Which is easier to read?
This?
Or this?
Proper phase alignment in the field is hard. It’s one thing to measure in the near field when you can get a nice clean trace, but once you get into the far field and add reflections it gets a lot harder.
One thing you can do to make the whole alignment process easier on yourself is to flatten the phase trace around the area of interest. To do this, add more delay to the delay finder.
A standard main+sub phase alignment goes like this:
- Measure Main solo. Auto set delay. Save trace.
- Measure Sub solo.
- Make the pictures match.
The easier it is to read the graph, the easier it will be to match the pictures. Let’s find out how.
Normally, using the procedure above, you’ll start out with a flat phase measurement in the high end (on time arrival relative to the reference) and it will slope up and wrap around as you move to the left in the low end (late arrival relative to the reference).
We want to move that flat portion to the left and we’ll do that by adding delay to the delay finder in Smaart.
But how much delay?
Start by adding delay with the arrow keys so you can see the gradual change. How much delay do you need to move the flat portion of the phase graph to 100Hz?
If I click down incrementally, I find that it takes 10.5ms to move the flat portion of the graph to 100Hz. Let’s see if we can do it faster with a quick calculation.
number of wraparounds * 10ms = additional delay
current delay + additional delay = total delay for an on-time arrival at 100Hz
Let’s try it.
Here’s a measurement of a Meyer Sound UPJ-1P. I can see that the phase is flat starting at 2kHz, but we want it to be flat around 100Hz. It looks like there are about 1 wraparound starting at 100Hz and moving left.
1 * 10ms = 10ms
90.52ms + 10ms = 100.52ms
Looks good. Much easier to read.
Let’s try a real measurement from the field. This one is more difficult to read, but it looks about 1 wraparound.
Let’s add 10ms just like in the last example.
74ms + 10ms = 84ms
What have you found that helps you read the phase graph? Comment below and let me know.
Credit to Timo Beckman for introducing me to this idea.
Can i translate this for our Korean friends??
Sure. Of course. Do you want to publish it on your own site or on Sound Design Live https://sounddesignlive.com?
Well, you’ve found the reason for the Low Delay Finder 🙂
A linear phase has a constant group delay. A flat phase has no constant group delay, except it is zero. Is that right? Why we want a flat phase?
Hi Markus!
> Is that right?
Hmmm, I don’t think I can comment. I’m not an expert with group delay, but I think you are correct.
> Why we want a flat phase?
This article is about making the phase slope less steep so that it is easier to read, not for any reasons of quality.