-
Why is "beating speaker frequency response into submission" via EQ a bad thing?
I've seen this mentioned a few times, where someone says that your "beating your frequency response into submission" by "over-EQ'ing" the response. I'm kind of curious what is really meant by this and why it's a "bad" thing.
Let's say that you just installed new speakers and you are EQ'ing each speaker to a particular curve, whatever that curve may be. Why is it a bad thing to use as many filters as needed to "shape" the response of the speaker exactly how you want it? At the end of the day, I would think that you'd want to have the response of each speaker match the curve of your choice as closely as possible. Why not use as many filters as needed to accomplish that goal? Why would you NOT want to make it match your target curve as closely as possible if you have the bands to spare? Are there any "cons" of doing this?
I mean who cares if it takes 20 PEQ filters to get the response exactly how you want it? Regardless of *why* you need to use so many filters, doesn't the ends justify the means? I mean if you can reach your goal with EQ instead of having to make physical changes, why not? Why not take advantage of the functionality offered by modern DSP's to do most of the work for you, if possible?
I'm genuinely curious to hear peoples opinions on the matter.
For what it's worth, I'll use as many of my 31-bands-per-channel of PEQ filters as needed to correct the response of each speaker - and it seems to work extremely well for me. Even if I use 5 filters just to make small .5dB corrections in 5 different places - why not? Maybe it helps, maybe it doesn't - but what does it hurt? Just curious why it seems so "frowned" upon to use "too many" EQ filters...
Thank you!
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
jtrosky
I've seen this mentioned a few times... I'm kind of curious what is really meant by this and why it's a "bad" thing. ....
Seems clear enough to me:
Quote:
"beating your frequency response into submission. Meaning EQ'ing whatever the hell is in your path to match some arbitrary curve without regard for actually optimizing the system."
https://www.diymobileaudio.com/threa...stakes.420955/
/thread
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
There have been volumes written on this and similar topics by people waaaay smarter than me, so I'm certainly not going to sum it up here. But, regarding signal manipulation, I think it has to do with "how" you EQ more than "how many times". For example, it seems to be "generally" accepted that boosting a signal via EQ is more detrimental than cutting a signal... when you introduce energy into a signal by boosting, you also introduce phase shift which becomes counter-productive to what you are trying to accomplish. There are tons of variables that come into play, speaker size, distance from the listener, early/late reflections, the volume of your interior, etc... hell, what makes one speaker sound different from another?
I know I haven't answered your question really, but if you've heard the phrase you quoted, you've probably also heard that installation is everything. Meaning that the more work you put into the proper design of a system to begin with, the less you will have to EQ the heck out of it later on.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
I think because people don't understand what they can and can't fix. People just try to get it to match a curve perfectly which usually results in them trying to "fix" dips that shouldn't be bothered with.
Like I had one at 400hz when I was 2 way. It was something that never really responded to eq. To match the curve, I would have had to either lower the curve alot or go crazy with boost. Either one of those is bad for the overall system.
Instead, you need to actually analyze your system and figure out what is and isn't fixable, and then make it match the curve the best only on what is actually fixable. That includes setting your curve at a level that makes sense and ignoring those things that can't be fixed.
Also, just because you matched a curve, all that means is that the levels are all worked out, there are quite a bit more parts to tuning than just frequency levels.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Sucking the life out of what could be an otherwise great tune is a real problem in my opinion. I've heard systems that imaged so tight they lacked personality. This is what I consider "beating a curve into submission". It's easy to make a system sound over processed when you have a ton of power at your fingertips. I still think my best tune for overall presentation was done with a processor where I had to make every eq band count and even move stuff around when I ran out of empty bands to work with. I beat my tune into submission with my Helix and it's imaging pretty tight but it lacks that personality needed to really rock out. I'll be taking a less is more approach next time to see where that gets me.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Grinder
Sorry, let me clarify my question... :-) I understand what they phrase *means*, but I'm trying to understand why it's considered a "bad" thing. Sure, in a perfect world, the physical install and speakers used would give you the perfect frequency response that you are looking for without any EQ. Obviously, we don't live in a perfect world. Before powerful DSP's were readily available for little $$$, I can understand putting tons of time, effort and money into the physical install in order to get the best response that you could.
But nowadays, with inexpensive, powerful DSPs available that allow you to easily "shape" a response however you want via powerful parametric EQ, I'm trying to understand why, if the end freq response is the same, is it "bad" to use EQ to get to the same result that you could/would have gotten to with better physical install/drivers, etc. Basically, why NOT beat the response into submission if it ends up with a better overall freq response in the end?
It is my understanding the that cars are a "minimum phase" environment, for the most part - and that if you EQ so that the freq response the same for all speakers, the phasing pretty much "falls in line". Again, that is my understanding - at a very basic level - which is why I asked the question. Working with the install that someone has, why NOT EQ until you get the freq response that you are looking for? What are the cons vs. not shaping the response to where you want it (via DSP)?
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Thanks for all of the replies - I appreciate it!
THe reason I asked is because I think I have my system sounding fantastic - even though I'm using a LOT of PEQ bands to get there. I'm having a hard time not fixing an obvious "peak" or "valley" that *can* be fixed via EQ - simply because people say that you shouldn't beat the response into submission. :-) Again, just wondering "why not?" if it seems to work and you're not boosting a bunch of stuff (Overall, only a few areas of the freq response actually have a net boost - and even then, only by a dB or two.
In the end, I'm going to stick with what works for me, but just wanted to hear what others had to say on the subject.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
I don't have any scientific answers, and even though I have an EE degree I'm often confused when people start throwing the term phase around. I know what phase is in EE terms, but when you start talking complex musical waveforms and pseudo science I really don't always understand it. But what I've heard several times is that an EQ, being a filter, introduces phase shift and in fact some people say this is the whole point. Either way I think we can agree you are altering the source signal. And just like we (or at least home audiophiles) strive to simplify the signal path, it seems to me that the more your beat the curve into submission, the more you are altering the original "pure" signal source. (At this point I can almost hear Skizer chiming in to say that's the whole point, to make the signal at your ears match the source).
Now if you are talking about making your coarse adjustments and then using another few bands to really fine tune things here and there by 1-2 dB I can't really call that beating it into submission, ie subtle changes throughout the curve. But when you start thworing in +4 and -6 or more adjustments sometimes next to each other, you are clearly altering the raw signal significantly. And again, EQ is not some pure signal boost/cut, as they say it introduces phase shift, which surely introduces other artifacts to the pure signal.
I haven't brought any hard facts in here, these are just my audiophile fairy dust opinions. However, my experience has been if I go crazy on the EQ (crazy being defined more as large boosts and cuts vs how many bands I'm using) to match some arbitrary line on a screen, I've several times gone back through during listening tests and picked each band and "softened" the inputs I made based on REW. And in most cases, it ends up sounding better to me if I reduce the large inputs on the EQ.
So my partial answer to your question is I don't think the number of bands you use is as important as how large a change you are making in any given band, or how much boost/cut you swing between adjacent frequency bands that really introduces a lot of "unnatural change" to the pure signal.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
preston
I don't have any scientific answers, and even though I have an EE degree I'm often confused when people start throwing the term phase around. I know what phase is in EE terms, but when you start talking complex musical waveforms and pseudo science I really don't always understand it. But what I've heard several times is that an EQ, being a filter, introduces phase shift and in fact some people say this is the whole point. Either way I think we can agree you are altering the source signal. And just like we (or at least home audiophiles) strive to simplify the signal path, it seems to me that the more your beat the curve into submission, the more you are altering the original "pure" signal source. (At this point I can almost hear Skizer chiming in to say that's the whole point, to make the signal at your ears match the source).
Now if you are talking about making your coarse adjustments and then using another few bands to really fine tune things here and there by 1-2 dB I can't really call that beating it into submission, ie subtle changes throughout the curve. But when you start thworing in +4 and -6 or more adjustments sometimes next to each other, you are clearly altering the raw signal significantly. And again, EQ is not some pure signal boost/cut, as they say it introduces phase shift, which surely introduces other artifacts to the pure signal.
I haven't brought any hard facts in here, these are just my audiophile fairy dust opinions. However, my experience has been if I go crazy on the EQ (crazy being defined more as large boosts and cuts vs how many bands I'm using) to match some arbitrary line on a screen, I've several times gone back through during listening tests and picked each band and "softened" the inputs I made based on REW. And in most cases, it ends up sounding better to me if I reduce the large inputs on the EQ.
So my partial answer to your question is I don't think the number of bands you use is as important as how large a change you are making in any given band, or how much boost/cut you swing between adjacent frequency bands that really introduces a lot of "unnatural change" to the pure signal.
Which is why FIR filtering is a different beast than IIR.
(One can achieve phase EQ along with amplitude EQ.)
If the output matches the input signal more closely we have a metric to define quality or fidelity.
I suppose can argue whether phase is important or not.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Holmz
I suppose can argue whether phase is important or not.
I would like to see those arguments as to why it wouldn't be...to me, phase is super important. You can't have destructive interference happening and expect great results. If volume is equal and you have 180 out of phase, you get literally no sound.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Which is why FIR filtering is a different beast than IIR.
(One can achieve phase EQ along with amplitude EQ.)
I've heard that Dirac Live uses FIR, and that the Helix people say its not worth using in car.
Other than that I have no idea what that means and while I can look up the acronyms I really don't understand what these filters are really doing.
Also, other than using Dirac Live, we have no ability to utilize an FIR filter. But interested to hear more.
(When I say have an EE degree, it was from 30 years ago and I wasn't the best EE student which is why I ended up in software).
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
preston
...
(When I say have an EE degree, it was from 30 years ago and I wasn't the best EE student which is why I ended up in software).
I like morning humour... that is a classic.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
I'm not sure I can add anything of value to the discussion, that hasn't already been said. If you are using REW auto EQ and imputing the values from REW into your DSP, you can limit how much boost you let the feature use. I have tinkered with the preferences in the past, and limited the amount of boost that can be used when calculating the frequencies. The helix allows up to +6db boost, but you can tell REW to "pretend" it can only boost +2 or +3 when calculating the frequency adjustments. This will make you use more bands but will minimize the (sometimes) huge differences between neighboring bands on the EQ.
Hopefully this comment isn't incoherent babble, but try limiting the availabliity of boost in REW and it might take some of the large fluctuations away. I know sometimes REW tells me to cut 1055 -7db then the next band is 1080 +6. Its my understanding that this is what you want to avoid as much as possible, not how many bands used.
Full disclosure though, I have not noticed tonnes of audible difference between many of the different REW/Helix setups I have experimented with. Despite extensive tinkering, as blindly as can be at times, I feel sometimes the tune goes great and other times it's a complete failure. Strangely, I cannot pinpoint why that is, still learning.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
For me over eqing and cutting everything in sight wrecks dynamics, sure take down nasty peaks in treble, but you don’t need to flatten a response to match a straight line, for example... a pillars often have a peak around 200-1k or thereabouts and then level off with a lot flatter upper mid and treble... well first off does your brain hear everything as the mic picks it up? Direct sound is prioritised by the brain... is most of that energy your mic is picking up reflected? How many times have I turned all eq off and actually listened and enjoyed listening as I’ve effectively castrated the system and took 8-10db (think about the power that required if you were to boost the same?) is there any wonder it sounded a bit flat?
i guess I now look at a curve and try and smooth it rather than have done pre determined line I must make everything match, as long as everything sounds in proportion and dynamics aren’t stifled I have done good! And easy test is listen to a live session and decide if it actually sounds live or flat as hell
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
mumbles
There have been volumes written on this and similar topics by people waaaay smarter than me, so I'm certainly not going to sum it up here. But, regarding signal manipulation, I think it has to do with "how" you EQ more than "how many times". For example, it seems to be "generally" accepted that boosting a signal via EQ is more detrimental than cutting a signal... when you introduce energy into a signal by boosting, you also introduce phase shift which becomes counter-productive to what you are trying to accomplish. There are tons of variables that come into play, speaker size, distance from the listener, early/late reflections, the volume of your interior, etc... hell, what makes one speaker sound different from another?
I know I haven't answered your question really, but if you've heard the phrase you quoted, you've probably also heard that installation is everything. Meaning that the more work you put into the proper design of a system to begin with, the less you will have to EQ the heck out of it later on.
I gotta stop you.. boosting is not inherently bad. You just run the risk of clipping depending on how your gain structure is set up. With IIR filters, any swing in response changes phase. But, a given frequency response has a given phase response so those changes in phase are still (mostly/usually) fixing the phase response as well.
Sent from my SM-G975U using Tapatalk
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Well, I have mostly 2-way active systems and those are really hard to tune. I tune with a DSP that has just 10 bands of parametric EQ per channel and sometimes I don’t even use all of them so if some people are using all 30 bands of EQ per channel then I would guess they could be either doing something wrong like trying to fix drivers that are out of phase with each other or equalizing in REW using too high of a resolution and obsessing over stuff that doesn’t actually matter. As stated, trying to EQ things that cannot be equalized away like cancellation nulls or phase issues will cause you to use excessive bands of EQ.
Conclusion: beating the curve doesn’t necessarily mean it’s bad but it could be a sign that something else went wrong along the process of tuning and that person is trying to fix a problem with EQ that was caused by a mistake in the process. Some drivers, especially ones that are being asked to do things they aren’t really designed to do may also require more EQ.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
JCsAudio
equalizing in REW using too high of a resolution and obsessing over stuff that doesn’t actually matter. As stated, trying to EQ things that cannot be equalized away like cancellation nulls or phase issues will cause you to use excessive bands of EQ.
This sums up the problem I was having in the first few attempts at a good tune. Your entire comment is well said.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
jtrosky
I've seen this mentioned a few times, where someone says that your "beating your frequency response into submission" by "over-EQ'ing" the response. I'm kind of curious what is really meant by this and why it's a "bad" thing.
It's hard to understand the intricacies of tuning so I think many folks try to follow a tuning recipe, and like many recipes people get frustrated when there isn't enough pepper for their tastes. Or like if I followed advice word-for-word in a Home and Garden magazine I would end up with a really beautiful yard that I don't like.
I'd like to think that over-using the dsp is a sign that the user has more to learn.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
JCsAudio
Well, I have mostly 2-way active systems and those are really hard to tune. I tune with a DSP that has just 10 bands of parametric EQ per channel and sometimes I don’t even use all of them so if some people are using all 30 bands of EQ per channel then I would guess they could be either doing something wrong like trying to fix drivers that are out of phase with each other or equalizing in REW using too high of a resolution and obsessing over stuff that doesn’t actually matter. As stated, trying to EQ things that cannot be equalized away like cancellation nulls or phase issues will cause you to use excessive bands of EQ.
Conclusion: beating the curve doesn’t necessarily mean it’s bad but it could be a sign that something else went wrong along the process of tuning and that person is trying to fix a problem with EQ that was caused by a mistake in the process. Some drivers, especially ones that are being asked to do things they aren’t really designed to do may also require more EQ.
When Eqing a single channel response other drivers won’t have an effect, however I suspect you mean eqing a side or a pair of drivers together at the listening position, for which I am a massive advocate of getting phase and timing bang on before adding eq
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
dumdum
When Eqing a single channel response other drivers won’t have an effect, however I suspect you mean eqing a side or a pair of drivers together at the listening position, for which I am a massive advocate of getting phase and timing bang on before adding eq
That is correct dumdum. The process I use involves EQ of each driver separately, but also together with other drivers involved covering the same octave, like tweeter to mid and mid to woofer and woofer to sub and I do that left vs right sides. I don’t really want to get into too much detail about that though.
Speaking of process, I know some people don’t actually have one (I was guilty myself and I’m still learning) where they can use it and get repeated results with the understanding of what that is. They may be blindly messing around until something appears to work and get close to the result they want. They may have all these methods/tricks that they learned can work but don’t necessarily know why and when to use them in the right situation, or why. It’s really tough for the enthusiasts though because they may only ever tune one or two systems a few times a year. That just isn’t enough experience to gain the knowledge and develop a process, so the enthusiast will always be at a disadvantage when compared to someone who does it for a living. Still, they can learn if they put the time into it, read a lot, listen to those that have more experience, and follow those that are experts in this area, and lastly actually put some time in tuning on a semi regular basis. The best way to learn is to read up and practice it. Failure/making mistakes is part of the process of learning that applies to all aspects of learning no mater who you are. Even the professional tuner who does it on a regular basis I bet sometimes screws up and has to go back.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Over EQ-ing is a bad thing ..... with rise of processors that allows to load house curve ant tune system to it it become even more problematic and more people do that mistake, but they do have almost ruller flat output (dont mind the sound, lol)
EQ IS NOT a band aid for poor install!!!! but just too many uses EQ to solve install related problems
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
First off - GREAT information - thank you all for your replies! As I'm rather new to all of this and still learning, I find threads like this very beneficial - both for me and other "noobs" that will eventually stumble upon it.
Just a few things I've noticed while tuning my system:
1. In my system, I've never ran into a frequency that doesn't respond normally to EQ. By "normally", I mean that adjusting the EQ by 1 dB (for example) causes 1dB of difference when I take the next measurement. I've never had something NOT respond to EQ? How can that be??
2. So far, I've just been EQ'ing each individual speaker to my customized target curve. Basically, I generated my target curve based on the overall EQ "shape" that works for me, with my hearing, in my vehicle, with my speakers - after lots of testing. The common curves just did not do it for me. But I place more importance on matching left/right than I do on matching the curve itself.
3. So far at least, I haven't found the need to EQ "whole sides", "speaker pairs" or "overall" responses. After EQ'ing each speaker individually and then taking measurements with "both sides" playing (both front door speakers, for example), the results are basically the same as they were for each individual door speaker (just louder), so I just didn't see the need to EQ anything after I EQ"d each speaker individually.
I'm NOT a "technical" SQ person - I just go with what sounds good to me (for example, I tried going without rear speakers and just cannot do it) - I like the effect my band-passed, reduced-level rear speakers give me, right or wrong. I also have very inexpensive Kenwood Excelon speakers and have no plans on upgrading to more expensive speakers (at this point - who knows what the future may bring). I also use stock speaker locations and am not interested in customizing speaker locations - just not interested in going that far - trying to make the best of what speakers I have and their stock locations.
Next I want to introduce some pictures to this thread... I think they may help get the point of my question across a little better. I recently decided to try EQ'ing my system manually instead of using the REW auto-EQ function - mainly to keep the number of PEQ filters to a minimum to see how I liked the results. However, as I took each measurement, I found myself using more and more PEQ filters to better match left/right really well, so I still ended up using a lot of PEQ filters to get to where i'm at now (which I'm actually very happy with - but always looking to make it even better). My wife even commented on how good my car sounds compared to hers - and she was sitting in the passenger seat of a drivers-seat tuned car! :-)
Next post will includes some pictures of what I did along the way - curious to see what you guys think about my decisions.
Again, thank you all for your input - it will be beneficial to me and other noobs in the future!!
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
On note 3.
If you don't check speaker pairs, then you are leaving some on the table. But if you are happy with that, then it is fine. Without checking the pairs, I find that I can get about 90% of the way there. Right now I haven't had time to get that deep into my truck or my wife's car and have literally just used the Helix auto RTA on my truck (I did spend some time on crossover settings to get my electrical and acoustical crossovers to be where they need to be), and REW on my wife's car.
I would say both vehicles are about 90% there even though the helix auto rta only uses 31 band graphical eq. What I notice is that depending on the singers voice and certain instruments, the center images in both vehicles will drift some left or right. This is why you should check pairs using limited band pink noise with your ears! You also need to make sure that your mid/midrange/tweeter are lining up in the crossover as the overlapping frequencies are really the only place they will hurt/help each other . Some of this could be due to the actual volume level being more on one side, and some could be that the speakers are out of phase at that point that isn't due to volume.
You can check this with the microphone, it is harder to see with REW, especially if you can't do a sweep in REW. But if you have your left mid measured and the right measured, then you measure both together, all frequencies should be above the individual lines (I believe about 6db above as you doubled the power and doubled the speakers, someone please correct me if I am wrong). If it is only 2db above (or really anything less than 6db), you have phase issues that need to be addressed. They are close to being in phase, but they are not 100%.
Don't get me wrong. I understand the theories of all of this but am still honing my ear skills and still learning when to ignore REW results. Tuning my wife's car with only 10 bands of PEQ really helps this.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Jdunk54nl
On note 3.
If you don't check speaker pairs, then you are leaving some on the table. But if you are happy with that, then it is fine. Without checking the pairs, I find that I can get about 90% of the way there. Right now I haven't had time to get that deep into my truck or my wife's car and have literally just used the Helix auto RTA on my truck (I did spend some time on crossover settings to get my electrical and acoustical crossovers to be where they need to be), and REW on my wife's car.
I would say both vehicles are about 90% there even though the helix auto rta only uses 31 band graphical eq. What I notice is that depending on the singers voice and certain instruments, the center images in both vehicles will drift some left or right. This is why you should check pairs using limited band pink noise with your ears! You also need to make sure that your mid/midrange/tweeter are lining up in the crossover as the overlapping frequencies are really the only place they will hurt/help each other . Some of this could be due to the actual volume level being more on one side, and some could be that the speakers are out of phase at that point that isn't due to volume.
You can check this with the microphone, it is harder to see with REW, especially if you can't do a sweep in REW. But if you have your left mid measured and the right measured, then you measure both together, all frequencies should be above the individual lines (I believe about 6db above as you doubled the power and doubled the speakers, someone please correct me if I am wrong). If it is only 2db above (or really anything less than 6db), you have phase issues that need to be addressed. They are close to being in phase, but they are not 100%.
Don't get me wrong. I understand the theories of all of this but am still honing my ear skills and still learning when to ignore REW results. Tuning my wife's car with only 10 bands of PEQ really helps this.
Like I said, I actually have checked speaker pairs - it's just that I didn't see the need to EQ anything "in pairs" or "whole sides" after checking them - they are basically the same as the individual measurements, only louder. There is some *slight* variations when measured in pairs, but we're talking 1 or 2dB at most - and only at a few spots - which could even just be measurement variances (even individual speaker measurements vary some from measurement to measurement). I guess it wouldn't hurt to try and fix those few issues I saw when measuring speaker pairs, but that would just cause me to use even more PEQ filters. :-)
I'll post some pictures of what I have seen and done - and why - and see what you guys think - I'm always open to learning more and trying different things - and I'm *sure* this will be beneficial to others as well.
I still haven't tried the Helix auto-EQ yet - but do want to try it at some point - although, I have a feeling it will use even more filters than I do and end up with more left/right variance than I get with my manual PEQ filters. But hey, I've been wrong before! :-)
I've only tuned with pink noise. Have never even got into sweeps and looking at phase in REW - mainly because my pink noise measurements didn't seem to indicate any major phase issues (based on speaker pair measurements looking the same as individual speaker measurements).
I am pretty confident in my active crossover settings, speaker level settings and TA setting though - I definitely have a solid base that isn't causing me to use more EQ than needed. Originally, my crossovers weren't optimal - but out discussion in that Helix DSP.3 thread helped a lot and I was able to optimize my crossover settings to reduce the amount of EQ boosts that I needed.
The only thing that I haven't done yet - and really need to get done - is deadening my doors. At this point, I'm going to wait until summer comes around to help with installation (don't want to install it while freezing cold) - I know I could heat it up while installing, etc - but would rather just wait until summer - especially since I still need to get a very minor power window issue addressed before I do it.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
jtrosky
Like I said, I actually have checked speaker pairs - it's just that I didn't see the need to EQ anything "in pairs" or "whole sides" after checking them - they are basically the same as the individual measurements, only louder.
Just because they are louder, doesn't mean you don't have issues. The two could be lining up like this which would increase the overall volume but still slightly out of phase.
https://www.caraudiojunkies.com/imag...Zc5qC/IZCDAf/Zhttps://www.caraudiojunkies.com/imag...Zc5qC/IZCDAf/Zhttps://dosits.org/wp-content/upload...hase-1.500.png
https://qph.fs.quoracdn.net/main-qim...fdca22721a4b4c
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Jdunk54nl
I'm guessing this is what causes the summed response to be a little wonky compared to each side lining up when played solo? Stuff like this is what makes me bang my head against the wall when tuning:lmao:
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Jdunk54nl
Just because they are louder, doesn't mean you don't have issues. The two could be lining up like this which would increase the overall volume but still slightly out of phase.
It's not just that it's louder - the actual response is the same as well. So I get a particular response when playing a single speaker. When i play both left/right at the same time, I get the same response, except that the response is a few dB higher in the measurement graph. If there were significant phase issues, I would expect the response to be different at the point where the phase is problematic. Right?
This will be a lot clearer when I post some pictures to illustrate what I mean....
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
The idea of EQing each speaker, and then each side to address cross over regions makes sense.
i could imagine a case if the left hand side (LHS) tweeter had a different EQ, and particularly a different number of filters that the RHS... then there could be a phase difference between the LHS and RHS... so they would potentially sound good somewhere, and pull the stage in one band or another if the phase starts wandering off in a different range frequencies.
Maybe something like that is happening?
Quote:
Originally Posted by
blockrocker
...
Full disclosure though, I have not noticed tonnes of audible difference between many of the different REW/Helix setups I have experimented with. Despite extensive tinkering, as blindly as can be at times, I feel sometimes the tune goes great and other times it's a complete failure. Strangely, I cannot pinpoint why that is, still learning.
You are probably not alone. Going from "sounds different" and correlating it to the tune, is easier than finding the causal reason why that tune is making it sound different.
(And I am not pretending to know)
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
In post 25, I’m not understanding what the issue is if that is what is implied here. What does that slightly out of phase wave form picture represent in real world actual circumstances and in the tuning process? Is this something that is audible? Please give me an example? I feel like issues that are not really real world tuning issues are being brought up as issues leading to confusion here.
In post 28, stating that different EQ for each driver is causing phase issues from say the left tweeter to the right tweeter is not really an issue when it comes to what you hear after the tune. I’m not saying you are wrong here about phase changes but does it actually amount to an audible difference here? I’m strongly of the opinion that is does not. This is a new one for me and honestly I do not think this is even worth considering, and again will lead to confusion among those still learning the basics of tuning. These poor guys will be chasing their tales forever if they put any weight on that assumption.
I could be off here but based on my own experience it seems like these things are really not worth considering as issues when we follow a solid tuning process. We should set time alignment with a tape measure and forgot it forever after that. It’s just not worth fussing over fractions of an inch especially when you consider how much your head moves around during a drive and as Andy Wehmeyer pointed out in his blog. You set crossovers according to driver size (beaming) and capability to work within the passband that it was designed for and so that you are within the widest possible polar response of that driver (again below beaming). You use 24 db LR4 slopes to get the acoustical response to match the electrical response as closely as possible and so that drivers are in phase with each other and because 24 db LR4 crossovers are the easiest to implement. You set you HP filter for tweeters so that they are protected and 1.5 to 2 times the fs. You set the gain for the amplifier driving the subwoofer to clipping. You set the rest of the amplifier gains for mids/tweeters to your target curve using the amplifiers gain adjustment for those drivers according to what you like for bass level, which for me is a bass level 20 db above the midrange horizontal run or body of the target curve.You EQ each driver to a target curve and then EQ all left summed together while also checking for out of phase (polarity issues) drivers and then you match the right side so that it matches the left side as closely as possible so your stage is in the center. Measure everything together and make any final adjustments against your house curve and personal preferences by ear. You use real music you know well to test out your tune and also testing/tuning sounds like the one that comes with the AudioFrog UMI-1 (or similar) to verify your results and left, right, center. You make any final adjustments from left to right and with amplifier gains after EQ, if needed.
Hope this helps. Not trying to be confrontational here but I hate to see newcomers or members new to tuning with a DSP be possibly confused and steered into an endless abyss of confusion.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Think about as you adjust your time alignment. Your speakers probably didn’t start completely out of phase. As you adjust time alignment, you are taking them from a lot out of phase, to slightly out of phase, to being more in phase.
This impacts the centering of the sound. If it isn't really close to being in phase, your center will shift and the sound levels (volume) will change. How much and is it noticeable depends on how much out of phase. The lower the frequency, the more noticeable it would be.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
We'll talk about it tomorrow. It's hard to type out
Sent from my SM-G975U using Tapatalk
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
This is semi embarrassing to even post.....but for the sake of learning I am going to.
I don't have REW graphs but will just to check them out but I was paying attention to the Helix RTA graph and didn't notice anything really weird on it. I am guessing something will show up on REW when I check speaker pairs but will post again once I get the chance to get those graphs. I want to know if something would have shown up on REW...it should when playing pairs.
I threw my 3 way together, spent about 10 minutes waving a microphone in front of my face using the helix auto rta. I did left side together, right side together, sub, then everything together. I never checked individual speakers/pairs or did anything beyond that. I changed my process and didn't do other necessary things with the new process. According to the Helix RTA, my speakers summed greater than the separate sides.
Nick (skizer) got into my truck this past weekend and nearly instantly said your midbass and midrange are out of phase. He pulled up the helix software and went through flipping polarity and he was definitely right. It should have been an easy test and is normally in my tuning routine with REW. But the RTA didn't show anything weird and I didn't have the time to check. I should have known something was out of phase. Every time something is, my ears get the clogged feeling from listening to out of phase stuff.
Point is, check speaker pairs (by ear using mono/correlated pink noise) and....just because on the RTA it is louder...doesn't mean you don't have phase issues....
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
Jdunk54nl
This is semi embarrassing to even post.....but for the sake of learning I am going to.
I don't have REW graphs but will just to check them out but I was paying attention to the Helix RTA graph and didn't notice anything really weird on it. I am guessing something will show up on REW when I check speaker pairs but will post again once I get the chance to get those graphs. I want to know if something would have shown up on REW...it should when playing pairs.
I threw my 3 way together, spent about 10 minutes waving a microphone in front of my face using the helix auto rta. I did left side together, right side together, sub, then everything together. I never checked individual speakers/pairs or did anything beyond that. I changed my process and didn't do other necessary things with the new process. According to the Helix RTA, my speakers summed greater than the separate sides.
Nick (skizer) got into my truck this past weekend and nearly instantly said your midbass and midrange are out of phase. He pulled up the helix software and went through flipping polarity and he was definitely right. It should have been an easy test and is normally in my tuning routine with REW. But the RTA didn't show anything weird and I didn't have the time to check. I should have known something was out of phase. Every time something is, my ears get the clogged feeling from listening to out of phase stuff.
Point is, check speaker pairs (by ear using mono/correlated pink noise) and....just because on the RTA it is louder...doesn't mean you don't have phase issues....
So your midrange and midbass speakers were physically wired differently in terms of polarity? Is that why they were out of phase? Or was it a setting within the DSP that caused them to be out of phase?
Just curious how you got to that point.
I have one of those polarity-checking tools - very nice to have to make sure polarity is correct - but you really only need it when you physically mess with the speaker wiring - otherwise, you can check phase settings in the DSP. But for $10 off Amazon, it's a nice-to-have.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
jtrosky
So your midrange and midbass speakers were physically wired differently in terms of polarity? Is that why they were out of phase? Or was it a setting within the DSP that caused them to be out of phase?
Just curious how you got to that point.
I have one of those polarity-checking tools - very nice to have to make sure polarity is correct - but you really only need it when you physically mess with the speaker wiring - otherwise, you can check phase settings in the DSP. But for $10 off Amazon, it's a nice-to-have.
I have the polarity app which does the same thing and everything is electrically correct from that. Also every wire in the car for speakers is aftermarket and is easy to tell which is the positive side and which is negative. In other words, it is not an electrical issue but I do not know what would cause it. Someone that has more knowledge than me can chime in.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
I am actually willing to bet that I have had phase issues more often than I think. I am really starting to realize that when phase is good in my truck, I have ZERO issues with my ears (like right now). When I do have phase issues or am checking phase, my ears always get a clogged feeling.
I remember quite well many times when my ears would get this clogged feeling and I could never figure out what the issue was so I just re-did everything. I would bet money it was a simple phase issue now knowing what I know.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Seat at equal distance from 2 speakers playing the same music - everything is coherent and in phase
Seat closer to the left one and both speakers are playing the same music - you now have a timing issue and a level issue
Both equals to phase problems.
Starting with the correct polarity helps but is not the remedy to a phase problem. In a car, you rarely if ever seat at equal distance from both speakers. To make the matter worse, you have 5, 7, 12, 15 speakers playing different frequencies, some louder, all with added refletion, etc...
I get the same feeling when something is out of phase (clogged ears)
Kelvin
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
subwoofery
Seat at equal distance from 2 speakers playing the same music - everything is coherent and in phase
Seat closer to the left one and both speakers are playing the same music - you now have a timing issue and a level issue
Both equals to phase problems.
Starting with the correct polarity helps but is not the remedy to a phase problem. In a car, you rarely if ever seat at equal distance from both speakers. To make the matter worse, you have 5, 7, 12, 15 speakers playing different frequencies, some louder, all with added refletion, etc...
I get the same feeling when something is out of phase (clogged ears)
Kelvin
But isn't that what time alignment is for (the speakers being at different distances)? I'm sure he had time alignment setup (and I'm assuming correctly since he didn't mention having to change that).
Although - @Jdunk54nl do we know exactly what was changed to "fix" the phase issue? If so, that would probably tell us what the problem was.... :-)
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Polarity was swapped in the dsp software for midbass. Beyond that I am not 100% sure what Nick changed.
yes time alignment was set via tape measure.
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
I'm really curious to hear if anything else was changed. I can't think of any reason why a complete polarity flip on the DSP would fix the issue unless the midbass speakers were physically wired with the opposite polarity of the midrange speakers - but you mentioned that they were definitely wired the same... I'm really curious to find out if anything else was changed to fix it - as well as how to easily identify the issue to begin with (which was the main point of your post).
-
Re: Why is "beating speaker frequency response into submission" via EQ a bad thing?
Quote:
Originally Posted by
jtrosky
I'm really curious to hear if anything else was changed. I can't think of any reason why a complete polarity flip on the DSP would fix the issue unless the midbass speakers were physically wired with the opposite polarity of the midrange speakers - but you mentioned that they were definitely wired the same... I'm really curious to find out if anything else was changed to fix it - as well as how to easily identify the issue to begin with (which was the main point of your post).
It was easy for me to hear when you played pairs of speakers with correlated/mono pink noise. Nick picked it up without even doing that. I think it was the "lack" of midbass he noticed. I know he did do some further EQ work, but you could tell a HUGE difference just by flipping the polarity of the speakers.