Erin's 2006 Civic Sedan

[erinh]

Tuning tips that might be worthwhile...



Just a spoonful of sugar...

At the Vinny I got some good feedback from the David Hogan. Hogan is probably one of the best critical listeners I've ever met. Not only does he give poignant critiques, but he also provides excellent suggestions on how to improve the sound.



He gave me a couple pointers here and there on my sheet (which I've linked at the bottom of this post). I had a few minutes last night to implement those suggestions. 1dB cut between 1-1.6khz on the right. A 1dB cut at 400-500hz on both sides. Saved the result. Toggled between my 'Vinny' setting and the newly updated one. While not a 'night and day' difference, it's still pretty apparent the focus tightened up. And when it did, there was more detail in the tracks I used to compare the tunes (Natalie Merchant - Wonder & Depeche Mode - Enjoy The Silence. IMO, you can't expect to make little changes have a great impact right away. Usually these kind of minute changes with a solid affect happen when you've gotten pretty far down in the tune (or assuming you already have an excellent baseline). If you have a hodge-podge setup and T/A, phase, levels are out of whack then adjusting a single dB here or there isn't really going to do much. It's like adding $1 to $1million. But, once you get to a point where your setup is blending well and things are pretty well tuned - not perfect, but good - these kind of changes can really help out.





Sub/Midbass tuning. Remember: It's all about the frequency wavelength.



Now, here's another thing that's quite interesting. I dumped my EQ settings on my midbass and subs a few weeks ago. I realized that the way I had gone about EQ'ing them was backwards. I had EQ'd each individually. What I should have been doing was looking at them together. Why? Wavelength. You've surely heard about staggering midbass/subs to help smooth the response. It's nothing new. The problem is, when you EQ a driver by itself in this passband you're not accounting for the interaction and influence of the other driver. That can both hurt and help you. The key is, you don't know unless you evaluate the system as a whole. Midbass to midbass this is more important in the entire passband. Sub to midbass, this is really only important at and just around the crossover (maybe 1/2 octave +/- the crossover point).



So, I measured the response of the subs + midbasses. I then measured the response of each midbass by itself and subs by itself. I did this last bit so I could better analyaize the influence of each speaker. For example, if I had a bump at 150hz, was it a single driver problem or both? For the most part, what I found was the response was pretty nominal, when evaluating at 1/24 octave*. I only had to make a few adjustments here or there.

* using 1 octave or 1/3 octave smoothing will cause you to fix things you can't repair and stuff that doesn't need repair. I'd suggest only using 1/3 or 1 octave smoothing for midrange frequencies and above (namely, tweeter).





Then I played the sub and evaluated the crossover point. But to do this, I went back to my phase adjustments. There's no sense trying to EQ something if the phase is out of whack at the crossover. Using the method I laid out here (Subwoofers And Time Alignment), I focused on measuring the response while adjusting the actual phase on my Helix DSP.

If you don't have a phase delay, just use time alignment; it's the same thing. The phase adjustments just allow you to do T/A in larger steps. To have an idea how much T/A you need for the sub/midbass adjustment you can use this site:

Use this link to calculate the relationship between phase angle/frequency/time:

Phase angle calculation time delay frequency calculate phase lag time shift between voltage difference time of arrival ITD oscilloscope measure two signals formula angle current voltage phi phase shift time difference - sengpielaudio Sengpiel Berlin



For example: to adjust the phase at the crossover point of 80hz by 15 degrees you'd need to delay in time 0.50ms.



This is what's important to understand:

15 degrees at 80hz requires 0.50ms... Think about that... that's a big number in time for such a small delta in degrees. To put it in perspective, think back to the time where you were adjusting your T/A values on your midranges to line them up. Typically, 0.5ms can mean the difference in too far right or too far left; you definitely move a lot of space laterally between midranges with 0.50ms. You typically moved in those 0.03ms steps and worked within a small window of probably 0.20ms to really tighten up your center focus.



Now compare that to when you were doing T/A on your sub by a measly 0.03ms and couldn't get things to sound right. It's because you need to adjust the delay for LOW frequencies by a LOT more to get the phase to swing just a little bit, as compared to a tweeter who's phase at 2khz can be a full 360deg out with that same 0.5ms.



Make sense? Good!

So, if you want to really adjust your subwoofer, I'd suggest doing so in 0.5ms chunks. Or, even using 2ms chunks (about 45 degrees at 60hz). Adjust incrementally until you've gotten to the point where smaller steps will help. Then switch over to 0.50ms steps.



I did the above, again with the Phase adjustment on my Helix (and again, this is the same as using time delay in large chunks). I put the mic at the seat and swept. Saved. Adjusted phase again. Swept. Saved. Rinse, wash, repeat. On the RTA screen, I zoomed in so I could focus on the 30-200hz area and I focused solely on the crossover point. I found the phase adjustments that got me the highest SPL. Then I went outside the crossover point and looked at how that blended with the rest of the system.



After picking the phase angles I wanted to use, I swept the system with an impulse. I then looked at the group delay for the 'better' phase angles just to see if anything was out of whack. To my surprise, the GD was very benign. When I compared the new phase setting vs my old one with EQ it was a huge difference in smoothness.



It's worth noting that, again, thanks to the long wavelength of low frequencies, you're likely not going to get the sub in phase at the crossover AND frequencies outside of this. In my case, I just worked with the trade-off. I took the setting that provided me a high SPL at 70hz (my crossover point) and also good SPL below and above that. Make no mistake, you WILL run in to this same issue. Regardless of your crossover point. More so if you overlap; less so if you underlap. But you'll see the trade offs.





Now, back to EQ. If you read the scoresheets I posted, you'll notice a couple comments regarding the low frequency response. Matt Roberts made a note of 30hz being out of phase. He was dead on. I noticed this in the GD data but at the time just dealt with it. What the culprit is was a low frequency boost I had; the boost was set at the lowest frequency I could and ran to about 35hz; this was just an attempt to add some extra low end. What happened, however, was a phase null at the seat. When I dropped that particular boost from my tune, the null practically went away. I'm still having to deal with a bit of an out-of-phase characteristic there right now but I don't think it's anything anyone else besides a well-trained judge would have picked up on had I not discussed it here.





I'll try to update this thread with some graphs when I get the chance to illustrate just what I'm talking about with regard to the frequency response and group delay results.


Score sheets in PDF format:
https://www.dropbox.com/s/la9xj8spfn...nny%202014.pdf



Hope this helps some folks out. The most important factor here is realizing the relationship of frequency to time/phase. And how you waste your time (no pun) trying to adjust subwoofer delay in very small increments.

[erinh]

I'll also add this:
You don't have to have an RTA to adjust the subwoofer delay. It's helpful because it takes a lot of guesswork out of the equation. But, simply, you can play a test tone at the crossover frequency and adjust delay until you have the highest SPL by ear. Listen for the bass to be up front. You'll likely find an area within about 45 degrees that seems to be in phase. To help really narrow it down, go outside of the passband to your midbass by one tone (ie; if you cross your subs/midbass at 60hz, go up to 80hz) and repeat the steps to fine tune the delay. You should now have your sub DIALED.

FWIW, I can now make my dash mat move at the front of my dash. Before when it was less in phase, it couldn't do that.

[chad]

The phase thing you note is the difference between "sounds good" and hells yes."

It is also evident in all facets of audio. Some never ever get it.

[chithead]

Interesting advice. Thank you sir for this and everything that you post. I'm impressed the amount of dedication you put into making each and every audio project absolutely incredible.

[erinh]

The phase thing you note is the difference between "sounds good" and hells yes."

It is also evident in all facets of audio. Some never ever get it.

Indeed. Even though I've understood it for a while, sometimes you kind of forget about it. Or you think of a different way to go at something and it improves. I feel with each iteration the system gets better. But the one thing that really makes or breaks the enjoyment factor is the sub/mid integration.

The common thing I see people doing is adjusting their sub by itty bitty increments... and then giving up because the results weren't what they wanted. When you break it down and realize how much delay needs to be used to really "move" the woofer due to its wavelength, it makes it MUCH easier to get things blended.

Interesting advice. Thank you sir for this and everything that you post. I'm impressed the amount of dedication you put into making each and every audio project absolutely incredible.

Thanks, man. I appreciate it.

[BigAl205]

Thanks for posting that, Erin...great advice as always

[keep_hope_alive]

awesome information as usual, Erin. well done.

i've been trying different ways over the years to try to get a handle on phase issues and have seen that getting phase correct at the listening position greatly reduces EQ needs and certainly helps develop a nearly-coherent combination of drivers. i look back only 4-5 years ago and realize i was spending time trying to "fix" phase issues with just EQ and Crossovers without actually understanding the phase issues i was having.

you've helped give me more ways of quantifying phase issues, which is greatly appreciated.

For those reading that would like some more information on group delay and phase shifts at crossover points, this may be useful: http://www.rane.com/pdf/ranenotes/Li...ers_Primer.pdf
Page 9 discusses drawing vector diagrams for evaluation of phase at the crossover point. I'd like to see this adjusted for car audio to factor in differences in distance - maybe see if it can derive your end conclusion developed through RTA and T/A.

[Kevin K]

Thanks Erin for sharing that, now for all that to sink in and then re-read and then try out in real time.

[erinh]

thanks for the feedback. now... GO TRY IT!

[chad]

Indeed. Even though I've understood it for a while, sometimes you kind of forget about it.

I've always been picky about phase response, people go "wow it must be really important with a big rig live." Well, it's MORE important when you are in a smaller venue and you are trying to work with sound that is already present coming from stage. Delaying to the backline helps but if your phase response is all over the place like a shit sandwich then it will never be right. If it ain't right you have to fight it, if you have to fight it then it's too loud, if it's too loud you lose audience and gigs. I've seen guys bring way too much PA into a club to fight a phase response issue that happens when blending with the stage volume.... I just bring too much because I'm a dick. I have also seen guys downsize a PA AFTER adding really good processing and getting a good tune going... Most, if not all, modern manufacturers will help you attain this... My rig is OLD so it took some time in a pasture to get right... Fortunately, now I own my own pasture

i've been trying different ways over the years to try to get a handle on phase issues and have seen that getting phase correct at the listening position greatly reduces EQ needs and certainly helps develop a nearly-coherent combination of drivers. i look back only 4-5 years ago and realize i was spending time trying to "fix" phase issues with just EQ and Crossovers without actually understanding the phase issues i was having.

Hammer--->Nailhead.

WAY too many people go at the EQ like a a window licker in a glass factory. The goal, flat, or a certain target EQ... Yeah they end up getting it, or damn close, but it sounds like hammered dogshit... No life, no realism, nothing CLOSE to what it is capable of because the thing is just fighting itself. And they keep going, and going then adding more EQ, more processing when in reality, just backing up to the starting point and thinking about it yields better results... Less is more.

And really, it's really not their fault, it's human nature... People get it in, set it up to "something" to get audio, then they go nuts, forgetting that "something" may not be right and they are literally polishing a turd that should not be a turd.

I try to use as little EQ as possible, even on a channel strip.. I'll move a microphone anywhere it will not get destroyed to get to as minimal EQ as possible unless I'm using the EQ as an effect, etc.