can be a generous guy at times. And he really is quite intelligent.
I'm sure you're right about that but it is the step that costs $$ rather than time. And then when you add in the fact that I'm impatient and want some sweet mid bass ASAP, it's hard not to focus there.
I'm feeling pretty good about the AE TD12S, even though I wasn't planning to spend $700 on the drivers right away. Ignoring the cost they seem like a really versatile and high quality driver that I could use in many different ways. Could I really go too wrong with them?
I actually think a set of those Dayton RS270 woofers will still be in my future at some point. But I see them in a classic 3-way like speaker for a different room.
Thanks,
Chris
Last edited:
1) why a dedicated midbass like the TD12M is frequently used.
2) why a large midbass covering ~1300Hz tweeter crossover is typically crossed ~100Hz - 150Hz to avoid audible Doppler distortion, and to also improve clarity around the baffle step + room modes transistion.
3) why high-efficiency 3-way main speakers covering 25-18Khz have advantages over 2-way plus remote 80Hz woofers.
P.S. When DSP with time-delay alignment is used, the "improved" pattern control provided from the 5" deep SEOS-15 or 5.9" deep H290C horn from pi-speakers is sometimes favored over the 3.9" deep SEOS-12. Worth some study.
------------
When a tweeter horn/waveguide is used, the polar directivity of the midbass at the crossover frequency should match the polar directivity of the horn/waveguide at that crossover frequency for a smooth, seamless transistion. A significant mismatch will garble the soundstage from the abrupt wide-narrow and narrow-wide directivity changes. The SEOS-12 or SEOS-15 waveguides require matching their 90H-polar directivity at a frequency where the midbass also produces 90-degree directivity. Study of common 12" -to- SEOS12 crossovers shows 1200-1400Hz is typically used by top designers for this match.
High-frequencies riding on low-frequencies can produce intermodulation distortion(IMD), often called Doppler distortion. Simulations and listening test confirm that a 12" midbass crossed at 1300Hz to the tweeter waveguide can generate audible IMD when used below 100Hz. Modeling shows a rapid increase in IMD even from 100Hz-> 80Hz. A smaller cone area midbass can be used from 1300Hz down to 80Hz(THX/Dolby spec) with acceptable IMD, but large cone area midbass require extra attention.
If a 12" midbass has a 100-150Hz low frequency crossover, physically adjacent woofers are required to avoid confusing the listener's brain from shifting sound-stage queues. Well studied in the literature, and why THX set 80Hz as the highest remote woofer frequency.
At audio wavelengths long enough to wrap around the 180-degree baffle and split-off some of their energy into the 360-degree room volume, your DSP room equalization will sum-in a roller-coaster of compensation power. Some designers report improved midrange clarity if the larger SD woofers are "wacked around" on this roller-coaster instead of the midbass trying for audio perfection up to 1300Hz. Many "acceptable width" cabinets in "small rooms" experience substantial baffle step and room-mode equalization around 150Hz, so a 100Hz-150Hz crossover gives the big beating to the big woofers.
Attachments
I did not wrote this about you Chris: this is more a general observation that many design in diy field start with that. There's reason about that fact (being budget, driver reputation, esthetical choice, design wise,...) and i do think that it can be a nice start point... if you already know most of the points discussed here and in private.
This is an interesting step and one which is important, but not to close 'doors' from a design point of view i think it should appear latter in the design process, once you have things about 'layout' already sorted cause as you seen some small choices can already compromise or open way to good results. But you must know them and how they will interact whith other parameters.
Well crap. I see it now. I was looking at the SEOS-12 being able to cross over at 950Hz and thinking I would be in good shape with a mid bass running up to that point. But you are saying that crossing over high enough to take advantage of beaming prevents an abrupt change in directivity at the crossover point.
Remi actually touched on this but I didn't get it until now.
So how noticeable is the transition if the crossover point is down where the 12" woofer hasn't started any beaming? It seems like the rolloff from the crossover would help to smooth things in a similar way.
I'll do some reading on IMD this evening.
So, basically, if I want to use a SEOS waveguide, and take advantage of DIY community knowledge on integrating them, I need to throw away my current subs? That gets pretty hard to swallow when I haven't even heard mid range and treble through a SEOS.
I definitely see the logic behind the 18 and 12 in your examples though.
All of it seems to be a chain reaction from the desire to match radiation patterns between the mid bass and the SEOS. Which makes my question above even more important. How audible would it really be to cross my 12" woofer over at about 1000Hz rather than 1300? And why would it sound much different than the rolloff from a low pass filter?
Thanks,
Chris
The mentioned Faital Pro is a very well built and very low distortion driver.
They also measure exactly as spec'd.
I have worked with clients AE drivers. And my last experience with them was about 4 years ago. Not knocking the fellow. He is a very good guy. ( Getting drivers can be a pain, one I have been up against myself in providing drivers. So black pot calling kettle black.) But the driver was quite a bit off spec.
I'll echo LineSource is spot on for the benefits of dividing the frequency span per driver.
If this is done is a fashion that takes into consideration the simple truths in polar response that Remi has discussed you can make a very effective setup.
LineSource's comments about DSP and the driver enclosure are also very important.
You listen to the entire system. Not a driver. Not a crossover point or even a DSP unit.
DSP has been hailed as to much of a good thing in my opinion.
You can make a crappy driver measure really nicely with a decent DSP unit.
Does that mean you are getting high fidelity?
No.
It means that at one watt with a fairly easy signal you are getting a nice measurement.
Try that same driver DSP system with a proper pink noise signal and see what you get in terms of distortion!
Then you are seeing what is actually happening.
On the topic of this endeavor.
Do you have a mic of any kind?
It need not be expensive. It's more a situation of what you know how to do with it, versus the complexity and cost of the microphone system.
ARTA, REW they are both free.
And a bit of reference on how to use a measurement system is in order.
- Status
- Not open for further replies.