Alignment Tool Phase/Impulse - Help

[banshee]

So from what I can tell i think this new tool would very quickly and accurately help setup phase and/or IR so I really want to learn how it works!

Using the latest beta build, and I think I have my setup correct. I posted previously how its setup but basically a Scarlett with Output 1 (Left), Output2 (Right-Reference) and Ref Input as Output2. This seems to be working and measuring right (I can see delays from loopback, but not sure.

I am only testing with this setup as I will re-do the entire tune once I completely understand and have this Alignment part understood and working!!

So I figured the phase would be close since they follow the XO curve properly however I have all delays set to 0 rights now on the DSP so expected to have to ADD delay to at least everything but the SUB, however according to the way I have setup the tool, its nearly perfect with no delay, maybe a very small amount to match perfectly.

The sub delay from referenced loopback is 26ms, and midbass is only 9ms, so figured at least this is the delay I would need to put into the DSP as a delay, but according to the tool, its not needed?

Maybe I just dont understand this? I will include the mdat as well.

Can someone please help out here?

 

[FargateOne]

@jtalden is your master here! my young padawan!!

 

[jtalden]

The sub delay from referenced loopback is 26ms, and midbass is only 9ms, so figured at least this is the delay I would need to put into the DSP as a delay, but according to the tool, its not needed?

The delay listed in the measurement notes can be misleading. When using the REW alignment tool this delay is not used.

Well done on the analysis.
A delay of -1.0 ms is the most favorable for SPL and phase tracking. This is a trivial SW delay shift from the posted 0.0 ms setting.

Optionally, a delay of -9.8 ms with the SW inverted is slightly more favorable for group delay and is only slightly different in SPL and phase tracking.

Either of these 2 settings is very favorable. I attached my modified mdat file. Measurements 1, 2 in my file are your measurements 5, 6.

And yes, since the SW cannot be delayed from 0.0 ms you will instead increase the delay of all main speakers.

 

[banshee]

Very nice, thanks! Will continue to look into the details of that mdat. Its interesting how one can invert a speaker and add some delay to fix things like group delay. I dont know much about group delay but heard about it.

Also, I was/am still unsure how the SUB being all the way in the trunk and the MB being right under the drivers seat can be in near-perfect phase with little (1ms) to no delay. Was I interpreting this correctly? Maybe its right.

I am also re-doing the entire tune this week, so once I finish the FR stuff will be ready to get back into the IR/Phase stuff!!

Really looking forward to getting this system tuned nicely! I think I am getting very close. Will post again soon................

 

[sm52]

It's amazing that midbass and sub are almost the same in time. Maybe in addition to the DSP, the playback device has its own delays?

 

[banshee]

It's amazing that midbass and sub are almost the same in time. Maybe in addition to the DSP, the playback device has its own delays

Right! I think I ended up inverting polarity of the MB and then ran again and it gave a delay that made more sense. It also seems to have worked just as good. Interesting for sure.

 

[banshee]

So I have a few more questions for using this alignment tool as I am most likely doing something wrong.

I am turning my car system (tweeter, midrange, midbass, and single sub).

I worked my way through both sides, I think for some reason the left seemed to work better, but that is most likely due to what I will describe here as my issue, LOL.

Process:

Take sweep of SUB (10-200hz) using the above setup all using output 1, ref output 2 and timing loop2 (out 2).
Then mute and do same with MB. Use AT and it quickly gives delay (12ms) for MB. Measure both speakers and very nice 5-6db at XO, so perfect alignment!

Questions:

1) Next, do I have to take sweep of the MB again now that I added delay to it (to align to sub) or can I use the same (original) sweep?
2) Lets say I get through ALL of left and they are each perfectly aligned with each other and sum correctly. Currently I am reversing my Sweep outputs to shift to measure the Right side as well as the reference and loopback. MAYBE this is not right and causing my issues? I will show the screenshots below of how I run Left vs Right.
3) So, maybe #2 was wrong and causing my issues as once I was done with all this each side summed very good individually but when all played together I had pretty much NO TA or staging, if something was not 100% L or R, it was playing strongly in the other side. I do feel the measurements I took on the Right side were WAY off and not adding the proper delay based on the left so I hope if I get this right, it solved most of my issues.
4) lastly, I know I can also try to adjust each pair of drivers (L MR vs R MR) but if each side was done "properly" would there be much to gain from adjusting both pairs of same drivers?

I really appreciate all the help!!

Left sweep setup:

Right sweep setup:

 

[sm52]

To align the impulses of all participants (there are 7 of them), you need to make 7 time-referenced measurements to one speaker (tweeter included in the system of 7 or additional, the eighth). Then you will see everyone on the timeline at the same time. The delays also need to be set relative to one, the one that was the reference for all 7 measurements.

 

[banshee]

To align the impulses of all participants (there are 7 of them), you need to make 7 time-referenced measurements to one speaker (tweeter included in the system of 7 or additional, the eighth). Then you will see everyone on the timeline at the same time. The delays also need to be set relative to one, the one that was the reference for all 7 measurements.

Oh, I was doing it differently, each successive speaker would measure against the next. I figured if the MB was now aligned with SUB (first speaker-refernce) this method would work, maybe its just "not right"?

 

[banshee]

Also to add, when I am aligning, I am aligning to phase for everything up to the final MR > TW, where I switch to IR.

 

[banshee]

Any advice here? Basically when I move to the right side, do I need to reverse all my outputs and inputs for those measurements?

 

[sm52]

Too much time needs to be spent to understand what you are going to do and whether it is right. I used a different method to measure speaker latency in a car. I burned a disk with sweep files and sync impulses on a CD, You can do the same on a flash drive if there is a usb for the flash drive in the player in the car. Read this thread. https://www.avnirvana.com/threads/acoustic-timing-reference-measurement.11571/post-87789 My post #9. This is not the whole process. This is part of him. If you follow this path, I will explain what to do next.

 

[banshee]

Thanks, Interesting, looking into this option now.

 

[banshee]

So this is the Acoustic timing reference method correct?

If so, I tried to follow this exactly as you posted. I muted all channels except for both L/R tweeters and played the L_ref_L.wav file on the media player and started the measurement sweep. I heard the 2 sweeps and it made it to the end, however had multiple errors, so I think something was def off.
I also then tried to run a sweep with REW playback, however seems to be the same issues, so maybe I did something wrong?

Errors:

Setup:

 

[sm52]

Is there something wrong. The signal-to-noise ratio is 7-10 dB. This is very little. You should hear two reference whistles, at the beginning and at the end. If you are measuring tweeters, why is REW picking up the low frequency signal but not the high frequency.

 

[banshee]

Yep, I hear them in fact I have the gains all the way up on the mic gain and they even go Red! Loudness is prob 80-90db. I agree something is wrong, will work on it again today.,

I do appreciate the help!

 

[banshee]

Just wanted to give an update on this. I think I figured out the underlying issue. My setup and process. I think by changing the Ref when shifting from L/R was causing issues and just "wrong" from what I can tell. I re-did everything leaving Ref on R and all speakers using L output and got measurements from each. So still tweaking the process, but everything is at least working as expected. Thanks sm52.

 

[sm52]

leaving Ref on R and all speakers using L output

Yes, the reference must be the same for all measurements in the session.