Home Theater Forum and Systems banner
1 - 18 of 18 Posts

·
Registered
Joined
·
22 Posts
Discussion Starter · #1 ·
I use a DEQX to PEQ my front two speakers and to perform crossover duties between my L/R mains and subs. I use REQ to determine PEQ settings.

The extra digital processing inserts a LOT of incremental time delays (could be 38ms) for my fronts and subs that all the rest of my speakers don't incur (a Center and four surrounds don't go through those delays.)

I'm wondering if I can use REQ to time align my Center and surrounds with the mains, and, to confirm time alignment between my subs and L/R mains? I figured I could take a measurement of each speaker, individually, and use the time differentials measured to compute effective distance differentials.

As I understand it the place to look is the time vs. energy chart. When I eyeball charts I see a "steady decline" from the initial peak/impulse down to around 3ms, then the chart "levels off" from a steep vertical to a more horizontal slope, and, begins to "wiggle" or "modulate" up and down quite a bit.

I assume the point in time at which that transition occurs is the "time gap"? ... that if I calculated that point for each speaker then the differences would be the effective distance differentials for each speaker (eg, if one speaker were at 6ms and another at 2ms then the first speaker was 4ms further away than the second speaker?

A related question. When run full range my surrounds have very little energy below 60hz, so I suppose to get this right I'd have to begin the measurement of all channels at a frequency high enough that all speakers would be "active"? That is, I assume the time vs. energy chart "begins" at the low frequency beginning of the sweeps?
 

·
Registered
Joined
·
7,514 Posts
I'm wondering if I can use REQ to time align my Center and surrounds with the mains, and, to confirm time alignment between my subs and L/R mains? I figured I could take a measurement of each speaker, individually, and use the time differentials measured to compute effective distance differentials.
Yeah, you can use the All-Energy-Time graph to determine speaker distance, but it's a bit tricky I find.

You can't use absolute time since the graph uses relative time from 0ms. So you need to have all the speakers on at the same time, or at least reference each one to the sub in seperate measures.. So you could align the mains to the sub and then the center to the sub and so on.

You only need to look at the first 15ms or so since the sound travels at about a msec per foot.

To identify the peaks as to which speaker it is (and it is hard to see, given reflections etc) you can adjust the time distance on your receiver and see which peak moves. The sub peak is fairly wide and the mains will be narrower.

Anyway, it works, as I've done it, but it takes some getting use to....

brucek
 

·
Registered
Joined
·
7,514 Posts
The only way I know other than what I described is with simple measuring of distance and subtract the delay of processing. A delay of 38msec is quite large - are you sure it's that much? Even if you mastered the Energy-Time graphs (which you should indeed give a try to), how do you propose to get 38msec of time difference from a receiver time trim. Do receivers offer that large of a value for each speaker?

brucek
 

·
Registered
Joined
·
509 Posts
hi brucemck

this is something I too have wanted to do, and for the life of me I can't work out how to do it. I've stared and stared at all the different options, but could make nothing of it.

I too use a deqx, but have gone a bit past that and have added many more channels, and it's now something like a six way, all thru the deqx but using some extra behringer devices.

I'm running three seperate sets of subs, and it was for this I wanted to set the delays (as they are different distances from the LP, and I wanted accurate delays on them to get the arrival time of the bass the same).

Bruce K gave a very similar answer to his above, but I was unable to work it out (I'm in no way anywhere near his league though, so if he 'has a bit of trouble' with it, it's way beyond my ken).

If you have success please report back!! I tried also using WINMLS for this exact reason, but my computing skills let me down there. According to Lars it is a simple matter to do time aligning with it, on some proviso of 'setting up the souncard correctly'. I didn't get that far, as I say my computing nous leaves a bit to be desired.

It is not a free program however.

Some where along the line, I indirectly asked Doug Plumb about whether his program had this capability, he didn't respond. In all likelihood he didn't see it, even tho it was in a thread he had participated in. I would not like to ask or start a direct thread on that in the REW forum!! ha ha, maybe a PM to him might settle the matter.

Again, that is not a free program.

have no other thoughts, but please let us know how you go.
 

·
Registered
Joined
·
22 Posts
Discussion Starter · #7 ·
TerryJ -- here's what I worked out.

I used the DEQX's room measurement functionality to measure the delay time for each channel. It's fairly easy to see on the impulse response graph and on the step response graph in the viewer.

Based on that, it was obvious that my left was .1ms quicker than my right.

It was also obvious, although less so, that there was a 2ms to 4ms difference between the subs and the mains. Based on that I created four profiles, each of which differed b 1ms in the delay on the fronts relative to the sub. I then tested each of those using REQ, and observed the frequency response around the crossover point. There was a 15db swing between 0ms and 4ms, with 4 being flattest +/- 30hz either side of a 100hz 60db/octave crossover.

BruceK -- my Meridian offers that much correction. I can get 20' out of the distance settings, and the rest out of the ms settings for each individual surround processing option. The DSP filter in the DEQX is 24.8ms long. The A/D and D/A (it's inserted in between my processor and my amp) adds 6.9ms to that.

My issue now is that there's a bit, just noticeable, of lip sync delay that I've got to figure out how to get rid of. When listening to music it's not an issue, so now I'm off to find some sort of video processing delay device.
 

·
Registered
Joined
·
438 Posts
So you could align the mains to the sub and then the center to the sub and so on.
That doesn't work. The sub is bandwidth limited so the peak in the Energy Time Curve is going to be naturally delayed....aligning the peak of the other speakers to that of the sub will delay all of the other speakers for too long, which is going to introduce phase cancellation in the crossover passband.

The guys over at Syn-Aud-Con had a great article describing this limitation of the ETC. I'll see if I can't find it and post it. The nice thing about it is that it provided a way around it.

Since it seems you're more worried about the delay through the crossover, a very easy way to measure delay would be to measure your crossover directly. Download Audacity, which is a free multi-channel recording tool (or pick any stereo recording device of your choice). Then what you do is create a sound sample of a square wave, or click track or whatever....something that has a vertical rise at the beginning. Burn this sound to a CD and play it through your system. On the left channel of your recording device, record the direct output from the CD. On the right channel of your recording device, record the direct output from the crossover. You will first see your sound show up on the left channel and then later on the right channel. The time readout of your software will be able to give you the time offset. It'll probably take you 15 minutes to get everything set up and ready to test, and then just a few seconds for every channel of your crossover.
 

·
Registered
Joined
·
509 Posts
That doesn't work. The sub is bandwidth limited so the peak in the Energy Time Curve is going to be naturally delayed....aligning the peak of the other speakers to that of the sub will delay all of the other speakers for too long, which is going to introduce phase cancellation in the crossover passband.

The guys over at Syn-Aud-Con had a great article describing this limitation of the ETC. I'll see if I can't find it and post it. The nice thing about it is that it provided a way around it.

Since it seems you're more worried about the delay through the crossover, a very easy way to measure delay would be to measure your crossover directly. Download Audacity, which is a free multi-channel recording tool (or pick any stereo recording device of your choice). Then what you do is create a sound sample of a square wave, or click track or whatever....something that has a vertical rise at the beginning. Burn this sound to a CD and play it through your system. On the left channel of your recording device, record the direct output from the CD. On the right channel of your recording device, record the direct output from the crossover. You will first see your sound show up on the left channel and then later on the right channel. The time readout of your software will be able to give you the time offset. It'll probably take you 15 minutes to get everything set up and ready to test, and then just a few seconds for every channel of your crossover.
Hi Mike

haven't seen you for a while!

I'd love to be able to do this accurately, alas my computer skills in using some of these programs are my achilles heel. Do you know of any 'tutorial' which, with a bit of application on my part, may get me through it?

Maybe most people learn this stuff through osmosis..I doubt it's taught in school. Or their work? Anyway, as i only got a computer itself about two years ago, this stuff just sends me gaga. A good book? Beats me how people are able to pick it up.

I'm proud that it only took me a couple of months to work out REW!! ha ha ha
 

·
Premium Member
Joined
·
6,711 Posts
I've added the ability to use the left channel of the soundcard as a timing and cal reference in my current beta build of REW so that will be in the next release, might even get it out over xmas :)
 

·
Registered
Joined
·
509 Posts
I've added the ability to use the left channel of the soundcard as a timing and cal reference in my current beta build of REW so that will be in the next release, might even get it out over xmas :)
meaning we can measure accurately the delay of a sub compared to the mains...and correct it with the tools at our disposal?

AWESOME!:T
 

·
Premium Member
Joined
·
6,711 Posts
Well, yes and no. You can determine the delays in the system being measured based on the location of the peak of the impulse response, but the width and risetime of the peak depend on the range of frequencies the system producing it can reproduce. Main speakers, when tested with a full range signal, produce nice sharp peaks and easily measured delays. On a subwoofer the impulse response does not look much like an impulse, it has a slow rise and a broad peak because it cannot reproduce high frequencies, so using the time to the peak of the sub response and trying to compare it with the main speakers is not comparing like with like. You may find that using a sweep that only covers the range both sub and mains can reproduce would give impulse responses from each that are more readily comparable, otherwise it will be interesting to see if Mike can find the article about this topic.
 

·
Registered
Joined
·
438 Posts

·
Registered
Joined
·
1,384 Posts
I read it, but it went way over my head... I need to read up on this, I'm afraid.

Is there something I can look for in the impulse response graph or elsewhere to optimize subwoofer delay? I have set it for best FR now, but apparently that isn't nescessarily best.
 
1 - 18 of 18 Posts
Top