I have a little experience building other people's speaker designs and a bit of experience designing my own. Mainly this was with passive crossovers and now in the world of DSP & Dirac I wondering what is considered the right approach to designing.
For example, we used to look for drivers that had basically flat response in the passband. Is that even needed anymore because as far as I can tell Dirac is going to EQ that out. Level matching, time alignment, baffle step do we need to think about this anymore? If I wire my tweeters out of phase will Dirac fix that too?
I would assume inconsistent radiation patterns would lead to issues that could only be fixed for a limited seating area in the room. Break up modes would need some correction/notch still?
I probably am not thinking about this correctly, so what's the right way to think about this?
For example, we used to look for drivers that had basically flat response in the passband. Is that even needed anymore because as far as I can tell Dirac is going to EQ that out. Level matching, time alignment, baffle step do we need to think about this anymore? If I wire my tweeters out of phase will Dirac fix that too?
I would assume inconsistent radiation patterns would lead to issues that could only be fixed for a limited seating area in the room. Break up modes would need some correction/notch still?
I probably am not thinking about this correctly, so what's the right way to think about this?
Only linear distortions can be fixed with DSP. For example, uneven radiation pattern cannot be fixed. "Dips" on frequency response due to poorly designed passive crossover cannot be fixed as well.
Too many DSP can sound unpleasant. For example, fixing a dip that's absent off-axis will make the sound more harsh. Strong phase unwinding can sound harsh too.
So, my opinion is that's still very important to design a good loudspeaker in the first place.
Hi Alex,
Unfortunately there's a really bad pattern I have seen in some active speaker DIY systems.
It is something like this:
Pick some crossover points.
Measure the final result
EQ the living hell out of it to hammer the system into submission.
The best active systems are fully analyzed, carefully integrated.
Unfortunately there's a really bad pattern I have seen in some active speaker DIY systems.
It is something like this:
Pick some crossover points.
Measure the final result
EQ the living hell out of it to hammer the system into submission.
The best active systems are fully analyzed, carefully integrated.
The right approach is to design for smooth, flat response using standard measurement, Dirac or no Dirac. What you can do that would be hard with passive is use drivers of very different sensitivity, and apply lots of EQ to a sealed box, or above tuning to a vented/PR box.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.