Alignment tool difficulties

[HenrikasMT]

Hey guys, my suggestion or Idea for easy time allignment for speakers which im missing on REW, is not using allignment tool, but having possibility to run pink noise and live phase trace on top of current one. for example at the same time id like to have my right speaker measurement taken, seeing phase graph and then measuring my rear right speaker and seeing actual live phase trace on top of the front one, so I could add delay and see what will change and how, alignment tool is very limited to its things, because having either you measure with furthest timing reference or you have to type it in manually, long process, which smaart does over 2 steps.
What would be your thoughts guys?

 

[John Mulcahy]

either you measure with furthest timing reference or you have to type it in manually, long process

There's no need for typing, just drag the alignment tool fine delay slider and look at the result, which updates as you move the slider.

 

[HenrikasMT]

There's no need for typing, just drag the alignment tool fine delay slider and look at the result, which updates as you move the slider.

If you look at wrapped phase you will not see anything proper in region of 500-6k region. Thats why it has to be IR alligned with furthest driver. But its just bold, to make additional measurements for just time alligning in with reference timing of the furthest driver. For that it would be much easier to have a window where you run one driver and on top you live run another to have it properly alligned. Much better imo and much faster.
Additionally, I care more about phase allignment of drivers at any point than seeing sumation on the graph without proper phase values.
For example my system delay is 100ms, and If I want to allign 2 drivers with different distances, the phase plot in allignment tool will be just wrapped phase with not much info.

 

[John Mulcahy]

Thats why it has to be IR alligned with furthest driver.

You can use Impulse alignment in REW if you prefer, it has the advantage of automatically detecting when phase should be inverted.

For that it would be much easier to have a window where you run one driver and on top you live run another to have it properly alligned. Much better imo and much faster.

Smaart does live so well I don't really see why you would want to use anything else.

For example my system delay is 100ms, and If I want to allign 2 drivers with different distances, the phase plot in allignment tool will be just wrapped phase with not much info.

Are you using V5.31.3? In the V5.40 beta there's a "Remove time delay" option to remove any delay common to both responses so their relative phase plots make more sense. That feature was suggested by d&b.

 

[HenrikasMT]

You can use Impulse alignment in REW if you prefer, it has the advantage of automatically detecting when phase should be inverted.


Smaart does live so well I don't really see why you would want to use anything else.


Are you using V5.31.3? In the V5.40 beta there's a "Remove time delay" option to remove any delay common to both responses so their relative phase plots make more sense. That feature was suggested by d&b.

Yes I do use it. For example my tweeter is showing after IR delay proper phase graph but in time allignment its completely wrapped phase gap, it gets worse if you tick off the remove time delay box. Thats why I am confused about this.

 

[John Mulcahy]

Probably easiest if you attach an mdat with the two measurement, or email it to the feedback address if you don't want to share it publicly.

 

[HenrikasMT]

There you go, no worries. Public its fine aswell

 

[John Mulcahy]

I can see 3 issues, two related to the setup and environment and one an issue in REW.

1. Your measurements have very different timing offsets, some shifted by 170 ms and some not. You should use the timing offset control on the Measure dialog to remove the 170 ms or so delay that is common to your measurements when you make them, that on its own should make future measurements easier to understand.
2. You are measuring in an environment with a lot of very strong reflections. In most of the measurements reflections are larger than the direct sound, making alignment more art than science - algorithmic solutions cannot really accommodate that kind of environment. It isn't meaningful to talk about aligning the phase of responses when those responses are actually multiple slightly time-shifted versions of what came out the driver that are summed at the mic.
3. REW is using a different phase smoothing algorithm in the alignment tool window that doesn't cope well with the phase jumps that are caused by reflections. I'll update it to use the same algorithm as used elsewhere in REW, but there may be a small penalty in responsiveness of the tool as the processing load will be increased.

I'll move these posts into their own thread.