-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
3 Way bookshelf the easy way
- Thread starter modulardesign
- Start date
OP
- Joined
- Mar 17, 2022
- Messages
- 190
- Likes
- 19
- Thread Starter
- #24
I have incrementally improved the crossover design. Crossover calculators do a meh job. Its all in the crossover after the box and drivers have been picked, quite a lot of fiddling here, rounding the edges of the baffle does wonders for higher frequencies
OP
- Joined
- Mar 17, 2022
- Messages
- 190
- Likes
- 19
- Thread Starter
- #26
Speaker Dispersion, Psychoacoustics, and Room Interactions in Real-World Applications
In real-world environments, loudspeakers must balance dispersion characteristics with room acoustics. Horizontal and vertical dispersion patterns are critical: overly wide dispersion in reflective spaces (e.g., outdoors or untreated rooms) can exacerbate unwanted reflections from walls, floors, or nearby structures. Conversely, narrow dispersion may reduce off-axis interference but risks shrinking the "sweet spot" for stereo imaging.
Psychoacoustics and Listener Preference
Human auditory processing actively suppresses reflections to focus on the direct sound source—a phenomenon known as the precedence effect. This explains why some speakers with imperfect measurements (e.g., uneven frequency response) may still subjectively please listeners: the brain prioritizes clarity in critical midrange frequencies (300 Hz – 4 kHz) while tolerating minor aberrations in less sensitive bands. However, excessive high-frequency energy or chaotic off-axis radiation can cause listener fatigue, as the brain struggles to resolve conflicting spatial cues.
Vertical vs. Horizontal Dispersion Tradeoffs
Vertical dispersion is often engineered to be narrower than horizontal to mitigate floor/ceiling reflections, which are typically more disruptive than lateral ones. However, ceiling-mounted speakers (e.g., in theaters) require broader vertical coverage to maintain even sound distribution. Center-channel speakers, tasked with anchoring dialogue, prioritize wide horizontal dispersion to accommodate off-axis listeners.
Room Acoustics and Speaker Types
Nearfield/midfield listening dominates home setups, favoring bookshelf or tower speakers. Bookshelves rely on stands or placement flexibility to minimize boundary interactions, while towers leverage extended cabinets for deeper bass (augmented by room gain) and multi-driver coherence. Dipoles or open-baffle designs, though praised for minimizing rear-wall reflections, demand careful room treatment to avoid comb filtering from uncontrolled rear radiation.
Key Design Considerations for Bookshelf Speakers
In real-world environments, loudspeakers must balance dispersion characteristics with room acoustics. Horizontal and vertical dispersion patterns are critical: overly wide dispersion in reflective spaces (e.g., outdoors or untreated rooms) can exacerbate unwanted reflections from walls, floors, or nearby structures. Conversely, narrow dispersion may reduce off-axis interference but risks shrinking the "sweet spot" for stereo imaging.
Psychoacoustics and Listener Preference
Human auditory processing actively suppresses reflections to focus on the direct sound source—a phenomenon known as the precedence effect. This explains why some speakers with imperfect measurements (e.g., uneven frequency response) may still subjectively please listeners: the brain prioritizes clarity in critical midrange frequencies (300 Hz – 4 kHz) while tolerating minor aberrations in less sensitive bands. However, excessive high-frequency energy or chaotic off-axis radiation can cause listener fatigue, as the brain struggles to resolve conflicting spatial cues.
Vertical vs. Horizontal Dispersion Tradeoffs
Vertical dispersion is often engineered to be narrower than horizontal to mitigate floor/ceiling reflections, which are typically more disruptive than lateral ones. However, ceiling-mounted speakers (e.g., in theaters) require broader vertical coverage to maintain even sound distribution. Center-channel speakers, tasked with anchoring dialogue, prioritize wide horizontal dispersion to accommodate off-axis listeners.
Room Acoustics and Speaker Types
Nearfield/midfield listening dominates home setups, favoring bookshelf or tower speakers. Bookshelves rely on stands or placement flexibility to minimize boundary interactions, while towers leverage extended cabinets for deeper bass (augmented by room gain) and multi-driver coherence. Dipoles or open-baffle designs, though praised for minimizing rear-wall reflections, demand careful room treatment to avoid comb filtering from uncontrolled rear radiation.
Key Design Considerations for Bookshelf Speakers
- Dispersion: Target ±60° horizontal dispersion for stereo imaging, with vertical controlled to ±20° to limit floor/ceiling bounce.
- Driver Selection: A 5.25–6.5" midwoofer paired with a waveguide-loaded tweeter balances directivity matching and power response.
- Cabinet Design: Braced, non-parallel walls reduce internal resonances; a front port minimizes boundary coupling.
- Power Handling: 50–100W RMS (86–90 dB sensitivity) ensures headroom for dynamic peaks in small-to-medium rooms.
- Room Adaptation: Include boundary EQ (e.g., -3 dB LF shelf below 150 Hz for wall-mounted placement).
Look, directivity isn‘t all. You‘ve got a lot problems with the design. The inductors in bass, barely available, inducing big wiggles in the bass, suckouts in the vertical, doesn‘t fit a real shelf. Frankly, do you think this project would push the envelope, so would be worth the work in DIY?
OP
- Joined
- Mar 17, 2022
- Messages
- 190
- Likes
- 19
- Thread Starter
- #30
Thanks, what looks like bass suck out is an attempt to go lower in bass, in a real room the speaker may measure differently. The vertical losses at 3k maybe a blessing in disguise where acute hearing loss is not an issue. There is a also some variation in what we hear vs what we measure sometimes the computer on our shoulders doesn't cooperate with the computer in our hands. The project is open and your free to contribute your improvement. At any one point one can only know so much.
Simple, what are the inductors in bass are going to cost? I guess them alone would be more than the speaker. Still you've got a re-action with the falling (capacitive) shoulder of the woofer's impedance, which makes roughly +/-3dB of a wiggle, two times.
Today a concept like that would be rendered active. Yesterday it wouldn't have been done at all. I suggest a low x/over to a dedicated midrange, but do it active with DSP.
a4eaudio
Senior Member
@modulardesign - Here are two of the things @Heinrich is trying to tell you...
1) Your component values are not values of inductors or capacitors you can actually buy. So while your xo may technically be correct, it is best practice to go back and tweak the components to readily-available sizes, like what I have added in black text. Also, your two inductors circled in yellow, are so big that they aren't going to be easy to purchase and will be very expensive. More on that in #2.
#2) Your low-pass filter in blue (and somewhat less, your high-pass filter on the midrange in green) should be much smoother. The bumps and dips in the blue woofer filter are a result of your component choice - that is, it is not inherent in the driver in anyway. Try to tweak the crossover components to give smoother filter functions. While your final SPL may "look" smooth, you may be getting that smoothness from resonances your are creating in your filter that will not "sound" good, despite the SPL smoothness. I think you are likely way off with the components circled in yellow above. Try to start with one of the two low-pass filters below and see if you can tweak the components to get a better result.
Note, below 500uH is 0.50mH
1) Your component values are not values of inductors or capacitors you can actually buy. So while your xo may technically be correct, it is best practice to go back and tweak the components to readily-available sizes, like what I have added in black text. Also, your two inductors circled in yellow, are so big that they aren't going to be easy to purchase and will be very expensive. More on that in #2.
#2) Your low-pass filter in blue (and somewhat less, your high-pass filter on the midrange in green) should be much smoother. The bumps and dips in the blue woofer filter are a result of your component choice - that is, it is not inherent in the driver in anyway. Try to tweak the crossover components to give smoother filter functions. While your final SPL may "look" smooth, you may be getting that smoothness from resonances your are creating in your filter that will not "sound" good, despite the SPL smoothness. I think you are likely way off with the components circled in yellow above. Try to start with one of the two low-pass filters below and see if you can tweak the components to get a better result.
Note, below 500uH is 0.50mH
As was stated by somebody else in more clarity than I dared to convey, the project is an utter fail.
OP
- Joined
- Mar 17, 2022
- Messages
- 190
- Likes
- 19
- Thread Starter
- #35
Thanks for the hints , here are some results, am hoping winding the inductors with 18gauge wire will bring down the costs of the crossover, for capacitors maybe use several in parallel to bring to better tolerance levels or just initially work with 5-10%
The biggest uncertainty is the edges of the housing. You don't need to worry about the capacities.
It now also turns out that the midrange driver is a little too low-volume. The tweeter is used far too high up in the frequency range, it can be used from 2kHz or even from 1.5kHz. Then you will also see whether the radiation is pointing upwards. Now it's pointing downwards, which is always a bit out of line. Many people do: obviously they haven't thought about it ...
Good luck!
This is the same poster from over on DIYaudio that used traced files and AI to come up with a design. The thread ended up being locked when nothing more positive came from it. The OP called his design faultless, and was not open to potential improvements.
