New: Hang Loose DSP Processor

[Kal Rubinson]

I'm hoping to still make the co-located stereo subs be dominant aurally.

??? Why would you co-locate stereo subs?

 

[john61ct]

Not the subs co-located with each other! :-D

I meant in the context of reinforcing the main front L/R pair with separate enclosures to get Full Range from those locations.

As to whether the experience is better feeding stereo rather than mono to the deepest-frequency enclosure, that will be objectively tested.

Does 'envelopment' exist? Can it be measured?

I hear trusted people use envelopment a lot. Can someone tell me what it is? How would I create more of it, assuming it's a good thing.

www.audiosciencereview.com

 

[levimax]

I thought FR EQ could only go so far with room effects?

I was hoping to tame room modes more with free placement of multiple mono subs, iterating REW testing. Do those need to be BRuTeS (Bone Rattling fUll-size TruE Sub)? or would smaller ones be enough?

I'm hoping to still make the co-located stereo subs be dominant aurally.

I'm thinking, shortening relative delays between my main front L/R and the various subs would be the most likely "need DSP" trigger.

The most effective use of DSP is to knock down room mode peaks. One of it's biggest limitations is trying to fill room mode dips. If you want minimum DSP as you mentioned I would just use speaker placement and room mode peak attenuation at least to start. Narrow dips are not very audible even though they look bad in REW. Look at response with psychoacoustic smoothing to get an idea how dips actually sound. Regarding subs most would say remote placement and sum to mono for best FR smoothness. I prefer tightly colocated stereo subs for stereo bass, better time domain performance, higher crossover / no location of sub issues, and less DSP. Probably best if you tried both ways and decide for your self.

 

[john61ct]

Good stuff thanks!

I prefer tightly colocated stereo subs for stereo bass, better time domain performance, higher crossover / no location of sub issues

That's my starting point too.

I'd happily sacrifice "smoothness" away from MLP for better precision / clarity / soundstage, envelopment, "speakers disappearing" for just me.

Watching films as a family, hopefully easy to switch via "profile settings" or easy re-routing the signal paths e.g BobWire

> Probably best if you tried both ways and decide for yourself.
Yes I plan to test all such factors, and not just with tools like REW. Blind ABX, multiple runs getting F&F involved.

 

[welwynnick]

Yes, of course it's possible. But whether it's a good idea is another story.
First, the rising group delay at low freqs is due to the minimum-phase nature of the loudspeaker. If you study that GD plot closely and compare it to the frequency response, you will see that the reason the GD is rising is because the SPL is falling. In order to fix this, you need to equalize that speaker to flat down to 20Hz. Now, ANY speaker can be equalized to flat down to 20Hz, even tiny bookshelf speakers the size of a drinking mug. The only problem is that you lose 50-60dB of headroom and you get a tonne of bass distortion. Your speakers will be flat to 20Hz, but they will also be so quiet that you can't hear them.

I read all that and found myself nodding in agreement without thinking about it, but there are two points:
(1) The midrange is most important to me, and I'm not worried about LF phase EQ too much, as I plan to tackle that with a form of infinite baffle subs:

Infinite Baffle Double Bass Array

Wave Forming, Dirac ART and Double Bass Arrays (DBA) are a revolution in bass management that goes beyond equalisation. They use “subtraction subwoofers” to take bass energy away, as well as adding it. So I wondered how much energy it takes to take bass away? To keep the box size reasonable...

www.audiosciencereview.com

(2) I know that group delay rises at the LF roll-off. I tried to describe the mechanism in post #66 and in the thread above. However the piece about equalizing a speaker flat down to 20Hz made me wake up. AIUI that's how active subs work - put a large driver and a powerful amplifier in a relatively small box, and equalize the LF extension. The driver will have a system resonance at a higher frequency than it would in a "normal" passive speaker, and the extension is achieved with EQ and a lot of power. Those active subs are part of what I'm trying to avoid.

(Unless you use an HLP) that EQ doesn't change the driver's LF roll-off mechanism. The resistance to motion is still largely from the acoustic suspension, which means the displacement is closely in phase with the applied signal. The cone acceleration (and SPL) is the second differential of that, and very much out of phase. That's the same situation as below the natural LF roll-off without any EQ. It's just that the level is higher. Therefore I don't think the EQ will reign in the GD.

 

[levimax]

Those active subs are part of what I'm trying to avoid.

The other issue with small sealed subs is that due to the heavy bass boost they also need a "protective high pass" right around 20 Hz to stop them from self destructing in case they tried to reproduce a "heavily boosted" signal below 20 Hz. A steep high pass filter like this causes large time domain issues which are in addition to the other issues you already mentioned.

 

[john61ct]

Can't such protective rumble HPF be pushed a bit further down to infra Hz so the impact is inaudible?

Also, would what you state also be true for an effectively ported bandpass + passive radiator design like @Wolf 's "Overdrive10 / Kilauea" design?

I built a sub - the Overdrive10

Over a year ago I posted here that was searching for a design for a DIY sub. There were several great suggestions, but I ended up going with the Overdrive10 designed by @Wolf. It's a bandpass design with a 10-inch Dayton driver and a passive radiator. I bought the hardware a year ago and finally...

www.audiosciencereview.com

 

[levimax]

Can't such protective rumble HPF be pushed a bit further down to infra Hz so the impact is inaudible?

Also, would what you state also be true for an effectively ported bandpass + passive radiator design like @Wolf 's "Overdrive10 / Kilauea" design?

I built a sub - the Overdrive10

Over a year ago I posted here that was searching for a design for a DIY sub. There were several great suggestions, but I ended up going with the Overdrive10 designed by @Wolf. It's a bandpass design with a 10-inch Dayton driver and a passive radiator. I bought the hardware a year ago and finally...

www.audiosciencereview.com

It is not audible as a filter rather it causes time domain issues like group delay and ringing..... the audibility of which is another questions but I don't think anyone wants more group delay or more ringing as compared to less.

 

[3ll3d00d]

(2) I know that group delay rises at the LF roll-off. I tried to describe the mechanism in post #66 and in the thread above. However the piece about equalizing a speaker flat down to 20Hz made me wake up. AIUI that's how active subs work - put a large driver and a powerful amplifier in a relatively small box, and equalize the LF extension. The driver will have a system resonance at a higher frequency than it would in a "normal" passive speaker, and the extension is achieved with EQ and a lot of power. Those active subs are part of what I'm trying to avoid.

we're going off topic but this has nothing to do with active or passive, it's just a function of enclosure size. Sealed subs that have a native q=0.7 response can be infeasibly large so it's a normal thing in this day and age to apply minimum phase eq to bring the response towards something that can neatly mesh with room gain. The issue here is that you need to drive those subs harder to produce the same output which means they will compress/distort sooner. The solution to that is to add more subs or turn it down.

A steep high pass filter like this causes large time domain issues which are in addition to the other issues you already mentioned.

if you drive any sub hard enough, it's enough go to melt or it needs some other form of protection. Obvious solution here is DIY then you can take full responsibility for melting your own sub when you blast it into oblivion

 

[levimax]

if you drive any sub hard enough, it's enough go to melt or it needs some other form of protection. Obvious solution here is DIY then you can take full responsibility for melting your own sub when you blast it into oblivion

I guess I should qualify that we were speaking about small sealed subs. If it wasn't for the DSP bass boost you would not ordinarily need a high pass protection filter on a sealed sub because the air pressure would help prevent over excursion. You are right of course that enough power will melt any sub so that has to be part of the design as well.

 

[welwynnick]

we're going off topic but this has nothing to do with active or passive, it's just a function of enclosure size. Sealed subs that have a native q=0.7 response can be infeasibly large so it's a normal thing in this day and age to apply minimum phase eq to bring the response towards something that can neatly mesh with room gain. The issue here is that you need to drive those subs harder to produce the same output which means they will compress/distort sooner. The solution to that is to add more subs or turn it down.

Assuming you have the same driver and the same motor, but a smaller box and a bigger amplifier, why would that increase the distortion?

 

[john61ct]

Not necessarily audibly significant

but pushing closer toward the power handling limits to get high SPL maybe deeper extension from a small box

can brings the driver closer to where performance is no longer linear.

My limit is an external 13"² HxW maybe 15" D, so 1.5 cuft internal? Bit of bass boost needed to get SPL below 30Hz, try to keep the LF sub group's box count low...

 

[welwynnick]

can brings the driver closer to where performance is no longer linear.

If the small unit is equalized for the same output, the driver motion and hence the non-linearity should be similar.

To answer my own question - with a smaller unit, I think there are several factors at play.

• The change in internal pressure will be higher, and the difference between compression and expansion changes will be slightly higher.
• The acoustic suspension forces will be greater , and mechanical suspension forces the same, so the latter will become less important.
• The motor forces will necessarily be higher, so motor non-linearity will become slightly more important.
• The motor will get hotter.

I'm not sure where that leaves us overall. The answer is probably "it depends".
As ever my interest in all of this is trying to achieve linear phase response over a wide frequency range, which is where external DSP would come in.
However I'll steer away from subs with ports, ABRs, coupled cavities and heavy EQ, rather than relying on the HLP to fix everything.
In other words, try to get the system fundamentals right first, instead of lashing anything together and then trying to fix everything afterwards.

 

[3ll3d00d]

Assuming you have the same driver and the same motor, but a smaller box and a bigger amplifier, why would that increase the distortion?

More power, more heat = Re rises which can impact Bl and hence impact the response (theoretically at least). You are going to have to run a sub pretty hard for this to be an issue but if you typically listen to that sort of content at spirited levels then it can be.

However I'll steer away from subs with ports, ABRs, coupled cavities and heavy EQ, rather than relying on the HLP to fix everything.
In other words, try to get the system fundamentals right first, instead of lashing anything together and then trying to fix everything afterwards.

In the end, you need to specify the clean output required to what frequency and this combined with the available space/budget tends to dictate the budget required. In this context dsp *is a* system fundamental not some hacked in afterthought.

 

[john61ct]

However I'll steer away from subs with ports, ABRs, coupled cavities and heavy EQ, rather than relying on the HLP to fix everything.

I agree wrt DSP, but a Bandpass enclosures with PRs relieves the issues you mentioned - gets higher deeper SPL from a small box. Yes high power is required.

But it only takes a little "bass boost" at the bottom to extend even deeper, without adding audible distortion. Does not even "need" to be DSP

 

[toobs]

Is it possible to output multiple channels via HDMI?

 

[toobs]

Also, if I may ask:what fpga board are you using? I am trying to figure out how to implement Volterra neural network predistortion.It sounds like you are underutilizing the FPGA and since you have the io figured out already, I might be able to implement the predistortion code in the FPGA.

 

[toobs]

Just read...it's a Spartan 6. Nevermind.

 

[welwynnick]

Is it possible to output multiple channels via HDMI?

While this is a stereo unit, please note for USB, the unit provides 8 channels of I/O. So it is possible, for example, to select the stereo TOSLINK input (or any input) and then 8 channels of USB output to be used for a stereo 3-way or 4 way digital XO system. Or up to 7.1 multichannel surround. Also note HDMI up to 8-channel 192kHz/24bit PCM and 6-channel DSD64 audio extraction.

 

[toobs]

Thank you for the kind reply. I have a few more questions, if you will:

1) the filtering is done by the dsp so that the raspberry pis could be free to run VSTs?
2) when will the 16 channel with aes67 be out?
3) 32 channel?