DiyMobileAudio.com Car Stereo Forum banner
1 - 20 of 26 Posts

·
Registered
Joined
·
639 Posts
Discussion Starter · #1 ·
I've found a few threads but they left me confused. Can anyone point to a good thread (or add info here) about the best way to use the phase adjustment in my Helix DSP ? (its the older unit but I have the latest tuning software version 4.x or something).

I know phase and time delay are intrinsically linked which is what confuses me. I'm mostly unclear if I'm supposed to use the phase adjustment to adjust the image or to improve coherence between the different drivers. For example, am I trying to use phase to bring the LW and LM into coherence (primarily at the xover point) or am I using phase to focus the image between left and right ?

I've read that phase is especially important for the sub, so do I enter the proper time delay for the sub and then adjust the phase on the sub ?

I've used the calculator on this page but I"m not sure how to utilize the numbers it gives. I think I'm supposed to enter the high pass xover value of the lower frequency driver and the time delay currently applied to the lower frequency driver (or would it be the delta time delay between the lower and higher frequency dirver ?) and then it spits out the phase of that signal. So does that mean delay the phase on the lower frequency driver by the number it calculates ? Is the phase control on the DSP "delaying" the phase ? (as opposed to "pushing" the phase forward in time.

To further complicate things, the Helix software does not allow phase adjustments on a channel labelled "low" ie they expect you to adjust the phase with the woofers being the zero point I guess. I could just label my woofer channels something else that would enable the phase shift on them.

Anyway if anyone knows any good discussions on this or wants to add some advice I'm all ears. Ha ha little audiophile pun there for you.
 

·
Registered
Joined
·
19,894 Posts
That "phase" adjustment is just a second order all past filter that uses degrees relative to the crossover point. Dont use it unless you know how to use an all pass filter. The polarity button is a different story.

Sent from my SM-G950U1 using Tapatalk
 
  • Like
Reactions: chasingSQ

·
Registered
Joined
·
639 Posts
Discussion Starter · #3 ·
I appreciate you chiming in Skizer but that doesn't really help me much other than "don't use it". I did get an EE degree a few decades back and I understand the general sense of the all pass filter is only affecting phase and also has an inflectin point and an associated q, but I don't understand how that translates into car audio tuning.

I read the AudioTec Sound Tuning article again, my approach will be to maximize SPl at the x-over points by altering phase. It seems to work out as the phase reference for the sub is the high pass xover point and the mid and tweets is the high pass x-over point.
 

·
Registered
Joined
·
19,894 Posts
I read the AudioTec Sound Tuning article again, my approach will be to maximize SPl at the x-over points by altering phase. It seems to work out as the phase reference for the sub is the high pass xover point and the mid and tweets is the high pass x-over point.
You dont need to do this. Really. Google the "audiofrog tuning pdf". If your fiddling around with phase, something is wrong. If anything, you should be fiddling with timing instead of phase, but even then, not much.

An all pass filter comes into play hardly ever in a one seat tune, but they can be helpful for 2 seat tunes where there is mo center channel/upmixing available. Your essentially using the all pass to fix the comb filtering that path length differences cause instead of time alignment.

Sent from my SM-G950U1 using Tapatalk
 
  • Like
Reactions: chasingSQ

·
Registered
Joined
·
639 Posts
Discussion Starter · #5 ·
Okay thanks that's very unambiguous advice - I will follow it.
 

·
Premium Member
Joined
·
258 Posts
You dont need to do this. Really. Google the "audiofrog tuning pdf". If your fiddling around with phase, something is wrong. If anything, you should be fiddling with timing instead of phase, but even then, not much.

An all pass filter comes into play hardly ever in a one seat tune, but they can be helpful for 2 seat tunes where there is mo center channel/upmixing available. Your essentially using the all pass to fix the comb filtering that path length differences cause instead of time alignment.

Sent from my SM-G950U1 using Tapatalk
So, I happen to run across a video showing how to "TA" the sub using only phase and not time alignment. I figured it sounded easy, so I tried it. It was rather easy to adjust the phase up/down and get mostly upfront bass and it sounds decent.

I am wondering if I could get better results using actual TA with the sub also. The rest of my system of course is TA'd pretty good.

Thoughts on this? Sounds like I should consider trying TA instead? :D
 

·
Registered
Joined
·
19,894 Posts
So, I happen to run across a video showing how to "TA" the sub using only phase and not time alignment. I figured it sounded easy, so I tried it. It was rather easy to adjust the phase up/down and get mostly upfront bass and it sounds decent.

I am wondering if I could get better results using actual TA with the sub also. The rest of my system of course is TA'd pretty good.

Thoughts on this? Sounds like I should consider trying TA instead? :D
an all pass filter is just going to eliminate ONE comb filter, not all like time alignment. Its not the correct way.
 

·
Premium Member
Joined
·
937 Posts
I read the AudioTec Sound Tuning article again, my approach will be to maximize SPl at the x-over points by altering phase. It seems to work out as the phase reference for the sub is the high pass xover point and the mid and tweets is the high pass x-over point.
Thank you preston. This so far has been the 1st stab I’ve seen on this thread at offering a methodology to actually using this feature on the Helix DSP.

I too yearn to know what are the appropriate steps or methods to follow when utilizing the phase adjustments on the Helix DSP.

What are others using out there to measure and verify the phase adjustments they make on their Helix units? Systune? SMAART?

Might get more help from reading in this thread too:

So, I can tell you this. The basic concept can be explained with words. If you're adept at using measurement software, then the logistics of pulling off these basic concepts can be figured out.

So, contrary to popular belief, REW DOES measure the actual phase, it CAN calculate the phase based on the magnitude (this is what happens when you click "show minimum phase.... It does a Hilbert Transform of the magnitude to show the minimum phase response that such a system could, in theory, have).

The logistical problem with performing phase alignments in REW is that after EACH AND EVERY sweep, you have to go in and manually set your reference delay (by clicking in the little gear in the upper corner of the IR window and inputting a previously arrived at time in ms)

It is imperative that you use the same reference delay number for each measurement that you're attempting to align.

Then, of course, you use the overlay window to compare each set of measurements and decide which one needs more or less delay/phase shift.

An all pass filter IS NOTHING MORE than frequency dependent delay.

Finally, when comparing two traces to see which one needs adjusting, the slope of the phase trace shows you relative arrival time. The steeper the slope, the greater the delay.

That's really it. match the slope of the traces, use polarity to get them to overlay, if necessary.

a lot of this basic concept I was able to capture on video before the lighting went to ****. Unfortunately, I found myself explaining all of this stuff using Smaart and SysTune, then when I would try to show how you'd do it in REW, I'd get so lost in the logistics of trying to pull it off with REW that I'd loose my train of thought....

I'll take a look at the videos I did get today, because it sounds like some of it might actually still be of use to some folks.
 

·
Premium Member
Joined
·
782 Posts
i've used helix's partial phase adjustment on midranges and tweeters. the use case i've had for it is lower midrange frequencies (as approaching the crossover frequency) pulling left.

one usual solution is centering tones at those frequencies to center them up, but sometimes end up losing left presence/pressure for the sake of a solid center.

so i added a little partial phase (11-33*) on the left and subtracted a little TA.

this helped me keep the left midrange presence/pressure good while keeping a solid center.

tweeters can play into frequencies that are phase-localized so the same effect works with them as well.

too much partial phase adjustment and i've obliterated my LC/RC, causing LC to pull way left and RC to pull right. it might seem like a wider stage but LC/RC are pushed out.

i found doing partial phase equally to both left and right doesn't change anything, which makes sense when you think about it :)
 

·
Registered
Joined
·
3,677 Posts
I've found a few threads but they left me confused. Can anyone point to a good thread (or add info here) about the best way to use the phase adjustment in my Helix DSP ? (its the older unit but I have the latest tuning software version 4.x or something).



I know phase and time delay are intrinsically linked which is what confuses me. I'm mostly unclear if I'm supposed to use the phase adjustment to adjust the image or to improve coherence between the different drivers. For example, am I trying to use phase to bring the LW and LM into coherence (primarily at the xover point) or am I using phase to focus the image between left and right ?



I've read that phase is especially important for the sub, so do I enter the proper time delay for the sub and then adjust the phase on the sub ?



I've used the calculator on this page but I"m not sure how to utilize the numbers it gives. I think I'm supposed to enter the high pass xover value of the lower frequency driver and the time delay currently applied to the lower frequency driver (or would it be the delta time delay between the lower and higher frequency dirver ?) and then it spits out the phase of that signal. So does that mean delay the phase on the lower frequency driver by the number it calculates ? Is the phase control on the DSP "delaying" the phase ? (as opposed to "pushing" the phase forward in time.



To further complicate things, the Helix software does not allow phase adjustments on a channel labelled "low" ie they expect you to adjust the phase with the woofers being the zero point I guess. I could just label my woofer channels something else that would enable the phase shift on them.



Anyway if anyone knows any good discussions on this or wants to add some advice I'm all ears. Ha ha little audiophile pun there for you.
I think what everyone here is saying is, get your time alignment dialed in properly first, then if there is some sort of phase alignment issue in the crossover regions, that's when you use the phase adjustment. As I understand it, phase adjustment is really only useful in the areas where the sound from one driver interacts with the sound from another, so xovers between drivers, or tweeter-tweeter, mid-mid, etc. You can get an overlays of measured phase and predicted phase in REW from the "overlays" button at the top of the window, of course you'll need sweep measurements to get that data. Go to the "eq" window, play with an all-pass filter, go to the predicted phase overlay and see the results. You should also uncheck all of the drivers you're NOT trying to blend to clean up the graph. Play around in REW so you can at least get an idea of what you need to do in the Helix.
 

·
Registered
Joined
·
5,111 Posts
Here’s the real storey that you won’t find on audio frog tuning page.


The “phase” ona helix moves the entire magnitude of the speaker channel
By however many degrees it says it’s moving it.

It has a built in calculator to take an “all pass filter” and set it down the highpass slope using the phase shift of the high-pass as a insertion point.

What happens when you put an all pass at the “knee” of a high-pass or a lowpass? Well they do the inverse of each other except on 12db LR or even order LR crossovers, than they behave the same.

If you add a all pass at a point where the phase is moved by x in the highpass , whoever x is the allpass center is 90deg on both sides of the all pass. They call that quadrature.
The 2nd order all pass will eventually go to -360deg from 0 degrees. If you insert the all pass where the phase is advanced 11degrees it will return to -360deg plus 11 because it has no idea where zero truly is. Where you insert it where ever that phase angle is at it will go -360 away from that at the end of the filter, thus shifting the entire magnitude 360deg plus 11deg , so a 11deg shift plus one cycle.

The helix has a calculator to use delay to realign to the other side compensation for the single cycle by using delay.

Have you ever tryed moving the phase slider really fast? That channel will cut out and come back on why? Because it’s crunching a bunch of math as you move the slider.


The way helix does it is so superb and is a true constant group delay( minus the inband if the allpass ) I’m fairly sure the Q is set to something extremely high in the stop band and somewhere down far enotthat you don’t hear the inband of the allpass , it goes into quadrature and back to at a extremely high angle.

In a nutshell it works and absolutely can be the one thing that makes your system image way better. But it also could be completely useless and not needed.

That depends of the transfer function (the acoustic all pass) of left vs right from your listening position.

So I’m a one seat tune the way helix does it is absolutely the best way to do it.

It’s not “just an all pass filter” it’s also a very complicated calculator that does a lot in the background to net you a constant group delay that works very good imo
 

·
Registered
Joined
·
5,111 Posts
here is a sim so you can see I added a all pass at 150hz where it was crossed at 300 and got a smooth constant phase shift on the in band
the stop band has a much higher angle of change, the use of delay would move a speaker to a higher angle if it is moved ahead of another speaker

if delay is used and set behind a speaker the phase gets pulled back


this is the magnitude and phase of a mid with no phase shift other than the shift caused by crossover

'

here is with adding an allpass one oactave below
you should see a linear phase shift in the inband of the speaker
santander bank store locator



they look almost the same, however if you look closely at the phase it is moved forward in the midrange...by only a few degrees about ten I would guess
I didn’t look close enough but it’s a small amount....point being it’s a linear and constant shift on the inband (the part you hear)


All pass filters are extremely valuable and useful
If anyone says don’t use one they just simply don’t know how to use one properly. There so many ways an all pass can be used to offset time related issues between speakers. For us being left side biased and all pass could be exactly what is needed to get your system doing what it should.

Not using an all pass and just using delay will definitely line up the two channels
In time. However we do not listen at even angles, we listen at odd angles in a car, we can doninstall tricks to make things better like tow speakers in and such , but the spinorama of a speaker is directly related to the angle of incidence.

If your sitting 90deg from the axis of a speaker guess what.....

To better answer this if your sitting 180deg from the axis of a speaker guess what
 

·
Registered
Joined
·
606 Posts
here is a sim so you can see I added a all pass at 150hz where it was crossed at 300 and got a smooth constant phase shift on the in band
the stop band has a much higher angle of change, the use of delay would move a speaker to a higher angle if it is moved ahead of another speaker

if delay is used and set behind a speaker the phase gets pulled back


this is the magnitude and phase of a mid with no phase shift other than the shift caused by crossover

'

here is with adding an allpass one oactave below
you should see a linear phase shift in the inband of the speaker
santander bank store locator



they look almost the same, however if you look closely at the phase it is moved forward in the midrange...by only a few degrees about ten I would guess
I didn’t look close enough but it’s a small amount....point being it’s a linear and constant shift on the inband (the part you hear)
Thanks for posting this.
 

·
Registered
Joined
·
719 Posts
here is a sim so you can see I added a all pass at 150hz where it was crossed at 300 and got a smooth constant phase shift on the in band
the stop band has a much higher angle of change, the use of delay would move a speaker to a higher angle if it is moved ahead of another speaker

if delay is used and set behind a speaker the phase gets pulled back


this is the magnitude and phase of a mid with no phase shift other than the shift caused by crossover

'

here is with adding an allpass one oactave below
you should see a linear phase shift in the inband of the speaker
santander bank store locator



they look almost the same, however if you look closely at the phase it is moved forward in the midrange...by only a few degrees about ten I would guess
I didn’t look close enough but it’s a small amount....point being it’s a linear and constant shift on the inband (the part you hear)


All pass filters are extremely valuable and useful
If anyone says don’t use one they just simply don’t know how to use one properly. There so many ways an all pass can be used to offset time related issues between speakers. For us being left side biased and all pass could be exactly what is needed to get your system doing what it should.

Not using an all pass and just using delay will definitely line up the two channels
In time. However we do not listen at even angles, we listen at odd angles in a car, we can doninstall tricks to make things better like tow speakers in and such , but the spinorama of a speaker is directly related to the angle of incidence.

If your sitting 90deg from the axis of a speaker guess what.....

To better answer this if your sitting 180deg from the axis of a speaker guess what
I read this six times and I still don't know what it means.
 

·
Registered
Joined
·
3,139 Posts
I read this six times and I still don't know what it means.
I must admit me neither, how it’s helpful to anyone I’ll never know, I think he’s trying to say he’s made the phase graph a straight line if unwrapped by adding an all pass at the low end… but that’s not the same as making the phase flat like some very well designed home audio speakers are, so I’m not sure what it achieves 🤷🏽‍♂️
 

·
Registered
Joined
·
5,111 Posts
I read this six times and I still don't know what it means.
that was from 2019 .....
A lot has changed since then.
I’ve learned a lot more since then and so that would / could have errors , especially in the terminology. That was three years ago

I’ll have to edit (I need time on this tho) and read it again....
 

·
Registered
Joined
·
5,111 Posts
I read this six times and I still don't know what it means.
lmao.... yeah I might have been drinkin too much when I wrote that... lol I don’t understand most of it ...

i will say ive come to dislike APF, they move the wrong way to do anything meaningful, although it has its uses. This helix , like the mosconi PhaC, great idea, just can’t get it to sound right. And have subsequently discovered thistype of phase alignment leaves much to be desired in stop band information.

i think I was doing a sim on the allpass effect of shifting the magnitude of phase, got that part out and went off on a tangent that didn’t make any sense.
 

·
Registered
Joined
·
749 Posts
There’s a video on YouTube of the new JL Audio Measurement Max.

They used an all pass filter to correct some mid range phase issues that arises after they performed their EQ corrections. Goto 20min mark in the video. The Max could measure the amplitude and phase properly and while showing it in real time. It is way too difficult to do this in REW since the captures are not being displayed in real time. So yeah this stuff has big potential use to correct for phase issues in the acoustic domain. Phase issues are sometimes caused by reflections that EQ cannot correct for. IIR filtering affects the phase in the electrical domain which causes problems since now we’re sending out of phase left and right signals for the speakers to play. I’ve heard from someone on this forum that correcting the amplitude to match both sides will also correct for the phase, but I don’t think that’s true and the JL video kind of demonstrates that. The interior of the car will do stuff with the phase that will throw it off unless you use an all pass filter to bring it back in phase as close as possible. FIR filters solve for this so the phase remains true and flat throughout, but the car’s interior and speaker install location will ruin some of the perfect phase that was retained due to reflections. It’s better to send true in phase signals to the left and right side then applying all-pass filters where is needed to correct for phase issues in the acoustical domain.

I have also read that all pass filters could be used to finely dial in 3 way systems. Time delay will get it close but if you want to finely dial in the phase of the speakers you need to be able to measure phase between two speakers and align them using an all pass filter at the crossover point as the center frequency. If you don’t know how it’s probably better off to leave it alone as it could cause more problems than good if you’re just guessing.

Does anyone remember how many degrees is shifted for every few ms of phase adjustments?

I will be testing the new Bit Tune 3.0 software soon using the new MPC PERCEPT auto phase correction algorithm (designed for car audio vs home audio like Dirac Live) on the new Bit One HD v3.0 firmware. They advertised to auto correct amplitude and phase using the Bit Tune spatial averaging mic.

Here’s a video explaining the new software and phase:
 
  • Like
Reactions: vactor

·
Registered
Joined
·
719 Posts
There’s a video on YouTube of the new JL Audio Measurement Max.



Does anyone remember how many degrees is shifted for every few ms of phase adjustments?



Here’s a video explaining the new software and phase:

Phase is adjusted in degrees and delay is adjusted in mS. The number of degrees per delay adjustment varies by frequency because the number of cycles per second varies.

At 80Hz, there are 80 sine waves per second. So, one 80Hz sine wave takes 1second / 80 cycles = 12.5mS.
At 20Hz, it's 1/20 or 50mS
At 1kHz, it's 1mS

Sound travels at about 1132 ft/sec. So...

At 80Hz, the wavelength is 1132/80 or 14.5 feet.
At 20 Hz, it's 1132/20 or 56.6 feet
At 1kHz it's 1.132 feet

So, at 80Hz, one degree of phase is equal to 14.5 ft /360 degrees or 0.48"
At 20Hz, it's 56.6 / 360 or 1.86"
At 1kHz, it's 1.132 feet / 360 or 0.0377"

So, delaying the signal in mS adds more phase shift as frequency increases. A first order all pass filter rotates phase by 180 degrees below the filter frequency. At the filter frequency, the phase is 90degree.

A second order all pass filter rotates the phase a total of 360 degrees and the phase is 180 degrees at the frequency you choose. The graph looks like this one below. When you adjust the all pass filter, no matter whether the adjustment says it's in degrees or frequency, you're just shifting the frequency of the filter.
Rectangle Slope Plot Font Parallel


Sometimes there's also a Q adjustment and that changes the slope of the line, which you can think of as the rate of phase change or the number of adjacent frequencies over which the change occurs. The filter above has a Q of 0.707

This one is a Q of 2.

Rectangle Slope Plot Font Parallel
 
1 - 20 of 26 Posts
Top