I have built a pair of active two-way speakers and I would like to share the finished product with you. Thanks to Digital Sound Processing it is possible to implement very steep crossover filters without the additional phase shift. Furthermore, the discrepancies between the drivers and the diffraction effects can be corrected. If necessary, even room-modes can be partially compensated for but that has not been done thus far.
When it comes to the enclosure, a sealed configuration has been chosen for. Partially because the tuning is more forgiving but also because there is no delay between the sound coming from the cone of the woofer and the sound that would have come out of the port, if a ported enclosure had been chosen for. This is why a sealed enclosure is often considered ‘tighter’ in the low-end. Not only that, the bas reflex principle is a resonance principle. Therefore, the system has to get going when a frequency at which it resonates is being reproduced and that it takes a while for the resonance to die out when that frequency is not reproduced a e. Luckily, the human ear is not very sensitive at the frequencies where these phenomenon’s occur.
A bass-reflex enclosure allows for more output in the lower octaves from a relatively small speaker. These speakers however are large enough to provide some low-end and therefore I decided against a bass-reflex design to avoid the tuning issues that come with them.
The internal construction is quite complex. It has been designed as such to prevent the soundwaves from bouncing between parallel surfaces, thus from becoming standing waves. The bracing, though, has been designed in such a way that it is still a feasible option for a hobbyist. Almost all of the panels are rectangular in shape and some have a simple miter joint. Because of the use of dyed MDF, applying a simple transparent lacquer will give decent results.
The pictures below give an impression of the construction and the end result.
The frequency response and the phase response also give an impression of the results that were obtained
The end result sounds very good: Tight bass reproduction and a lot of detail in the highs. Compared to other speakers, the imaging can be considered sublime. You can really locate the numerous instruments and vocals reproduced, left, right, upward, down below etc. The three dimensional recordings such as the ‘’Virtual Barbershop’’ are also realistically represented, despite the fact that these are actually meant to be listened to using a pair of headphones. All of this is thanks to the linear phase response.
More information on the building process, measurements and DSP programming can be found in this PDF document.
For the time being, the speakers are still inside the mancave, my wife doesn’t want them in the house, but the kids love to ‘provide shelter’ for the speakers in their rooms.
Because the wife doesn’t want any visible speakers we have floating dresser hanging from the wall above which the TV is mounted. When not in use, the TV looks like a painting instead of a TV. Below all of this a soundbar is mounted to make sure it is out of sight.
The surround speakers have been implemented in a wall panel. This is a shallow TQWT design covered by a picture that has been printed on a special canvas. The advantage of this is that when we get tired of this picture we can simply order a new canvas for €50 to cover the enclosure of the surround speakers. The wall panel is less than 5 cm deep but the full-range speakers used were 4 cm deep. Therefore a small recess has been made in the rear panel of the enclosure to make room for the magnet of the drivers.
The baffle, of course, is huge, causing the effect of baffle step to be almost non-existent. Especially compared to tiny speakers that came with the soundbar, which sounded a little thin in the lower midrange. This magical picture; however, sounds considerably better than the tiny soundbar speakers.