converting revel f208 to active crossover with 6 amps

[NJ coffee roaster]

Already have 4 mono amps, considering grabbing 2 more (probably topping b100s) and swapping minidsp flex for htx to use the htx as an active crossover, one amp per driver, totally bypassing the f208 internal crossover. Curious how much improvement I could expect if I do this.

 

[Matt_Holland]

Have you identified which areas you want to improve, or that you think need improving?

 

[DWPress]

Not having looked up the specs of your Revels - what you propose is certainly doable but I'm not sure if you'd get an improvement. It's not a trivial thing to create active crossovers correctly in the bass/sub bass region.

 

[staticV3]

Biggest issue one could likely improve with an active crossover is the cancellation dip at 2.2k:


Still a major challenge, especially w/o a way to gather anechoic driver response data in-situ.

 

[Defi]

I would be surprised if you noticed a marginal improvement. That being said... it does sound like fun and I think you should do it and document it here.

 

[kemmler3D]

Before starting I would try and track down the raw measurements of the drivers used and see where the possible improvements actually are. Everything with speaker design is a trade-off and generally when you squeeze here, it bulges there. For example, you might be able to fix the off axis crossover dip, but how much distortion are you getting in the bargain?

Overall if you rework the crossover, I am sure you can come up with something different and even reasonable compared to the original... Whether you consider it better is not guaranteed.

One thing you can try is using the same crossover points, but much steeper filter slopes. Done right this is sometimes said to improve overall sound without really altering the overall FR. The idea is that having sound not overlapping drivers is better. I'm not aware that this has been rigorously tested but it's something you can do with DSP crossovers that doesn't also require you to outsmart Revel.

It's not that easy to improve a speaker like this... Keep in mind that xover was designed by a professional who probably spent the better part of a full work week on it, maybe more. Your advantage is having no restrictions on parts or filters by virtue of using DSP, but even with a head start, we are still talking about outdoing a real audio engineer with access to professional measurement equipment. Improving crossovers is much more a thing with speakers that are super cheap and were slapped together with a cost target and the sound being an afterthought.

Your Revels are not like that... They may have wanted a slightly fancier crossover but were limited by budget, but even so, whoever did the job had performance in mind.

As others have said, identify intended areas for improvement before buying more amps.

 

[staticV3]

For example, you might be able to fix the off axis crossover dip, but how much distortion are you getting in the bargain?

Steeper Xover slopes minimize cancellation issues and, compared to analog Xover circuits, come nearly for free with DSP.

 

[NJ coffee roaster]

Before starting I would try and track down the raw measurements of the drivers used and see where the possible improvements actually are. Everything with speaker design is a trade-off and generally when you squeeze here, it bulges there. For example, you might be able to fix the off axis crossover dip, but how much distortion are you getting in the bargain?

Overall if you rework the crossover, I am sure you can come up with something different and even reasonable compared to the original... Whether you consider it better is not guaranteed.

One thing you can try is using the same crossover points, but much steeper filter slopes. Done right this is sometimes said to improve overall sound without really altering the overall FR. The idea is that having sound not overlapping drivers is better. I'm not aware that this has been rigorously tested but it's something you can do with DSP crossovers that doesn't also require you to outsmart Revel.

It's not that easy to improve a speaker like this... Keep in mind that xover was designed by a professional who probably spent the better part of a full work week on it, maybe more. Your advantage is having no restrictions on parts or filters by virtue of using DSP, but even with a head start, we are still talking about outdoing a real audio engineer with access to professional measurement equipment. Improving crossovers is much more a thing with speakers that are super cheap and were slapped together with a cost target and the sound being an afterthought.

Your Revels are not like that... They may have wanted a slightly fancier crossover but were limited by budget, but even so, whoever did the job had performance in mind.

As others have said, identify intended areas for improvement before buying more amps.

I suppose I am more asking if anyone here thinks that keeping the same crossover slopes and points as the stock passive ones but done in the dsp with individual amps will yield a benefit vs the passive ones.

I have found some information suggesting that it will, but I am not sure if this will translate to anything I will actually notice.

 

[staticV3]

I have found some information suggesting that it will, but I am not sure if this will translate to anything I will actually notice.

Replicating the analog crossover response but in digital will not likely result in audible improvements.

Real gains would come from modifying the cossover in ways that weren't possible/practical for Revel egineers when they were doing it in analog.

 

[skankhobag]

The greatest improvement will come from using brick-wall, FIR filters. This will give the following benefits:

• Total elimination of any lobing due to the crossover and smooth out any XO related frequency response anomalies

• Require around 50% less power (maybe even less) to reach the same output level since passive crossovers waste 3-6dB of power as heat dissipation

• Totally eliminate the danger of amp clipping burning out a tweeter or mid. You can also run a limiter on each driver to make sure they don’t get too much power.

• Allow you to perfectly time align the drivers

• Reduce amplifier IMD since the frequency range each amp covers is much smaller

• Increase the maximum output of the loudspeakers by around 3dB

• Running a high-pass filter just below the port tuning frequency will help protect the woofers from bottoming out if sub-bass hits the speaker (a problem with any ported design). This will save power and reduce distortion a bit when you’re playing loud music.

There are 3 things to consider:

• You’re going to need to do a lot of research on your end and invest in a measurement microphone & software to get the absolute most out of this type of rig. It’s not going to be easy.

• Your source is now going to be your computer if it’s not already

• This is not going to be inexpensive if you’re going to do it right. Don’t cheap out on anything.

You may also want to get Dirac Live into the equation since MiniDSP supports it. And, oh, yeah, use the Revel crossover points.

When you have this entire process done, your audio system will sound as good as it can with those speakers.

Having done this myself, the end result is definitely better than the passive crossovers, but it’s certainly not a night and day improvement since you’ve only improved 20% (at best) of the speakers characteristics.

 

[jhaider]

I suppose I am more asking if anyone here thinks that keeping the same crossover slopes and points as the stock passive ones but done in the dsp with individual amps will yield a benefit vs the passive ones.

I have found some information suggesting that it will, but I am not sure if this will translate to anything I will actually notice.

No that will be a big downgrade. A competent crossover includes frequency shaping and not just rolloffs.

IMO, if you’re going to do this, use a lesser speaker. A long time ago I did something similar with a Boston Acoustics speaker I bought cheap because it had a nice looking waveguide. CR57 or something like that. But that was a $50 or $60 purchase used and just for funsies, not ripping up a very good speaker still worth low-mid 4 figures.

 

[Chrispy]

I'm more of the opinion to build your own speaker if you want to do something like this rather than adapt a speaker designed a particular way with a passive crossover. Generally seems it will be hard to better the current engineering but going to great expense to do so....

 

[NJ coffee roaster]

No that will be a big downgrade. A competent crossover includes frequency shaping and not just rolloffs.

IMO, if you’re going to do this, use a lesser speaker. A long time ago I did something similar with a Boston Acoustics speaker I bought cheap because it had a nice looking waveguide. CR57 or something like that. But that was a $50 or $60 purchase used and just for funsies, not ripping up a very good speaker still worth low-mid 4 figures.

true but shaping can be done even more so in minidsp

 

[kemmler3D]

keeping the same crossover slopes and points as the stock passive ones but done in the dsp with individual amps will yield a benefit vs the passive ones.

Very likely your crossovers are still working close to spec, so it would be surprising if it made even a 2dB difference anywhere. It could make an improvement, but a very subtle and quite possibly inaudible one.

You can do a quick gut check by playing mono pink noise and muting one channel at a time. This way you can quickly compare whether each speaker has the same frequency response. It might help to listen close to the speakers to reduce the influence of the room. If you hear changes from speaker to speaker, that will be proportional to the change you can make by going to active xover.

Even if you go to the extent of doing FIR brickwall filters, you will still be putting in a decent amount of effort to get another 5% out of already-good speakers.

@Chrispy makes a good suggestion. If you're willing to mostly build a pair of 3-way speakers, starting from scratch can be more interesting and rewarding.

 

[Wolf11Man]

The most simple way to activate an existing speaker is to measure each driver nearfield, first with the crossover connected, then, with identical settings/ mike position, but no crossover connected. This eleminates any room issues.
Next you start to program the DSP, same setting and mike position, to match the recorded curve.
From the recorded measurements you (or someone who knows) can see what filters Revel created in the passive x-over.
If you match them, the sound should improve a little, more clarity and attack usually. How much depends on the amp and the passive crossover. You eliminate any sound issues of passive components and ease the load on the amplifier(s). Even the most stuborn "passive is best" fractionist can not deny this.

If you do identical filter curves in the nearfield, the in room response should not change with this method. You keep all the advantages and faults the speaker had in passive mode.
Next step can be to optimise the filters if needed. If you want to to do this right, you need advanced measurements, not only near field. This is serious stuff and you need free field, different angles and distortion tested. Then enough knowledge to do better filters, crossover frequency etc.
In general, any passive crossover is a compromise of limitations the parts budget and the manufacturers policy allows.
Number and quality of x-over parts is a typical playing field for the guys with the red marker to save costs.
Part quality is very much overrated, once you remove the known junk, boutique parts don't improve anything noticeable.
Parts number and size is another chapter, the smallest PCB with the least components that still sounds acceptable wins in a commercial products. This is the area where a good developer can make huge improvements over the passive version. Be warned, this comes close to a complete new development!

 

[Ageve]

It might be worth noting that F208 has a 24dB/octave crossover, and the nearfield response looks really good out of the box (my measurements, drivers in the cabinet, crossover connected, same method as Amir uses):

 

[JohnnyNG]

I'm more of the opinion to build your own speaker if you want to do something like this rather than adapt a speaker designed a particular way with a passive crossover.

Bingo. Especially when your target is a truly great speaker like the F208 with no glaring flaws.

 

[NJ coffee roaster]

The most simple way to activate an existing speaker is to measure each driver nearfield, first with the crossover connected, then, with identical settings/ mike position, but no crossover connected. This eleminates any room issues.
Next you start to program the DSP, same setting and mike position, to match the recorded curve.
From the recorded measurements you (or someone who knows) can see what filters Revel created in the passive x-over.
If you match them, the sound should improve a little, more clarity and attack usually. How much depends on the amp and the passive crossover. You eliminate any sound issues of passive components and ease the load on the amplifier(s). Even the most stuborn "passive is best" fractionist can not deny this.

If you do identical filter curves in the nearfield, the in room response should not change with this method. You keep all the advantages and faults the speaker had in passive mode.
Next step can be to optimise the filters if needed. If you want to to do this right, you need advanced measurements, not only near field. This is serious stuff and you need free field, different angles and distortion tested. Then enough knowledge to do better filters, crossover frequency etc.
In general, any passive crossover is a compromise of limitations the parts budget and the manufacturers policy allows.
Number and quality of x-over parts is a typical playing field for the guys with the red marker to save costs.
Part quality is very much overrated, once you remove the known junk, boutique parts don't improve anything noticeable.
Parts number and size is another chapter, the smallest PCB with the least components that still sounds acceptable wins in a commercial products. This is the area where a good developer can make huge improvements over the passive version. Be warned, this comes close to a complete new development!

Thanks for the insight, from what I understand if you have 2 setups with the same crossover points and slopes, one done with DSP and the other done passively, the one done via DSP will have lower distortion, more uniform phase response, and lower resistance

Would you say this is accurate?

 

[dfuller]

Thanks for the insight, from what I understand if you have 2 setups with the same crossover points and slopes, one done with DSP and the other done passively, the one done via DSP will have lower distortion, more uniform phase response, and lower resistance

Would you say this is accurate?

Not necessarily.

 

[Wolf11Man]

If you do the same curve

Thanks for the insight, from what I understand if you have 2 setups with the same crossover points and slopes, one done with DSP and the other done passively, the one done via DSP will have lower distortion, more uniform phase response, and lower resistance

Would you say this is accurate?

No, dififferent. In general the audible result will be better. The amp likes a simple impedance curve better than a complex crossover. Second, the tweeter and mid amps have to deliver far less current.