time delay measurements are always different

[azookey]

I'm trying to measure the time delays of 2 subs using the acoustic references. I'm following the various guides online, specifically using miniDSP. Every time I measure a single sub, i am getting rather large differences. I can make 3 measurements in a row, one right after another, and I'm getting results as follows:

m1 -1.49
m2 -0.84
m3 -3.32


Also, maybe unrelated question, should i be measuring delay before or after I set crossovers in the miniDSP plugin?

Thanks.

 

[jtalden]

Are you using a main speaker with a tweeter as the reference channel? Please post the mdat file of the repeated measurements. That may provide some hints as to where the problem is.

 

[azookey]

Speakers are Zu Omen's. They have a tweeter, but it doesn't kick in until rather high as the main driver is considered full range. Mdat attached with measurements of mains, plus 3 measurements of each sub.

Thanks

 

[jtalden]

There is no difference in the actual delays of the repeated SW measurements. Below find 2 impulse overlay charts of the 3 measurements of each SW. The REW calculated delay estimates reported in the measurement notes vary a little when measuring SWs full range. These values are not used when using the REW alignment tool. If you are following another alignment procedure then using an average of the 3 measurement would probably be the best value to use.

 

[azookey]

Thanks for the feedback.

I was following this guide which just looks at the "system delay" of each measurement in the info window.

edit: I can't seem to paste a link, but its "Measuring time delay for home theater with the UMIK-1 and REW" from the minidsp site.

Thanks

 

[azookey]

So, in terms of figuring the delay, if my mains are at 0ms, which peak/dip do I use?

The first peak at approx. 13ms, or the big dip at 24ms? Or am I just doing this wrong


THanks

 

[jtalden]

There are 2 delays to choose. The SW to to SW delay and the SWs to mains delay. I cannot speak to what the MiniDSP setup procedure calls for as I have never investigated that method, but it is likely to work out just fine. Someone else will need to help clarify what to use.

I use the REW alignment tool and will go ahead and provide timing for both the SW to SW delay and for the SWs to mains delay. You can then compare it to the results of MiniDSP method if you want. There are usually several different settings that provide good results. The objective is a smooth SPL response in the crossover range that will allow for a reasonable amount of EQ. If different methods result different delay timings, but both meet the objective then it would be hard to argue that one solution is better than the other.

Analysis to follow.

 

[John Mulcahy]

The subwoofer measurements have very poor signal to noise ratios, it is the noise which is causing the variation in the timing figure. Try making the subwoofer measurements over a narrower measurement range, with an end frequency of 200 Hz, for example, rather than 20 kHz. That will improve their signal/noise ratio.

 

[jtalden]

azookey, Your current delay timing for the SW to SW and for the SWs to L,R are very good as they were measured. There is no need for changes. I looked at several changes to smooth the SPL as an option in case you want to experiment with other good timing settings. I settled on one that may be interesting to try. This one and the current one will both benefit a lot from from PEQ.

Optional Setup:

The only change here (last 2 columns) is to invert the polarity of the 2 SWs and reduce the delay of SW-R by 4.5 ms. If all the current delays are at 0 ms, the delay of the other 3 speakers can be increased by 4.5 ms instead. It's the relative delay and polarity between all speakers that is important. This optional setup intentionally mistimes the 2 SWs a little to help smooth the SPL and to help support the 80 Hz area a little better. With appropriate EQ to each setup the difference in SPL should be very small. You may, or may not, hear a sound quality difference. The optional setup (blue trace) vs the Current setup is shown in the chart below.

 

[azookey]

Wow, thank you so much for this! I will definitely try these measurements out and report back.

Could you link to a tutorial or page where I can read about what you did?

Thanks.

 

[jtalden]

REW help has the alignment tool controls identified.
Here is a thread that has some discussion.
There are other threads here that may also help. Just search for threads that I contributed to.
I am not aware of any other sources, but a web search may find something.
Attached is the file I created using your original data for this analysis.