Group Delay Questions and Analysis

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thx XRK. Looks like a lot of work.

I ordered the miniDSP 2x4. The plug in I ordered will give me the following:

dsp c.GIF

Should arrive tomorrow or Monday. Ill see how close to your sim plots I can achieve.

View attachment 483042

Here is what I get with all subs running and all XO points set at 180hz (their highest setting)

The 40hz problem, as previously anticipated, maybe room mode related. My room is 26.25' long which corresponds to 42.9hz. Not positive this is related. But a theory.
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Thx XRK. Looks like a lot of work.

I ordered the miniDSP 2x4. The plug in I ordered will give me the following:

View attachment 483041

Should arrive tomorrow or Monday. Ill see how close to your sim plots I can achieve.

View attachment 483042

Here is what I get with all subs running and all XO points set at 180hz (their highest setting)

The 40hz problem, as previously anticipated, maybe room mode related. My room is 26.25' long which corresponds to 42.9hz. Not positive this is related. But a theory.

The plugins are cheap ($10) and you can get them instantly via download. I personally like parametric EQ more than graphical EQ. Much more fine control and lets you implement Linkwitz transform with the sealed.

It is a lot of work, but I did say I would get you some models to see what is the best case scenario "low hanging fruit" to grab. Hope it's helpful. If you know what your high pass filter freq and slopes are, I can put them in for more accurate predicts. Similarly, cabinet damping affects Q is best measured with an impedance sweep. If you have that info, we really can get this dialed in. But a ballpark of 16ms to 18ms seems like a good ultimate number to work towards.
 
Last edited:
The plugins are cheap ($10) and you can get them instantly via download. I personally like parametric EQ more than graphical EQ. Much more fine control and lets you implement Linkwitz transform with the sealed.

It is a lot of work, but I did say I would get you some models to see what is the best case scenario "low hanging fruit" to grab. Hope it's helpful. If you know what your high pass filter freq and slopes are, I can put them in for more accurate predicts. Similarly, cabinet damping affects Q is best measured with an impedance sweep. If you have that info, we really can get this dialed in. But a ballpark of 16ms to 18ms seems like a good ultimate number to work towards.

I like PEQ better than GEQ also, but the plugins that offer PEQ only offer a max delay of 7.5ms.
 
I thought it was a hardware limitation of the 2x4 miniDSP. Not sure if the GEQ will actually have the large delays shown. Have you asked them if it does indeed do that ? I have and was told only way is to use miniSHARC.

The Center/Rear plugin is the only one that allows 27+4ms of delay. I am going to use the Rear Center channel as a sub channel.
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Can you show me the link where it says this plugin with 2x4 miniDSP will allow 24ms delay. I wanted large delays before and went through many emails with miniDSP tech support. And time and time again they said anything beyond 7ms was not possible with this unit. I only needed 9ms. Even considered daisy chaining out from one channel back into another to get 14ms. Would have 2 stages of ADC/DAC though. Just saying that, unless the hardware has changed, which I doubt, that plugin may be for the 2x8 miniDSP. Since you make it sound like it is multi channel surround it is at least 6 channels - so probably not 2x4 which cannot do 5.1 surround.
 
Can you show me the link where it says this plugin with 2x4 miniDSP will allow 24ms delay. I wanted large delays before and went through many emails with miniDSP tech support. And time and time again they said anything beyond 7ms was not possible with this unit. I only needed 9ms. Even considered daisy chaining out from one channel back into another to get 14ms. Would have 2 stages of ADC/DAC though. Just saying that, unless the hardware has changed, which I doubt, that plugin may be for the 2x8 miniDSP. Since you make it sound like it is multi channel surround it is at least 6 channels - so probably not 2x4 which cannot do 5.1 surround.

http://www.minidsp.com/images/documents/Product Brief-Rear-Center channel plug-in.pdf
 
Founder of XSA-Labs
Joined 2012
Paid Member
That does indeed look like its for a 2x4 miniDSP. Great! Wonder why they never told me about this and why they insisted it was a hardware limit ? The thing I need to delay is the tweeter and mid relative to the sub. This is for the sub but gives me hope that long delays are possible with the right software and it's not a hardware issue.
 
That does indeed look like its for a 2x4 miniDSP. Great! Wonder why they never told me about this and why they insisted it was a hardware limit ? The thing I need to delay is the tweeter and mid relative to the sub. This is for the sub but gives me hope that long delays are possible with the right software and it's not a hardware issue.

I read somewhere it was a memory and processing power issue. The Center/Rear plugin does less in other areas, allowing for more in the delay. At least, that was my surface read of the issue.
 
Here are some quick sims of the idealized system of a bass reflex and two sealed boxes using the T/S parameters (Acoustic Concepts AC10R) you provided along with the box alignments. I assumed a slightly overdamped system to extract the best transient response. Also assumed use of a 2nd order Butterworth high pass at 30Hz for the bass reflex and at 22Hz for the sealed boxes in order to keep the xmax under 7mm at 24 volts.

If you still have the data, could you post the xmax charts for the sealed and BR without any HPF?

Also, remember the BR has 22hz port tuning. ( I could tune a bit higher if that is advantageous)

Assume 75W for each sealed, 250w for the dual woofer BR.

Thz.
 
Last edited:
miniDSP results leaving everything in the same place, this is the best I have so far come up with.

05 16 GD DSP.jpg

Was able to make progress in the 36-41hz range. 40hz dropped from 25ms to 16.3ms.

05 16 15 FR DSP.jpg

Much smoother here :)

Basically, it worked out best continuing to use the plate amp LPF rather than those of the DSP. Polarity still reversed. Delay 0.5ms.

GEQ :

20hz: -1db
25hz: +1db
31.5hz: -1db
40hz: -3db
50hz: -1db
63hz: -1db
 
Last edited:
I think that was because I had the error of only a single driver. The new theoretical best is lower and quite a good number: 6.25ms at 40hz and 10ms at 30Hz

Well, you have disproved that I live in the Matrix ;)

On the other hand, I am happy with what $100 got me, especially in the FR extension. Cut the 40hz GD in half.

Theory and application are very different things. Especially since I am measuring at the LP (taking the room into account).

Just for grins, I Googled as many GD's as I could find. Most were horrible. Some had 20 or 30ms of GD at 100hz or higher :eek:
 
Last edited:
What did you use the miniDSP to do exactly that brought the GD down? Did it take over the sub woofer crossover?

The biggest single thing was the ability to apply the GEQ adjustments. The marriage around 40hz from the main sub to the mains wasnt ideal. There was clear overlap of their FR's in the 40hz region.

For whatever reason, the plate amps 4th order XO worked better than any of the options on the DSP. So I bypassed that feature on the DSP.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.