Helix P-SIX - EQ Linking Help

S

swrocket

Noob
[FONT=&quot]So I've been tuning my front stage (Still waiting on the Sub to arrive) and got it pretty good on each side.[/FONT]
[FONT=&quot]To do each side, I linked the left TW/MR/MB and applied EQ and then did the same for the right and level matched them.[/FONT]
[FONT=&quot]The over all front (both sides together) though, needs a bit more EQ. I have the bands, but I wasn't sure what would happen if I linked the entire front stage (all speakers) together and then applied the 4 eq bands...or should I just apply the same to each side separately?[/FONT]
[FONT=&quot]
[/FONT]
[FONT=&quot]Any tips would be greatly appreciated.[/FONT]
[FONT=&quot]Here is the REW response graph. (I'm using a Half of Whitledge curve)[/FONT]
[FONT=&quot]
1618676479747-png.298553
[/FONT]
[FONT=&quot]
[/FONT]
[FONT=&quot]and here are the EQ adjustments needed in REW[/FONT]
[FONT=&quot]
1618677436716-png.298554
[/FONT]
 
What frequencies are you crossing your speakers?

Ideally you would tune each speaker indvidually, level match them to each other at that time, but get the crossover's right (this really helps get phase issues fixed). Looks like you are having some issues at the 300hz region (Is this where you crossed your midbass to mid range?) and then from 700-2,000hz when summed. This is most likely a phase issue in some way to have that issue. Eq isn't going to solve a phase issue. It might "fix" the graph and make it look pretty, but that isn't how you want to fix that.
 
If you get dips while summed it’s likely phase related, so if you get a dip it’s better to apply an all pass to the good side to match the bad side rather than trying to correct the response with eq, as it’s phase related adding more eq won’t help and will have a negative effect on the response as any extra power/level you put in will likely be heard, but not actually do anything to your summed graph

Remember if you don’t get 4+db of summation (to the top end of mids, 5-6db in midbass) it’s due to phase being out due to reflections and standing waves, if you have a phase related narrow dip you don’t pile eq into it... so why would or should you when it’s between left and right sides?

your timing looks good from those three traces

don’t stress over phase related differences, add a little correction with a high q all pass perhaps, but otherwise that’s a pretty good overall response

try maybe a 2nd order all pass on your left side at 310hz to see if you can get better summation of the two drivers, your right side has an obvious dip caused likely by phase, so you deliberately mess the phase on the opposite side so the phase then matches better and you get better summation, it takes a bit of messing because we don’t have phase information but it doesn’t hurt to try

1.4k also looks like it has something phase related going on...

uneqd 1/48th octave graphs of the above drivers would reveal a fair bit more also
 
Last edited:
Thanks dumdum.
My x-overs (electric) are at 327Hz (MB to MR) and 2788Hz (MR to TW) all LR24.
I do seem to have some issue on the right mb to mr but not sure why...I did get ~4db summation for MR and TW and ~5db on the MB just summed with eq only on each side.
I have un-EQ measurements for all drivers individually, as pairs and each side, but only at 1/6th Octave. I'll retake at 1/48 and post.
 
So I tried the individual driver method (vs each side) and I am really at my wits end. I cannot understand what the issue is in this car...
Here are my measurements after 3 rounds of EQ (measured at 1.48 smoothed to 1/6) all using UMIK-1 moving mic method at half volume on the HU
Acoustic X-Overs at 90-330-2900

I really need some help here.

Here is left
View attachment 15642

and right
View attachment 15643

something is weird with the mid range...
Here are drivers and driver pairs
View attachment 15644
 
It's obvious that you have same phase issues with both L+R are playing (both in the midbass and midrange). You can probably fix those with a few well-placed allpass filters. You also don't seem to be getting very good summing at the crossovers. Usually, this is a timing-related issue. Absolute worst case, you could use an allpass fitler to get better summing at the crossover as well.

I had similar issues with dips in the combined L+R response - and a few allpass filters solved those dips completely.

You just have to experiment a little. For the L+R midrange dips, try a 2nd order allpass fitler at ~1100hz with a Q of 3 or so. Without being able to measure phase, you just kind of have to mess around with the values. I usually start with frequency and Q values that are similar to what I would use to correct the problem with parametric EQ. Most of the time, those values work pretty well for allpass fitlers.
 
No - just on one (left OR right). It may work better on one than the other. I usually try it on the left side first and then if I'm not getting any results, I'll try it on the right side. If you put in on both sides, it will do absolutely nothing (ask me how I know!). :-)
 
Also try a few on your midbass. Maybe try one at 310hz and one at ~410hz - both with a Q of 7 or so (both 2nd order). That might take care of those summation issues on the midbass as well.

Allpass filter are awesome when you need them. Some has mentioned that they can cause other issues, but I personally have had nothing but success with them.

If the wide Q=3 allpass filter on the midrange doesn't work well, maybe try two smaller Q ones instead - try to "fix" each dip individually instead of both with a single allpass filter.
 
Look forward to hearing your results. You will have to play with the exact frequency and Q value a little to get the best results, but I was amazed at what allpass filters did for my measurements.
 
So the 1100 AP helped some as did the 410. tried a few Q values and freq but only made it worse.
View attachment 15645

The overall response though is still horrible. I'll play around some more and see what I can do. I just don't understand the AP filters enough.
 
swrocket said:
So the 1100 AP helped some as did the 410. tried a few Q values and freq but only made it worse.

The overall response though is still horrible. I'll play around some more and see what I can do. I just don't understand the AP filters enough.
Click to expand...

They actually did help. Instead of the 1100hz AP, maybe try splitting it into two separate AP's - one at ~950hz, Q=7 and another at ~1300hz,, Q=6 (just rough estimates on the freqs). It really is a guessing game unless you can measure phase - but if you are persistent, you can usually find something that works.

Keep in mind that the small peaks and dips really aren't that big of a deal. It doesn't look as "pretty" on paper as a "perfect" measurement, but you probably can't even hear the little stuff.

Your right side seems to be summing OK at the crossovers, but not your left side. If you know the time delay is good, you can try an allpass to fic the big dip at the left midbass <-> midrange xover. Like I said, pretty much always use 2nd order allpass filters and just try the same hz and Q values that you would use if you were using parametric EQ to try and fix the dip. That is what has worked best for me. But I still had to spend a lot of timing getting everything to sum correctly (L+R pairs as well as speaker-to-speaker crossover areas).

Does each speaker like up with the target curve in REW properly? The crossover point looks a little low on the left side midbass -> midrange xover.

Don't give up - I don't thoroughly understand allpass filters either, but you'll get it if you mess with it enough. Basically, just use an AP filter to fix any of the dips in the combined L+R response that don't exist in the individual speaker measurements. Obviously, if there is a dip in the invidual measurements, it will show up int he combined L+R measurement as well (don't need AP filter for that - just better EQ on the individual speakers.

What does the overall measurement look like?

Most importantly, how does it sound? It's easy to get carried around with the measurements, trying to make them look perfect - but not really needed (I do the same thing though). :-)
 
Your right midrange measurement when midbass, mids and tweeters combined shouldn't differentiate from the single mid measurement in the passband at all.
This is not a timing or phase issue if it's in the passband of a single speaker between two different measurements.
The crossovers seem to have a timing/phase problem on the left side between midbass and mids and a not as severe timing/phase problem on the right side between same drivers.
You got REW, you got a UMIK-1, so you can measure acoustical phase in these areas with sweeps with acoustic timing reference and check the phase.
Recheck your time alignment for midbass and midrange speakers. Speakerpairs can easily be checked with impulse measurements. Crossovers from mid to tweeters aren't bad as you will almost never get full summation up there due to all the reflections that mess with phase very easily so i wouldn't worry about that area.

When doing the phase alignment do this for driver pairs, i.e. measure phase of both midbass drivers and both midrange drivers and adjust delay for all speakers upwards at once, i.e. if REW says add 1 ms delay to midrange, you also need to add this 1 ms to the tweeters to keep relative timing between midrange and tweeters in check.
 
You can use Open sound meter too with a internal loopback via audio midi in macbook or asio4all on windows and your umik1. Clock drift issues will mess with phase data above ~8,000hz, but sub to midbass and midbass to midrange and midrange to tweeter should all be easily viewable.

This includes midrange to midrange phase stuff (and all other speakers besides tweeter to tweeter higher frequencies, but that is very hard anyway.)


https://www.caraudiojunkies.com/showthread.php?5156-Open-Sound-Meter-RTA-Phase-ETC-Program
 
Measuring phase is the best way to use all pass filters period, guessing can have you chasing your ass for a while, it’s far better to measure phase for both sides and then apply all pass to match them

if you don’t get summation in a crossover it’s likely you have a phase issue there... timing is good in his drivers as he gets steady summation through the audible bandwidth

if you don’t get summation in a crossover you try moving the crossover first... you can also try expanding the bandwidth or driver and checking summation of both, if you don’t get 4-6db of summation in certain freqs it’s simple... don’t put crossovers there ;)

if you can’t go lower or higher and need to implement it, for example if midbass have an allpass applied on the left at 350 and your crossover is 360 then you have previously had one driver with bad phase due to reflections at 350 and made the other side have bad phase at 350 to match with an allpass... simples... but then when the two drivers don’t sum with the mids at 360 it’s NOT timing if you have good summation as in this case through the rest of the pass band... it is the fact you now have two midbass with ‘bad’ phase and you want to match a pair of mids up with good phase around the crossover... so add a pair of all pass filters to the mids at 350 also... then you have ‘bad’ phase on all four drivers so you then get summation &#55357;&#56397;&#55356;&#57340;

it’s all about matching phase to other drivers
 
Thanks all. I will try to figure out how to measure phase, and also try different XO points. The electric xovers are a mixed bag and then I EQd each driver to the acoustic LR4 crossovers
One strange thing is that if I just set LR4 electric xovers, I need to flip the midranges to get summation in the xovers, but using the mixed crossovers (not LR4) I didn't do that...
I am crazy with work the rest of the week but will spend all Saturday on this and will post back.
 
swrocket said:
Thanks all. I will try to figure out how to measure phase, and also try different XO points. The electric xovers are a mixed bag and then I EQd each driver to the acoustic LR4 crossovers
One strange thing is that if I just set LR4 electric xovers, I need to flip the midranges to get summation in the xovers, but using the mixed crossovers (not LR4) I didn't do that...
I am crazy with work the rest of the week but will spend all Saturday on this and will post back.
Click to expand...

thats not weird, the natural roll off phase of the drivers is likely changing 180 (ie a 12db slope) already so as you add the 24db slope you then effectively make a 36db slope and have to flip phase due to this... when you create an acoustic 24db slope you then get the correct phase through the crossover as it’s the acoustic that determines the phase shift and therefore what polarity changes need making
 
You must log in or register to reply here.
Back
Top