While messing around with a center channel this week, I accidentally created the most realistic soundstage I've ever heard in a car. Considering that I've been studying this subject for the better part of twenty years, I thought some others might enjoy reading about what I did, and some further experiments to improve things.
A lot of my previous work on the subject is over on diymobileaudio.com. However, the more that I work with car stereo, the more I realize that the tricks we use in the car can be used at the home and office.
For instance, I sit in front of a computer for 8-12 hours every day, and the work day would be a lot more enjoyable with a high quality stereo. Many of the tricks featured in this thread will also work in very small rooms, for instance a home office or a computer gaming room.
What I am going to explore in this thread is using psychoacoustic tricks to fool the listener into believing that they hear speakers which aren't there.
We've all seen these HDTV soundbars, which attempt to recreate a five-channel experience using as many as a hundred speakers in a single enclosure. The way that the good ones work is that they create a beam of sound which is focused at specific locations in the room, creating reflections which fool the listener into believing there are speakers there (which don't exist.)
Here's the thing though - they suffer from distortion, due to the small size of the drivers, and their goals are quite ambitious - five channels from one enclosure.
What I am going to document in this thread is the use of a single enclosure to cover stereo high frequencies, using waveguides to control the beam, and compression drivers to lower distortion down to inaudible levels.
If my engineering goes according to plan, we'll have two virtual speakers which seem to emanate from the left and the right.
If you're interested in reading further, see these threads:
Creating a Soundstage with Waveguides and Psychoacoustics - diyAudio
Creating The Perfect Soundstage - DIY Mobile Audio
Anyone Tried Using One Tweeter? - Page 2 - DIY Mobile Audio
A lot of my previous work on the subject is over on diymobileaudio.com. However, the more that I work with car stereo, the more I realize that the tricks we use in the car can be used at the home and office.
For instance, I sit in front of a computer for 8-12 hours every day, and the work day would be a lot more enjoyable with a high quality stereo. Many of the tricks featured in this thread will also work in very small rooms, for instance a home office or a computer gaming room.
What I am going to explore in this thread is using psychoacoustic tricks to fool the listener into believing that they hear speakers which aren't there.
Here's the thing though - they suffer from distortion, due to the small size of the drivers, and their goals are quite ambitious - five channels from one enclosure.
What I am going to document in this thread is the use of a single enclosure to cover stereo high frequencies, using waveguides to control the beam, and compression drivers to lower distortion down to inaudible levels.
If my engineering goes according to plan, we'll have two virtual speakers which seem to emanate from the left and the right.
If you're interested in reading further, see these threads:
Creating a Soundstage with Waveguides and Psychoacoustics - diyAudio
Creating The Perfect Soundstage - DIY Mobile Audio
Anyone Tried Using One Tweeter? - Page 2 - DIY Mobile Audio
Anyone Tried Using One Tweeter? - Page 2 - DIY Mobile Audio
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
OT: I used my ATC mids and one Stage Accompany ribbon tweeter as a sound bar under my 42" TV for a couple of months. Sounded good, but I prefer stereo, with 2 tweeters.
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
OT: I used my ATC mids and one Stage Accompany ribbon tweeter as a sound bar under my 42" TV for a couple of months. Sounded good, but I prefer stereo, with 2 tweeters.
Just read your other threads. Wow, great work.
I'm going to try out what you did with the 2" drivers mounted in the middle of the dash. Are these the drivers you used Parts-Express.com: Peerless 830970 2" Full Range Driver 4 Ohm | Peerless P830983 2" Full Range Driver tymphany vifa danish speaker technology dst lat250 Symphony of Savings Peerless070109 npa00453? Either way I think that's what I'm going to use.
Keep up the great work!
I'm going to try out what you did with the 2" drivers mounted in the middle of the dash. Are these the drivers you used Parts-Express.com: Peerless 830970 2" Full Range Driver 4 Ohm | Peerless P830983 2" Full Range Driver tymphany vifa danish speaker technology dst lat250 Symphony of Savings Peerless070109 npa00453? Either way I think that's what I'm going to use.
Keep up the great work!
Patrick,
Have you ever tried just mono? I've been doing it for years and it's still my preferred method for listening to music. The sweet spot is everywhere you'll practically be and set up is like falling off a log.
Dan
Have you ever tried just mono? I've been doing it for years and it's still my preferred method for listening to music. The sweet spot is everywhere you'll practically be and set up is like falling off a log.
Dan
I am getting very good results with a Peerless 830970. I'm feeding it about four times as much power as it's supposed to get, and it's taking it like a champ. Then again, I have it crossed over at 3khz, so that helps 🙂
While it's marketed as a woofer, due to it's very small size, it's makes a compelling tweeter. It has a copper cap and an underhung motor. So if you use it down to 200hz, it's distortion performance is good considering it's size. But if you limit it to high frequencies only, you're basically looking at a tweeter that can run with the big boys that cost $50 and up.
As a bonus, it's smaller than 80% of the tweeters out there, and it has enough excursion to be used with very shallow slopes.
I am also using it dipole, as I've found that dipole tweeters sound very 'open'.
Mono, I've been thinking on and off over the years. Eventually I never do because it's hard to integrate with video section of the system 🙁
The height of (TV) screen is at eye-level, which is almost the same as ear-level the speaker(s) should be. And it's not so practical (for me) to use projector/screen all the time. So I've still not figure out how to put them together🙁
Two separate systems? Good idea, but the space is not allowed🙁
The height of (TV) screen is at eye-level, which is almost the same as ear-level the speaker(s) should be. And it's not so practical (for me) to use projector/screen all the time. So I've still not figure out how to put them together🙁
Two separate systems? Good idea, but the space is not allowed🙁
Hello,
How big is this one enclosure going to be? And what would the maximun listening distance be?
Thanks.
Your goal is to make the listener to hear the speaker? 😀 I think what you like to do is to make a phantom image at the location where there is no speaker?
Some say a stereo setup can do this, but I'm not convinced 🙄
For this, you might find Ambiophonics satisfying.
Stereo from one speaker for home listening:
http://www.diyaudio.com/forums/multi-way/10962-stereolith-loudspeakers-question.html
- Elias
You can do it with just traditional stereo. You can get phantom out of speaker localization up to about 90 degrees ime. BUT it is not as strong as just sticking a speaker in that location. It always seems to be bandwidth limited even if the source isn't. And on some speakers it can sound "phasey". I've never had much luck with the front speakers that bounce sounds of the side walls though. I have had success with rear Bipolars that are similar but I really don't think it has anything to do with the side wall as much as it is summing localization.
I tried ambiophonics in the past, and found that the processing sounded unnatural and distracting. As you know, it's possible to do the same thing with a physical barrier.
In this thread I'll be doing something similar. But instead of a physical barrier, I will use a waveguide to narrow the response of the high frequency drivers.
Theoretically, this should give us the best of all worlds:
- Using a wall to eliminate crosstalk is effective, but it looks ridiculous
- Using DSP to eliminate crosstalk *should* work, but sounds unnatural to me. It also reduces dynamic range.
- A carefully designed waveguide can limit radiation even more effectively than a physical barrier.
Stay tuned...
😀 Got my drivers today!
I have no tools right now(they are all back at my parents 🙁) So I made a mock setup using a plastic container, hot glue, electrical tape, and toiletpaper rolls.
The CTC spacing is only 3" or so, and yet sitting on either side of them(at my computer desk, but trying to mimic the car) it still produces a decent stereo image.
Cool
Especially with the ambiophonic test tracks...
Anywho I can't wait to build the actual waveguide 😀
I'll be crossing at 3kHz since that's where my crossover is(it's passive 🙁)
I have no tools right now(they are all back at my parents 🙁) So I made a mock setup using a plastic container, hot glue, electrical tape, and toiletpaper rolls.
The CTC spacing is only 3" or so, and yet sitting on either side of them(at my computer desk, but trying to mimic the car) it still produces a decent stereo image.
Cool
Especially with the ambiophonic test tracks...
Anywho I can't wait to build the actual waveguide 😀
I'll be crossing at 3kHz since that's where my crossover is(it's passive 🙁)
Are you using the Peerless 2" drivers? I posted some polar measurements on Sunday. They're in another thread I started here. The Peerless has very nice response in a shallow waveguide, though the response above 15khz is wonky. Luckily we're not very sensitive to peaks and dips in that range.
Yep the 2" peerless ones.
😀 I'll check the other thread. Thanks for the update!
I've been using them for about 4 hours as fullrange dipoles now. They sound pretty good for 2" drivers. And definitely have no problem playing up high.
Edit to clarify
830970 is the peerless driver I'm using.
😀 I'll check the other thread. Thanks for the update!
I've been using them for about 4 hours as fullrange dipoles now. They sound pretty good for 2" drivers. And definitely have no problem playing up high.
Edit to clarify
830970 is the peerless driver I'm using.
Last edited:
I think eliminating the cross talk is a perfect goal! If there is no cross talk you have all the possibilities to steer the image as you like.
With waveguides you are limited strictly above 1kHz, or so. Ideally cross talk cancellation should go about two octaves lower for best results. Propably you could use normal stereo below 1kHz? After all, Blumlein stereo only works at low freqs as intended 🙄
In order to proceed in a straigthforward way, one could first measure the cross talk cancellation obtainable from a physical barrier, and then see if a waveguide design is possible with comparable performance.
What level of cancellation (in dB's) you are planning for?
- Elias
With waveguides you are limited strictly above 1kHz, or so. Ideally cross talk cancellation should go about two octaves lower for best results. Propably you could use normal stereo below 1kHz? After all, Blumlein stereo only works at low freqs as intended 🙄
In order to proceed in a straigthforward way, one could first measure the cross talk cancellation obtainable from a physical barrier, and then see if a waveguide design is possible with comparable performance.
What level of cancellation (in dB's) you are planning for?
- Elias
Here's a status on the project:
In an ambiophonic setup, a physical barrier is used to prevent crosstalk between the two sides. I believe a waveguide can be used for the same purpose. Also, due to the way that our hearing mechanism works, the location of the midbass frequencies has a greater impact on our perception of stage width and depth than the location of the tweeters. Due to this, if we pay VERY careful attention to the design of the tweeters, we can relocate them to the center. We'll still have stage width, and we'll also have better definition of the soundstage because early reflections are reduced. Intelligibility is also improved.
Here is what I settled on. Two waveguides in the center. The driver is a Celestion CDX1-1425
A view of one tweeter. Note the careful integration into the curves of the dash
In order to "push" the tweeters as far back as possible, the compression driver was taken apart, and the first 4cm were replaced with a ninety degree bend. This allows a VERY shallow depth while maintaining pathlength
I will be posting subjective evaluations, as well as measurements. I also have pix of other things that I tried, including dipole waveguides and waveguides driven by flat pistons.
In an ambiophonic setup, a physical barrier is used to prevent crosstalk between the two sides. I believe a waveguide can be used for the same purpose. Also, due to the way that our hearing mechanism works, the location of the midbass frequencies has a greater impact on our perception of stage width and depth than the location of the tweeters. Due to this, if we pay VERY careful attention to the design of the tweeters, we can relocate them to the center. We'll still have stage width, and we'll also have better definition of the soundstage because early reflections are reduced. Intelligibility is also improved.
Here is what I settled on. Two waveguides in the center. The driver is a Celestion CDX1-1425
A view of one tweeter. Note the careful integration into the curves of the dash
In order to "push" the tweeters as far back as possible, the compression driver was taken apart, and the first 4cm were replaced with a ninety degree bend. This allows a VERY shallow depth while maintaining pathlength
I will be posting subjective evaluations, as well as measurements. I also have pix of other things that I tried, including dipole waveguides and waveguides driven by flat pistons.
What happened to the peerless driver 🙁?
Did it die?
😀 I hope I have time this weekend to play around with mind.
Did it die?
😀 I hope I have time this weekend to play around with mind.
Well the reason that I had the Peerless wasn't because I have some magical affection for dipoles. I see a lot of people on the board gush about the "boxless" sound, but the truth is that it has more to do with directivity than the lack of a box. (IE the lack of a box is what you see, but it's the change in directivity pattern that you're hearing.)
Long story short, I used a dipole to increase the efficiency of the Peerless.
There's a few problems with this approach however:
- When I tried a monopole in a sphere, it sound audibly superior to the dipole. I believe this is due to better control over the directivity.
- The efficiency dropped like a rock, severely limiting headroom
- The monopole had better "articulation" than the dipole, and that's very important to me
Basically, the monopole sounded better than the dipole, but I could never live with something that can't get loud. I messed around with some array sims, but the results were less than promising 🙁
In the end, it was obvious that the best solution was a monopole that could take some power, and that's where compression drivers come in 🙂
If you can live with moderate SPL levels, the Peerless in a sphere is easily the easiest solution, and sounds very very good.
If anyone is too lazy to build an enclosure for the Peerless 830970, I found a quick solution. I was at Fry's Electronics yesterday, and stumbled across this:
This isn't a driver that's similar to the 830970, or a close cousin. This is the exact same woofer. I took some pics with my camera phone, just to be sure. I'll upload those later. The speakers are named "Creative Gigaworks T3."
To give you an idea of how good these drivers are, the average set of computer speakers are $50. These are $280. So do the math 🙂
Interesting observations. Specifically about the midbass vs. the tweeter in re-creating the soundstage. Just to clarify, what do you define as midbass frequencies?
Also, about directivity vis-a-vis dipole vs. monopole in a sphere. What kind of directivity does a dipole have that gives it the "boxless sound"? And how does a monopole in a sphere have "better" control over directivity?
Also, about directivity vis-a-vis dipole vs. monopole in a sphere. What kind of directivity does a dipole have that gives it the "boxless sound"? And how does a monopole in a sphere have "better" control over directivity?
- Status
- Not open for further replies.