Dirac Live and bass management

[rcohen]

Is it better for the bass management to be done before or after Dirac processing?

 

I use JRiver for most stuff and play through a pre-amp via HDMI.

 

I can think of three options:

 

1) Use bass management in the pre-amp, after Dirac.

2) Use bass management in JRiver, before Dirac.

3) Use Windows 7 bass management, before Dirac.

[dallasjustice]

Your question is a lite too general. To me, the correct way to do 2CH with subs is to crossover R/L to at least 2 mono subs. I apply delay to each channel so that sound is arriving at listening position all together.

 

If you go this route and your "bass management" means crossover and delay, you MUST apply all of your bass management AFTER the DAP.

 

DIRAC will only apply the filter for the number of channels you tested with in DLCT. Even MCH DLCT doesn't create filters for crossover. If you try DLCT using crossover and delay into more than 2CH inside Jriver, DAP will only apply filters to 2CH; not all channels.

 

I tried this using an exasound MCH DAC using DLCT MCH version. Using a virtual cable, routed DLCT via Jriver. It doesn't work for the reason stated above. DIRAC really isn't designed for this purpose. I am now using a DEQX to apply crossover and delay in my 2CH system with a pair of mono subs. Once I got the delay and crossovers dialed in, I run DLCT for 2CH, and it works perfectly.

 

 

Is it better for the bass management to be done before or after Dirac processing?

 

I use JRiver for most stuff and play through a pre-amp via HDMI.

 

I can think of three options:

 

1) Use bass management in the pre-amp, after Dirac.

2) Use bass management in JRiver, before Dirac.

3) Use Windows 7 bass management, before Dirac.

[rcohen]

Your question is a lite too general. To me, the correct way to do 2CH with subs is to crossover R/L to at least 2 mono subs. I apply delay to each channel so that sound is arriving at listening position all together.

 

If you go this route and your "bass management" means crossover and delay, you MUST apply all of your bass management AFTER the DAP.

 

DIRAC will only apply the filter for the number of channels you tested with in DLCT. Even MCH DLCT doesn't create filters for crossover. If you try DLCT using crossover and delay into more than 2CH inside Jriver, DAP will only apply filters to 2CH; not all channels.

 

I tried this using an exasound MCH DAC using DLCT MCH version. Using a virtual cable, routed DLCT via Jriver. It doesn't work for the reason stated above. DIRAC really isn't designed for this purpose. I am now using a DEQX to apply crossover and delay in my 2CH system with a pair of mono subs. Once I got the delay and crossovers dialed in, I run DLCT for 2CH, and it works perfectly.

 

I am using Dirac full (7.1).

 

Currently, I am setting all my Windows 7 playback devices (Dirac and HDMI) to full range for all 7 speakers.

Then, it goes out to my pre-amp via HDMI.

From there, my pre-amp is doing delays to all speakers and subs, with all speakers set to small.

 

Incidentally, I have 4 subs, with the 2 at the front of the room using 1 sub out, and the 2 at the back of the room using the 2nd sub out. This allows me to time-align the front and back subs independently, but all subs are playing the same signal.

 

An alternative strategy would be to set my pre-amp so that all speakers are large, and do the bass management (crossover, not delay) in JRiver prior to the Dirac Processor. It would work. My question is which approach is better?

 

If doing the crossover prior to Dirac is best, another possibility would be to set the speakers to small in Windows, rather than JRiver. That is less flexible (and not sure about quality), but it would apply to all software, rather than just JRiver.

[rcohen]

Any more insight into this question?

 

Here's my playback chain:

JRiver Media Center -> Dirac Live Full -> Pre-amp

 

Currently, I am doing time alignment and bass management in my pre-amp and using DL to correct the combined response.

 

Alternately, I could do the bass management in JRiver, and use DL to correct the individual speaker channels and combined sub channel. Time alignment would still happen in the pre-amp.

 

I have heard conflicting stories on which is the better approach.

[3ll3d00d]

Does your current setup work well? By that I mean do you get a smooth transition from mains to sub with no obvious integration related response anomalies? If so, I don't know what you have to gain by changing anything.

 

FWIW I think miniDSP posted a good summary of the difference approaches on AVS - 8 - IN 8 - OUT HDMI Audio processor with Dirac Live® technology at CEDIA - Page 3 - AVS Forum

 

If you wanted to do bass management in jriver then I think you could try combining that with peq in jriver to attempt to fix any anomalies in the crossover range. I suspect this will be a fair bit more work though as you would need to iterate over Dirac then peq then Dirac to optimise it.

[rcohen]

Thanks for the link!

They both seemed to work fine, based on my brief experiments with it.

Someone recently suggested how to take REW measurements through JRiver, which helps.

 

I'm not sure what all to watch out for, but I imagine the main concern would be matching time/phase at the crossover frequency.

[rcohen]

Looking closer at that thread, it looks like a great discussion of this issue.

[dallasjustice]

Looking closer at that thread, it looks like a great discussion of this issue.

 

I know that there are a few "creative" ways to use DIRAC with bass management. IME, Acourate is designed to incorporate multiple channels in this way. It's a much better solution, IMO.

[rcohen]

I know that there are a few "creative" ways to use DIRAC with bass management. IME, Acourate is designed to incorporate multiple channels in this way. It's a much better solution, IMO.

Yeah, I need to experiment with Acourate and Audiolense XO one of these days, especially if Dirac doesn't add support for more channels!

[spdif-usb]

Yeah, I need to experiment with Acourate and Audiolense XO one of these days, especially if Dirac doesn't add support for more channels!

The Emotiva XMC-1 with Dirac Live supports both independently EQ'd stereo subs and independently EQ'd dual mono subs. If you must have independently EQ'd quad subs in an AV processor, expect to pay one heck of a lot more than a couple thousand smackaroonies (either that, or you can just crossover your fingers and wait for Emotiva to release the RMC-1).

[rcohen]

The Emotiva XMC-1 with Dirac Live supports both independently EQ'd stereo subs and independently EQ'd dual mono subs. If you must have independently EQ'd quad subs in an AV processor, expect to pay one heck of a lot more than a couple thousand smackaroonies (either that, or you can just crossover your fingers and wait for Emotiva to release the RMC-1).

I want extra channels for Atmos, not independent EQ on subs. Not sure why you'd want separate EQ on each sub.

[rcohen]

Does your current setup work well? By that I mean do you get a smooth transition from mains to sub with no obvious integration related response anomalies? If so, I don't know what you have to gain by changing anything.

 

FWIW I think miniDSP posted a good summary of the difference approaches on AVS - 8 - IN 8 - OUT HDMI Audio processor with Dirac Live® technology at CEDIA - Page 3 - AVS Forum

 

If you wanted to do bass management in jriver then I think you could try combining that with peq in jriver to attempt to fix any anomalies in the crossover range. I suspect this will be a fair bit more work though as you would need to iterate over Dirac then peq then Dirac to optimise it.

 

After reading that thread carefully, it sounds like each approach has its own potential land mines, and the best bet is to try both and measure to make sure you didn't hit one.

[Kal Rubinson]

The Emotiva XMC-1 with Dirac Live supports both independently EQ'd stereo subs and independently EQ'd dual mono subs.

I didn't know that had been released yet.

[spdif-usb]

I didn't know that had been released yet.

The XMC-1 itself, yes. The firmware update that will enable Dirac Live on it, no not yet.

[Kal Rubinson]

The XMC-1 itself, yes. The firmware update that will enable Dirac Live on it, no not yet.

Exactly.

[rcohen]

Do Dirac's filter lengths made a difference here?

 

I imagine that Dirac's subwoofer channel gets reduced bandwidth and longer (measured by time) filters than the full bandwidth channels.

 

If that's true, is it advantageous to do bass management before Dirac, in order to take advantage of the longer filters?

[3ll3d00d]

Do Dirac's filter lengths made a difference here?

 

I imagine that Dirac's subwoofer channel gets reduced bandwidth and longer (measured by time) filters than the full bandwidth channels.

 

If that's true, is it advantageous to do bass management before Dirac, in order to take advantage of the longer filters?

AIUI Dirac uses IIR for the low frequencies so I don't think filter length makes any difference. The filter itself is ~15-20ms long which equates to a frequency resolution of ~25Hz in FIR terms hence pretty useless for a sub.

[rcohen]

AIUI Dirac uses IIR for the low frequencies so I don't think filter length makes any difference. The filter itself is ~15-20ms long which equates to a frequency resolution of ~25Hz in FIR terms hence pretty useless for a sub.

To be more specific, my room is 30 ft (27 ms) long. If I put subs at the front and back and sit in the middle, will Dirac be able to cancel front/back reflections? Does doing bass management prior to Dirac come into play, since it sounds like my room is longer than the FIR filter?

[3ll3d00d]

To be more specific, my room is 30 ft (27 ms) long. If I put subs at the front and back and sit in the middle, will Dirac be able to cancel front/back reflections? Does doing bass management prior to Dirac come into play, since it sounds like my room is longer than the FIR filter?

Cancelling a modal resonance at ~18Hz is a job for an IIR notch filter, I don't see the length/resolution of the dirac filter as being relevant to that (as dirac AIUI employs both FIR and IIR when creating the final filter).

[dallasjustice]

To be more specific, my room is 30 ft (27 ms) long. If I put subs at the front and back and sit in the middle, will Dirac be able to cancel front/back reflections? Does doing bass management prior to Dirac come into play, since it sounds like my room is longer than the FIR filter?

 

This is exactly my setup except my room is 24' long. Filter length isn't related to delay. You will need RL and rear sub delays applied. DIRAC is not able to do that.

[rcohen]

Cancelling a modal resonance at ~18Hz is a job for an IIR notch filter, I don't see the length/resolution of the dirac filter as being relevant to that (as dirac AIUI employs both FIR and IIR when creating the final filter).

I wasn't sure if mirroring front and back subs and sitting in the middle changed that, since there is an opportunity to do a 1-time cancellation.

[rcohen]

This is exactly my setup except my room is 24' long. Filter length isn't related to delay. You will need RL and rear sub delays applied. DIRAC is not able to do that.

Why would I need delay if I am sitting in the middle?

(unless I was inverting the signal to the rear subs, which I haven't tried, yet)

[3ll3d00d]

I wasn't sure if mirroring front and back subs and sitting in the middle changed that, since there is an opportunity to do a 1-time cancellation.

OK I see what you mean (I think). If you are sitting at the 50% (acoustic) distance in a cuboid room then the output from those subs will completely cancel the 1st axial length mode so you don't need EQ in that case.

[dallasjustice]

Why would I need delay if I am sitting in the middle?

(unless I was inverting the signal to the rear subs, which I haven't tried, yet)

 

Think about flight time of each speaker to your head. First, you will need each sub on front and backwall. Second, your RL speakers will be out into the room from frontwall some distance. If you have your RL out into the room, the front sub will hit your head AFTER RL.

 

You would only need a single delay for RL if you sit exactly mid point between frontwall and backwall. Once you set this system up, you will discover that you don't want to sit at the front/back mid point because you won't like sitting so close to RL speakers in such a large room. Most likely, you will sit slightly closer to backwall than frontwall. You would then need slightly more delay applied to rear sub than RL.

[rcohen]

Think about flight time of each speaker to your head. First, you will need each sub on front and backwall. Second, your RL speakers will be out into the room from frontwall some distance. If you have your RL out into the room, the front sub will hit your head AFTER RL.

 

You would only need a single delay for RL if you sit exactly mid point between frontwall and backwall. Once you set this system up, you will discover that you don't want to sit at the front/back mid point because you won't like sitting so close to RL speakers in such a large room. Most likely, you will sit slightly closer to backwall than frontwall. You would then need slightly more delay applied to rear sub than RL.

 

Yeah, that's how I had it before. Just experimenting. I'm using my preamp to set delays on the LCRs and subs. Interestingly, I have the LRs in the corner, now. They have big coax horns, so they are highly directional and are doing surprisingly well in the corners. The stereo imaging is rocks solid until you walk up to a few feet away from the center.

 

it is a little bit strange having subs a wide soundstage, though. I'll have to give it some time to see if I keep it.

 

Previously, I had my subs in the corners, but I wanted to experiment with 1/4 3/4 placement and inverting the phase on the back pair to cancel out all the room modes (except height). Then I started wondering if I could accomplish the same thing with double the power by sitting in the center and matching the phase and delay on my subs, then letting room eq cancel the reflections.

 

My room has very reflective walls, and a setup that actually removes the energy from the room, rather than compensating at the listening position might be a good thing.