XO and Time alignment

PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

XO and Time alignment

Post by PhilActiveAudio » 27 Feb 2021 18:26

Hi Uli,

One short question: do Time Alignment delays deducted with one XO type (e.g. Neville Thiele) can be applied to all other XO types without proceeding with a new full Time Alignment operation?

My guess: since Time Alignment is to correct delays between drivers related to their physical location in space, I presume I can apply the same corrections (i.e. rotations) with all XO types without having to proceed Time Alignment for each XO type.

Thanks in advance.
Phil
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
User avatar
UliBru
Posts: 442
Joined: 19 Oct 2019 13:58

Re: XO and Time alignment

Post by UliBru » 27 Feb 2021 20:04

Of course the physical distances between the drivers will not change when you switch the XO type.
But the XO itself may have a slight influence on the timing behaviour.
But I believe you can simply use the same delays as before. A new delay measurement is then just a game of finetuning.
Acourate also corrects the phase in macro4.

So simply go ahead and report your findings ;)
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

Re: XO and Time alignment

Post by PhilActiveAudio » 18 Aug 2021 15:41

Just made some checks today and I share with you my results with 2 Neville Thiele crossovers (order 1 and 2).

With the same crossover, delays may be slightly different between drivers of the left and right speakers. For me, some samples only for bass drivers. So the Time Alignment procedure is to be done for each driver of each speaker to be as perfect as possible.

With two different crossovers, delays may be slightly different between drivers of each speaker. For me, some samples only for bass drivers. So the Time Alignment procedure is to be done for each driver for each crossover to be as perfect as possible.
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
User avatar
Labdoc
Posts: 95
Joined: 19 Oct 2019 19:11

Re: XO and Time alignment

Post by Labdoc » 18 Aug 2021 15:56

So the Time Alignment procedure is to be done for each driver for each crossover to be as perfect as possible.
@PhilActiveAudio addresses the question of accuracy of a time delay measurement with driver-Crossover combination #1, when applied to driver-Crossover combo #2.

Another question to consider is the precision, or reproducibility, of time delay measurements. In may experience, consecutively repeated time delay measurements show variation. My reaction has been to do 5 or so consecutive measurements and average the result. If this observation of variability is reproducible in @PhilActiveAudio's setup, an alternative explanation of the difference reported between each driver-crossover measurement could be due to random variation of the measurment.
Acourate/AConvolver on Win 11x64 > Thunderbolt > RME UFXplus (as DAC/preamp) > Balanced cable > Wyred4Sound SX1000 Icepower mono amps > ATC SCM50PSLT and stereo JL Audio F113 subs
PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

Re: XO and Time alignment

Post by PhilActiveAudio » 19 Aug 2021 22:28

So far, I cannot say if these measurements are reproductible as I performed only one stereo measurement for each speaker and filter combination.

To be precise, I did as following:
1) Left : mic at 1,20m distance from speaker and 1m height. Logsweep with filter 1 and then logsweep with filter 2. When I compare the two measurements, Woofer delays are slightly different (some samples) but Medium and Tweeter delays are equivalent (normal for Tweeter at MaxM at 6000).
2) Right : mic at 1,20m distance from speaker and 1m height but at a different position in the room. Logsweep with filter 1 and then logsweep with filter 2. When I compare the two measurements, Woofer delays are again slightly different (some samples) but Medium and Tweeter delays are equivalent. When I then compare Left and Right measurements, Woofer delays are is slightly different (some samples) but Medium and Tweeter are equivalent.

It already looks quite reproductible for Medium and Tweeter, but needs to be formally confirmed by x different measurements as proposed.

Woofer difference between Left and Right could be partly explained by the shape of my speakers: 3 independent items (1 Jensen box, 1 Arai horn, 1 tweeter) that are well positioned one on another but probably not at 1 mm as far as left and right Jensen/Arai are concerned.
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
User avatar
Labdoc
Posts: 95
Joined: 19 Oct 2019 19:11

Re: XO and Time alignment

Post by Labdoc » 19 Aug 2021 22:57

My setup is quite different: 3-way main speakers + active stereo subwoofers. My measurements are from the main listening position. The time alignment is for the crossover between the subs and the main speakers. So, similar to your experience, it is mainly low frequency time alignment I am dealing with.
Acourate/AConvolver on Win 11x64 > Thunderbolt > RME UFXplus (as DAC/preamp) > Balanced cable > Wyred4Sound SX1000 Icepower mono amps > ATC SCM50PSLT and stereo JL Audio F113 subs
PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

Re: XO and Time alignment

Post by PhilActiveAudio » 26 Sep 2021 12:32

I performed several LogSweep measurements with 3 different filters: NT order 1, NT order 2 and UB jPol11 order 2. The 3 drivers of the left and right speakers have the exact same delays except the bass driver which have +1 sample with NT order 1.

Finally, it seems that delays measured with one filter may apply to all filters.
To be confirmed however.
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
User avatar
UliBru
Posts: 442
Joined: 19 Oct 2019 13:58

Re: XO and Time alignment

Post by UliBru » 26 Sep 2021 13:03

The linearphase XO filters do not change the time alignment by themself. So basically the relative time between drivers keeps constant.
But drivers also have their own (minpase) behaviour. Thus a driver will change its timing dependent on the corner frequency. Possible the timing may also change dependent on filter order. But I do not expect big changes, maybe a sample or two.
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

Re: XO and Time alignment

Post by PhilActiveAudio » 26 Sep 2021 19:22

Indeed, crossovers frequency were equals but filters were different. I will measure different crossovers frequency to complete my findings.
How much impacts in terms of samples crossover frequency changes may have?
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
PhilActiveAudio
Posts: 55
Joined: 21 Jan 2021 21:51

Re: XO and Time alignment

Post by PhilActiveAudio » 03 Oct 2021 19:22

To share with you my findings.

Below are the measured delays with 3 different crossover type / order combinations and different Mid/High XO frequencies.
Measurements were done with the mic at the exact same position for all combinations for left and then right voices.
The left part of the table is varying because the reference driver (the one that is set at 6 000 samples) is not always the same (Mid or High).
Results are very close to the ones done previously but with a mic at a slightly different position.

Drivers delays variation with crossovers.JPG
Drivers delays variation with crossovers.JPG (212.07 KiB) Viewed 2208 times

Relative delays calculated in the right section tells:
> Low (Supravox) / Mid (Radian): different delays with a stable XO at 600 Hz.
> Mid (Radian) / High (Fostex): stable delays with different XO from 6 000 Hz to 8 000 Hz.

It seems that it is better to measure each situation (XO type, order and frequency) to measure the specific delays to be applied.
This is finally what I did for each combination I am currently testing.
exaSound e68 DAC ==> class D / Accuphase / class D ==> Jensen Supravox / Arai Radian / Fostex
Post Reply