can REW measure sub delay?

[terry j]

hi guys

it is very possible that this question has been asked in another form or for other reasons.

I've recently been mucking about with integrating subs into my existing three way system that runs a deqx. The deqx only runs three channels, but can run up to six if two are daisy chained. Need I say that I'd rather achieve similar results without the expense of an extra deqx??

First up the reason, with my current set up I can get flat in room response down to 29hz, but due to the driver I'm using (which has a natural roll off in the box at around 70 hz) it takes quite a bit of boost to achieve that. The driver is an 18 inch pro driver so it doesn't really complain, but logic tells me that there must be extra distortion which must be a good thing to get rid of if possible. If I can get lower as well then it has to be better all round.

What I have done is run a splitter out of the bass output into a Behringer dcx 2496, I had one laying around and so can take advantage of the x-over points and slopes etc as well as use the inbuilt parametrics.

I won't go into the whole story as it's not germane to the question.

So far it has been successful. I now have a minus 3 db point around 13 hz, and I run the sub up to around 80 hz which means the main woofer has no boost at all.

Plenty of output (for now 4 peerless xls 10's, will build a matching sub for the other side, maybe overkill but will look more balanced). As I say, plenty of output but listening tonight to some bass heavy tracks, had the feeling that something didn't quite gel.

What I think is happening (finally to the point!!) is there must be some sort of processing delay when the signal goes through the dcx. I put the subs in the same plane as the mains, ie equidistant from the LP.

Now, if it's true that there is some delay thru the dcx, then of course the sub needs to come closer to the LP in order to time align.

Is there some method in REW which will tell me if there is a delay and if so how much?? I dunno, maybe the impulse tab, or the ETC tab or something. If that is true, can a quick explanation of how to read it be given??? (thanks if that is the case).

On further reflection, the same must be true of any of the BFD's, but I would imagine in most cases people here would use some sort of sub out jack, and for all I know there could be an inbuilt delay in the processor that could compensate. I of course can't do that with my arrangement.

If there is no means of determining this delay via REW, there is a large thread on DIY on the dcx 2496, I would imagine I could ask over there, and someone might be able to give me a figure. Once I have that, I suppose the trick would be to move my subs closer according to the rough formula of 1ms=one foot, would that ge correct??

Thank you all, hope that any answer given here might be useful to others as well.

cheers

[brucek]

there must be some sort of processing delay when the signal goes through the dcx.

The BFD and other DSP processors such as the DCX contribute a processing delay of about 1 msec. This calculates in time to about 1 sec (speed of sound 343 m/sec=~1000ft/sec = ~1ft/msec).

If you have a receiver with a time delay, you simply add a foot to the sub distance (so the receiver will subtract the msec and advance the sound). You can also use phase adjustment.

You can use the Energy Time plots of REW, but it takes some interpretation to get used to, and you have to remember that the plot is relative to the impulse and not an absolute measurement of time. This then shows you the sub pulse in time after the impulse. (i.e., the impulse is set to 0msec)

You need to zoom the horizontal time base to only look at the number of msec in distance that the sub is from the microphone. So if the sub is 6 feet away, perhaps use a scale of 10 msec (10 feet). You should see the sub pulse at about 6 feet (6msec). You can add a main with the sub and attempt to time align the two peaks, but it's quite tricky to interpret (IMHO). The best way I found to ensure I was looking at the correct peak was to change the time delay in the receiver and see if the peak moves in the next sweep and then I know I'm looking at the correct speaker. Or take a sweep of the sub and main, and then take a sweep of just the main and use that to interpret if the peak you feel is the sub is actually so...

Anyway, you can play around with it. Other things that are useful with the Energy Time curve is testing a main speaker and see where the reflections are that come after the main peak....

brucek

[terry j]

thanks Bruce

the thing of most import right now was the figure of 1ms delay.

I did a very brief listen with the sub pulled out towards me a bit, around a foot I'd say heh heh, and on first listen the impression was that it had tightened up.

That was on a completely different set of music than the one I thought sounded 'odd' for lack of a better word.

Will redo the listening tomorrow, and maybe even have a play with Energy Time plots you spoke of, see what (if anything) I can glean.

But, the data on 1 ms was most definitely what I was after, thanks.

EDIT just had a quick look at the energy time plot, and from a very fast play it seemed to me that %FS would be the easiest graph to interpret, would that be right?

Gee, I'm not holding out much hope after a second look ha ha ha. Might stick with a tape measure methinks!