Something in my timing reference settings has changed???

[AustinJerry]

I have taken measurements last week and again today and am looking at the Acoustic timing reference information. Nothing has changed--the mic is in the same spot, same speaker used as the timing reference, room correction off, etc. Yet the delay values are significantly different, 11.6ms to 27.8ms. One measurement uses the description "using IR start time relative to Acoustic reference" and the other measurement says "using estimated IR delay relative to Acoustic reference". I suspect this is a clue to why the measurements are different, but I can't find a setting that affects these descriptions. Any advice would be appreciated.

 

[AustinJerry]

Removed.

 

[John Mulcahy]

There is no setting which affects that, it was changed in 5.19 beta 10:

• The delay figure with acoustic timing reference now uses the Estimate IR Delay process (cross correlation with minimum phase response) for better results with measurements that do not have HF content (mids or woofers)

It also fixes a problem with tweeter measurements where the secondary peak could be slightly larger than the primary, giving an incorrect timing figure.

 

[AustinJerry]

Thank you for your reply, John. Can you help me interpret the measurement results as they pertain to the time alignment among four subs? My two front subs are equidistant from the MLP, as are my two rear subs, but the front subs are approximately eight feet further away. I am currently using a 7.1ms delay on the rear subs to time align them with the front subs. The measurements for the front subs show 30.1 and 30.4ms delay relative to the acoustic timing reference speaker, while the two rear subs are showing 11.3 and 11.4ms delays. How can these measurements be used to assess the accuracy of my current time alignment measurements?

 

[John Mulcahy]

I'd need to see the mdat file for the measurements to comment, but I'm away on holiday with the grandkids at the moment so won't be able to look at any measurements until the weekend.

 

[AustinJerry]

Thank you, John. Here is a link to the MDAT file, which contains four individual measurements of the subs. The labels are FL, FR, RR, and RL for front left, front right, rear right, and rear left, respectively.

https://www.dropbox.com/s/8zxhmre46n4phll/AustinJerry 4Subs.mdat?dl=0

I am interested in understanding how to interpret the Delay readings as it pertains to the sub-to-sub time alignment.

 

[jtalden]

Jerry,
Comments/observations aside from your question to John:

Thanks for posting this file. I have been wanting to see a file containing before and after measurements of Dirac applied to a setup. It would be interesting also to see the before and after of the mains, and also the before and after of the timing accuracy of the SWs to mains. I have seen plenty of charts posted that all look great, but this is the first measurement data and it tells the bigger story that there is no indication of any downside/tradeoffs in any of the other charts.

Your initial SW timing is very accurate for all 4 SWs. I have timed the 2 SWs in my 2 setups also although accurate timing is not necessarily a requirement to achieve a favorable combined response (as shown by Earl Geddes for example). Your individual SWs were inverted polarity initially. Dirac Inverted the polarity (now positive polarity) and reduced the SWs delay by about 4.5 ms. The resulting Dirac SPL/Phase/GD and other charts are all amazing good.

The pre Dirac SPL/phase/modal responses here was already very good. Thus, this example does not demonstrate how well Dirac handles other 1 or 2 SWs setups positioned poorly with several strong room modes. That would make the job more difficult. The charts I have seen suggests it does that well too.

I enjoy learning how to achieve very good results using DIY methods and am reluctant to automate for that reason, but it is amazing what Dirac can achieve at the LP. I will have to figure out how to experience Dirac myself at some point. Thanks again for posting the file.

 

[AustinJerry]

Jerry,
Comments/observations aside from your question to John:

Thanks for posting this file. I have been wanting to see a file containing before and after measurements of Dirac applied to a setup. It would be interesting also to see the before and after of the mains, and also the before and after of the timing accuracy of the SWs to mains. I have seen plenty of charts posted that all look great, but this is the first measurement data and it tells the bigger story that there is no indication of any downside/tradeoffs in any of the other charts.

Your initial SW timing is very accurate for all 4 SWs. I have timed the 2 SWs in my 2 setups also although accurate timing is not necessarily a requirement to achieve a favorable combined response (as shown by Earl Geddes for example). Your individual SWs were inverted polarity initially. Dirac Inverted the polarity (now positive polarity) and reduced the SWs delay by about 4.5 ms. The resulting Dirac SPL/Phase/GD and other charts are all amazing good.

The pre Dirac SPL/phase/modal responses here was already very good. Thus, this example does not demonstrate how well Dirac handles other 1 or 2 SWs setups positioned poorly with several strong room modes. That would make the job more difficult. The charts I have seen suggests it does that well too.

I enjoy learning how to achieve very good results using DIY methods and am reluctant to automate for that reason, but it is amazing what Dirac can achieve at the LP. I will have to figure out how to experience Dirac myself at some point. Thanks again for posting the file.

Thank you for your feedback and kind words. I am a big fan of Dirac Live, having converted from Audyssey several years ago.

Adding a little information as background to my earlier posts: I use a MiniDSP balanced 2x4 to consolidate the four sub signals into a single input on a MiniDSP 88A, which performs the Dirac Live calibration. The 2x4 is where I am applying delays to the individual sub channels for time alignment. So, Dirac sees a single sub input which has already been time-aligned. Like most automated room correction systems, Dirac does not optimize the alignment between the sub signal and the mains. I do that manually by conducting the well-known "sub distance tweak" procedure.

If you are interested in how I am using the 2x4, it is documented here:

https://www.dropbox.com/s/w7jmpozner6muzw/MiniDSP 2x4 Set-Up Guide Latest Version.pdf?dl=0

 

[andyc56]

One thing that's interesting about Jerry's data is looking at the excess group delay of each sub individually, then the four subs at once (GD button, Controls, Generate minimum phase). The two front subs each have significant deviations from flat excess group delay, so there's some non-minimum-phase action going on with them. The two rear ones have nearly constant excess group delay, so they're minimum-phase (neglecting "time of flight" delay). Then, when the four subs are all combined, the result also has flat excess group delay! This shows that using multiple subs can, under the right circumstances, "fix up" non-minimum-phase behavior of at least some of the individual subs. It's difficult to imagine how one could improve on this performance. Nice work.

 

[AustinJerry]

Nice to hear from you, Andy, and thanks for the kind words!

 

[John Mulcahy]

Hi Jerry, thanks for the file. I've made a test version that should better handle subwoofer measurements, I'll PM you a link to the installer to see if that gives better delay figures.

 

[AustinJerry]

John, I have re-measured the subs with the REW test version. The results are more in line with what I expect to see. Here are the results:

The numbers in the left column are what REW measures as the delay relative to the acoustic timing reference (the left speaker). Now comes what I assume can be done: If I calculate the delta from the right front sub (the furthest sub), I can then add this delta to the delay settings for each sub channel to reach "perfect" alignment. After making the adjustments and re-measuring, the new delays are indeed spot-on. Is this approach a valid one?

Here is a link to the new measurement file:

https://www.dropbox.com/s/g4jj2kgsokp0tou/Sub measurements with new REW beta.mdat?dl=0

 

[John Mulcahy]

Looks good and agree with the approach. As an aside, the sub responses look to be inverted, most evident in the combined response. Dirac seems to take care of that though.

 

[AustinJerry]

Yes, you are correct. The front two subs and the rear two subs have opposite polarity. Previously, I had flipped the polarity on the front two subs, not realizing that I should have corrected the rear two subs. I had scheduled a fresh Dirac calibration today after fine-tuning the sub alignments, so I took the opportunity to correct my previous mistake. The four subs now have the same polarity and Dirac no longer needs to make the correction. And the bass response sounds fantastic!

Thanks again for your help.