ProSoundWeb Community

Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => LAB Lounge => Topic started by: Josh Billings on November 08, 2019, 08:57:45 PM

Title: What should a Sound System frequency response look like?
Post by: Josh Billings on November 08, 2019, 08:57:45 PM
I'm in the process of dialing in our dance music Sound System, but it sounds way too bright if I have it EQed flat. I'm accustomed to much more sub-bass than mids and highs.

I'm wondering what you guys do when EQing a sound system with a measurement mic? Is there like a standard practice here for prerecorded / DJ music shows?

Josh Billings
Title: Re: What should a Sound System frequency response look like?
Post by: Ivan Beaver on November 08, 2019, 09:12:30 PM
I'm in the process of dialing in our dance music Sound System, but it sounds way too bright if I have it EQed flat. I'm accustomed to much more sub-bass than mids and highs.

I'm wondering what you guys do when EQing a sound system with a measurement mic? Is there like a standard practice here for prerecorded / DJ music shows?

Josh Billings
I would start by asking what is your measurement platform?  True dual FFT or RTA?

Is the whole response flat? or flat above a certain point.

If you look at the Fletcher-Munson curves, you will see that it takes a lot more low end to have the same "loudness" as the higher freq.  So there usually needs to be a "haystack" on the bottom.
Title: Re: What should a Sound System frequency response look like?
Post by: Josh Billings on November 08, 2019, 09:44:47 PM
So you think something like this should be the overall goal, especially since it's dance music???

I would start by asking what is your measurement platform?  True dual FFT or RTA?

Is the whole response flat? or flat above a certain point.

If you look at the Fletcher-Munson curves, you will see that it takes a lot more low end to have the same "loudness" as the higher freq.  So there usually needs to be a "haystack" on the bottom.
Title: Re: What should a Sound System frequency response look like?
Post by: duane massey on November 09, 2019, 01:08:03 AM
I have NEVER heard a system that was set "flat" via measurements that sounded good. After you get a starting point you should use your own ears and taste to set the system to your own desires. In a dance system you typically don't have the feedback issues that a live system should have, so don't read the graph, close your eyes and make your ears happy.
Title: Re: What should a Sound System frequency response look like?
Post by: Russell Ault on November 09, 2019, 02:14:56 AM
I'm wondering what you guys do when EQing a sound system with a measurement mic? Is there like a standard practice here for prerecorded / DJ music shows?

Try to think of a sound system like a painter's canvas. A white canvas (i.e. one that reflects all wavelengths of light basically evenly) is typically used because it allows the painter to easily do whatever they want with it. Painting a sunset on a pink canvas might be easier than on a white one, but trying to paint open ocean is going to be harder if your canvas is pink, since you'll end up having to correct for the pink tones in all of the blues you're trying to paint. Using a white canvas means you don't need to know your subject before you buy the canvas.

A sound system that exhibits an equal amplitude per octave (i.e. "flat") response is exactly like the white canvas. By having a clearly-defined neutral starting point you can do anything with it. This is especially useful for sound systems that have to do several different types of shows (e.g. with an equal-amplitude system you don't have to do drastic things on the channel strip to keep your talking head from sounding like a beatboxer).

Of course, very few people want to stare at a white canvas all day. Most live sound systems that are tuned flat will than have some low frequency boost added. Whether this is added in system processing or in the console is a matter of preference and use-case, but I will say that doing this sort of "toning" from the console gives the operator much more flexibility.

Here's what I'd suggest trying: tune your sound system flat (as Ivan said, make sure you're using proper dual-channel FFT measurement software and not just pink noise into an RTA); then, use the EQ for channel the music is coming through to make it sound the way you'd like it to. For dance music I'd guess a low shelf filter at ~80-100 Hz turned waaaay up. This way you're music will sound correct, but you won't spend nearly so much energy fighting to make announcements intelligible.

-Russ
Title: Re: What should a Sound System frequency response look like?
Post by: Chris Grimshaw on November 09, 2019, 03:02:13 AM

If you look at the Fletcher-Munson curves, you will see that it takes a lot more low end to have the same "loudness" as the higher freq.  So there usually needs to be a "haystack" on the bottom.

This feels like a misapplication of FM curves - they tell us about the sensitivity of our hearing vs frequency and SPL, and I can't for a moment see why that ought to factor in to the frequency response of a PA system.

For me, the only curve that sounds good is flat. Anything else is an FX box.

Chris
Title: Re: What should a Sound System frequency response look like?
Post by: Patrick Tracy on November 09, 2019, 03:08:46 AM
If you look at the Fletcher-Munson curves, you will see that it takes a lot more low end to have the same "loudness" as the higher freq.  So there usually needs to be a "haystack" on the bottom.

Sort of true if you're running the levels quite low. Loudness curves tell us how we hear at lower SPL relative to a higher SPL reference. Human hearing becomes less sensitive at the low and high ends referred to 1kHz when the levels are lower, though the effect is more pronounced in the LF.
Title: Re: What should a Sound System frequency response look like?
Post by: Mike Monte on November 09, 2019, 07:03:02 AM
I have NEVER heard a system that was set "flat" via measurements that sounded good. After you get a starting point you should use your own ears and taste to set the system to your own desires. In a dance system you typically don't have the feedback issues that a live system should have, so don't read the graph, close your eyes and make your ears happy.
+1
to add to the above: IMO you'd be best to use your ears and tune the rig to the room when there are people in the room.
Make your ears happy, then the patrons' ears will be happy.



 
Title: Re: What should a Sound System frequency response look like?
Post by: Mark Wilkinson on November 09, 2019, 08:07:33 AM
+2 to comments that say a FFT flat line response sounds exceedingly bright....

My sound curve preference.....a starting flat line response that tilts like a see-saw.
Fulcrum is the middle of the 10 octave spectrum, so in the 600-700Hz range.  Tilt to taste, raising the low end, while reducing the high end.
Add some haystack bass for certain types music/sound, if needed.
Simple, repeatable, fits outdoors, and most indoors.  Works 95% of the time for me...
Title: Re: What should a Sound System frequency response look like?
Post by: Keith Broughton on November 09, 2019, 09:14:37 AM

Fulcrum is the middle of the 10 octave spectrum, so in the 600-700Hz range.  Tilt to taste, raising the low end, while reducing the high end.
Add some haystack bass for certain types music/sound, if needed.
Simple, repeatable, fits outdoors, and most indoors.  Works 95% of the time for me...
Pretty much how I "tune" for a starting point.
Title: Re: What should a Sound System frequency response look like?
Post by: Mal Brown on November 09, 2019, 02:24:59 PM
+1
to add to the above: IMO you'd be best to use your ears and tune the rig to the room when there are people in the room.
Make your ears happy, then the patrons' ears will be happy.



 

This!
Title: Re: What should a Sound System frequency response look like?
Post by: Scott Bolt on November 09, 2019, 03:31:15 PM
I have NEVER heard a system that was set "flat" via measurements that sounded good. After you get a starting point you should use your own ears and taste to set the system to your own desires. In a dance system you typically don't have the feedback issues that a live system should have, so don't read the graph, close your eyes and make your ears happy.
THIS
Forget the graph!

I used to have a Drive Rack PA that had an auto-flat setup wizard.  It used pink noise and a flat mic to zero out the response curve of the system.  Problem was, it sounded awful when done (although it did show me that my old folded horns were putting out way too much mud and not enough thump).  I had to add to the bottom (Ivan's hay stack ;) ) to make it sound good again.
Title: Re: What should a Sound System frequency response look like?
Post by: dave briar on November 09, 2019, 03:54:13 PM
+1
Make your ears happy, then the patrons' ears will be happy.
This actually sums up my entire mixing philosophy. I’m really just selfish. I simply want to hear every last nuance clearly and distinctly. I consider myself lucky that patrons nearly always seem to agree. All the knowledge I gain here as well as analyses I employ are just tools to that end.  Not that a dual-channel FFT is useless, it’s just another tool.
Title: Re: What should a Sound System frequency response look like?
Post by: Bradford "BJ" James on November 09, 2019, 04:14:53 PM
Use the SMAART traces to expose system anomalies, or odd room/system interactions. If the system is accurately reproducing what is being input into it, then all is good. Final tuning by ear.
Title: Re: What should a Sound System frequency response look like?
Post by: Ivan Beaver on November 09, 2019, 07:16:52 PM
This feels like a misapplication of FM curves - they tell us about the sensitivity of our hearing vs frequency and SPL, and I can't for a moment see why that ought to factor in to the frequency response of a PA system.

For me, the only curve that sounds good is flat. Anything else is an FX box.

Chris
I never said that it should be the response of a sound system.

I was just pointing out the how much more low freq it takes to "sound the same".

Of course if you look around the 3-4K range, you will notice that our ears are very sensitive there  (It has been theorized that we are most sensitive there, because that is where a baby screams, so that we can hear the screams easier).

At low levels flat through that region is fine and adds detail and clarity.  But at higher levels, the system will sound "Bitty" and harsh.

So the tuning should also consider the SPLs that it will be producing.

Once again, it is not as simple as it would appear
Title: Re: What should a Sound System frequency response look like?
Post by: Ivan Beaver on November 09, 2019, 07:20:48 PM
So you think something like this should be the overall goal, especially since it's dance music???
Without a vertical scale, there is no way to guess what the actual curve would be.

It could be 1dB/division or 10dB.

Some people like a gentle rolloff of the highs above 1 or 2Khz, others like it flat.  Some like a boost.

It depends on their particular hearing, what their reference is etc.

Personally, when I am tuning a system, I will never add a boost on the high end, unless the customer requests it.
Title: Re: What should a Sound System frequency response look like?
Post by: Peter Morris on November 10, 2019, 07:40:02 AM
I'm in the process of dialing in our dance music Sound System, but it sounds way too bright if I have it EQed flat. I'm accustomed to much more sub-bass than mids and highs.

I'm wondering what you guys do when EQing a sound system with a measurement mic? Is there like a standard practice here for prerecorded / DJ music shows?

Josh Billings

Logically you would expect flat to sound perfect, but in practice it almost never does and sounds too bright as you noted.

There are various target curves out there that vary from about 3 to 6 dB roll off in the HF region from about 1KHz (+ LF boost) to the extremes of the X cure used for cinema applications. See below as an example of the -6dB curve.

I think there are various reasons for this – The vocal mic’s we use are not flat and have a high mid peak of about 5-6 dB combined with a proximity effect in the low end that we usually counter with a high pass filter; distortion in the HF compression driver; the phase response of the system; the power response of the system and the space it is used in.

Bottom line – expect to roll off the HF from about 1KHz and boost the bass if you want things to sound nice.
 


Title: Re: What should a Sound System frequency response look like?
Post by: Luke Geis on November 11, 2019, 12:12:00 AM
I think we often misunderstand the difference between what is meant by flat vs that of linear.

When you use an RTA, you are measuring the sound in a space and this includes reflections, nodes, comb filtering ( mic placement matters here ) and losses due to distance. So you EQ tills it's flat and it sounds like ass because the bass is low and the highs are exessive. The system is flat though..... The problem is that you don't have a reference. It is just noise in the venue that you assume is correct from the source and the mic picks up what is hears and you see a reading.

Linear is a different animal. Linear can mean two things. It can mean that what goes in, is what comes out ( which is the basic idea ) or it can mean that the system is devoid of large irregularities that are obviously not intended. Let's say that all you have is pink noise and an RTA setup. If you can get the RTA data from the source media you can compare it to the RTA data you get from your room mic. If there is a large hole around 250hz in your room reading, that does not correspond to your source media, there is a reason for that. You have to decide if it is mic placement, room modes, comb filtering or otherwise. But if your RTA reading is relatively smooth and tracks the source media pretty closely, chances are good that if it sounds good in the room, it is fine. Linear system response is ideal and the general basis is that it corresponds closely to what the source media is. The challenge to acquiring a linear system is a way to monitor and compare your source media to your measurement of the PA.
Title: Re: What should a Sound System frequency response look like?
Post by: Chris Grimshaw on November 11, 2019, 03:29:52 AM
The bit I don't get is that everyone seems to think a "flat" system sounds bad. I find they sound great, but of course my reference (home HiFi) is also flat so it's definitely what I'm used to.

Chris
Title: Re: What should a Sound System frequency response look like?
Post by: Miguel Dahl on November 11, 2019, 11:00:42 AM
The bit I don't get is that everyone seems to think a "flat" system sounds bad. I find they sound great, but of course my reference (home HiFi) is also flat so it's definitely what I'm used to.

Chris

Some years ago when I on occasion worked with philharmonics and other orchestras there was most often this one guy who mixed them. He came in early and set up his measurement rig and he was after a flat profile, like flat from bottom to top. Mic flat on the ground and sweeped and pointed at the screen and said "make this flat please".

He did this with acoustic bands also. It always sounded like a million dollars. It was how he liked to have the system set up before he started turning and pressing knobs.
Title: Re: What should a Sound System frequency response look like?
Post by: Frank Koenig on November 11, 2019, 12:38:46 PM
The bit I don't get is that everyone seems to think a "flat" system sounds bad. I find they sound great, but of course my reference (home HiFi) is also flat so it's definitely what I'm used to.

I think you are in the bass-averse minority :) Pretty much everyone in my experience, whether involved with pro-audio or not, seems to like a substantial bass boost when playing back commercially mastered music. This is irrespective of the playback level or style of music. When soft they want boost so they can even hear the bass, and when loud they want exaggerated bass to flap their trousers. I usually use a low shelf starting at ~500 Hz and  ~+8 dB at 100 Hz.

The exception is cinema. Movies mastered for home viewing always seem to have plenty of bass (low frequency effects?) and I play them back flat to get better vocal clarity.

My theory is that this goes back to phonograph recordings that had limited dynamic range at low frequencies and has persisted, perhaps because of the limited low frequency capability of small consumer payback systems (it sucks when the plastic box rattles). The folks who want bass, and have a system that can reproduce it, turn it up. I'd love to hear about this from the point of view of any mastering engineers here.

Now when it comes to developing speaker settings, I'm all for flat. White canvas and all that.

--Frank
Title: Re: What should a Sound System frequency response look like?
Post by: Brian Jojade on November 11, 2019, 01:36:44 PM
Recorded vs live are very different things.  Personally, I like to tune my overall system fairly flat, then adjust each thing in the mix appropriately.  That being said, I rarely rely on measurement mics for this anymore.  I can simply listen to the system and get it close enough for me to work with it. If it's off by a dB here or there, it doesn't much matter, since I'm compensating for things in the mix anyway.

For music playback, it's kind of interesting that a system tuned flat does seem a bit, well, flat.  You'd think that studios would be tuned flat and want the music to sound good on a flat system.  But they do seem relatively consistent at least.
Title: Re: What should a Sound System frequency response look like?
Post by: Eric Snodgrass on November 11, 2019, 07:58:25 PM
For music playback, it's kind of interesting that a system tuned flat does seem a bit, well, flat.  You'd think that studios would be tuned flat and want the music to sound good on a flat system.  But they do seem relatively consistent at least.
It depends on the music and also how and where it was mixed.  There are some well-known studio mixers who mix some projects using headphones and laptops.  I've read articles where they say they've mixed while traveling in airplanes or staying in hotel rooms.  Doing that takes the studio room out of the equation.  So they're mixing to the sound of their headphones and possibly whatever room simulation plugin they've put on the master bus. 

When I'm EQing a system I have never had a goal of creating a "flat" system tuning.  I'm just trying to find and lessen frequencies the p.a. is exciting in that particular space.  If I don't have the luxury of time to use a measurement system (which is 98% of the time), I'm using whichever type of "money" microphone that will be used the most (wireless handheld, lectern mic, lavalier, headset mic, etc.) to hear how the system is reacting to the acoustic space and correct any frequency anomalies I hear.  I aim for vocal clarity.  That method for me leads to good results for everything else played through the p.a., because I've learned that if a vocal mic is exciting frequencies around 150 in a room then everything else that can reproduce that frequency will also excite that frequency.
I've found that most new speaker systems that I've heard, from mid-level MI powered boxes to high-end arena systems, sound pretty darned good right out of the gate (thanks to manufacturer DSP and tunings).  So when I'm mixing FOH I'm finding it extremely rare these days (to the point of being almost non-existent) when I'm tweaking EQ for the mains because of the sound of the speaker itself. 
It's great when I do have the luxury of time to make measurements in a room.  It helps me pinpoint any issues with the reaction of the room.  I never expect a certain shape to the frequency response graph. 

So, the TL;DR - a sound system frequency response chart should look exactly like how the sound from the speaker system is reacting within whatever environment it is placed. 
Title: Re: What should a Sound System frequency response look like?
Post by: Kevin McDonough on November 12, 2019, 05:54:39 AM
I think we often misunderstand the difference between what is meant by flat vs that of linear.

When you use an RTA, you are measuring the sound in a space and this includes reflections, nodes, comb filtering ( mic placement matters here ) and losses due to distance. So you EQ tills it's flat and it sounds like ass because the bass is low and the highs are exessive. The system is flat though..... The problem is that you don't have a reference. It is just noise in the venue that you assume is correct from the source and the mic picks up what is hears and you see a reading.

Linear is a different animal. Linear can mean two things. It can mean that what goes in, is what comes out ( which is the basic idea ) or it can mean that the system is devoid of large irregularities that are obviously not intended. Let's say that all you have is pink noise and an RTA setup. If you can get the RTA data from the source media you can compare it to the RTA data you get from your room mic. If there is a large hole around 250hz in your room reading, that does not correspond to your source media, there is a reason for that. You have to decide if it is mic placement, room modes, comb filtering or otherwise. But if your RTA reading is relatively smooth and tracks the source media pretty closely, chances are good that if it sounds good in the room, it is fine. Linear system response is ideal and the general basis is that it corresponds closely to what the source media is. The challenge to acquiring a linear system is a way to monitor and compare your source media to your measurement of the PA.

this is the most sensible post in the thread so far.

Firstly, no one wants a "flat" system in terms of the exact same SPL for every frequency 20hz to 20khz. As many people have said it'll sound too bright and lacking bass. But a "flat" system and a flat transfer function are two very different things.

I think also, I rarely have a single system tuning I'll use, it often depends on the genre of music.

In all of the work I do, tuning a sound system usually comes in 3 stages:

1) measure each band of speakers (usually the tops and then the subs, but maybe also mid bass cabs as well if you have specific cabs for that etc) using the transfer function of Smaart, Systune etc. At this stage I'm looking to check that they're broadly flat, and that there are no obvious holes or dips. If there are try and see if it's a room reflection, positioning issue, etc etc and see if it's solvable.

2) Compare the crossover region where these different speaker bands meet, make sure they're delayed and in phase as much as possible.

Once I know that the elements of the sound system are all operating correctly then....

3) Tune to taste.

this is the most important step. The other two are just checking everything is working correctly and you have a blank canvas to start with, this step is now setting the system for the show. And here I'll be using my ears as much as the measurement system.

Firstly I'll play some of my common test tracks. These will check that the system sounds broadly OK and I'll make a few level/EQ adjustments here. Everyone loves to rhyme off their favourite test tracks, but the actual track itself is less important, what's most important is that you know it inside out, back to front, and know exactly how it should sound, that way if it doesn't sound like that you can quickly pick out any issues. Generally I'll probably put a bit of a boost on the subs, maybe pull out a few db around 2k. If it sound boxy I'll pull out a little around 400-800hz, boomy usually around 200hz. In these cases I will sweep the eq back and forth a bit until i find the sweet spot that evens it out. Will also listen to the top end and decide if I need to make any EQ changes there. And of course the room makes a big difference, the same sound system might have quite different EQ depending on the room.

Also, get to know your voice! I know most people hate the sound of their voice, but getting used to it is a big thing for system tuning. A big part of your tuning regime should be to talk through the PA with whatever vocal mic your artist(s) will be using, and make sure it sound natural and even at the tuning stage. Obviously with EDM or something this is less important, but assuming the vocals are a big part of the vocal sound, spending the time at this stage making sure your voices sound clear and natural in the PA will save a pile of hassle when it comes to actually sound checking and doing the show with the act.

But secondly I'll play some tracks that are relevant to the genre of the show, and probably adjust a little more from there. If I'm doing an acoustic/folk music show the subs might only be boosted a touch above the main PA. A rock show usually 6db or so. Big bass heavy hip hop might have the subs 20db up on the mains, and Dance/EDM music the subs at least 20db as well, but also probably a fairly noticeable scoop out the mids.


K

Title: Re: What should a Sound System frequency response look like?
Post by: Bill Meeks on November 13, 2019, 04:31:24 PM
Lots of good information from the various responders, but I take issue with the folks who say (paraphrasing) "...forget the meters and use your ears...".

The problem with that mindset, in my opinion, is those are YOUR ears. They are NOT the ears of your audience. What if you have been in this business for years and are old (I'm old, too, so take no offense please) and have listened at loud volumes for too long and your hearing is like -20 dB down at 10K compared to a young adult? Or what if you have hearing issues in other ranges and don't realize it? What sounds "great" to you might sound borderline awful to others.

Meters and analyzers are more objective. Yes, I understand all the details about room nodes, comb filtering, phasing and all that, but you can compensate for those limitations by taking several measurements around the venue and averaging. Certainly not perfect, but it helps.

Using just YOUR ears is akin in my opinion to letting a color blind person design a light show or choose paint colors for a room. The end result is not likely to match up with what the majority of non-color blind folks might choose.
Title: Re: What should a Sound System frequency response look like?
Post by: Caleb Dueck on November 13, 2019, 07:36:18 PM
Lots of good information from the various responders, but I take issue with the folks who say (paraphrasing) "...forget the meters and use your ears...".

The problem with that mindset, in my opinion, is those are YOUR ears. They are NOT the ears of your audience. What if you have been in this business for years and are old (I'm old, too, so take no offense please) and have listened at loud volumes for too long and your hearing is like -20 dB down at 10K compared to a young adult? Or what if you have hearing issues in other ranges and don't realize it? What sounds "great" to you might sound borderline awful to others.

Meters and analyzers are more objective. Yes, I understand all the details about room nodes, comb filtering, phasing and all that, but you can compensate for those limitations by taking several measurements around the venue and averaging. Certainly not perfect, but it helps.

Using just YOUR ears is akin in my opinion to letting a color blind person design a light show or choose paint colors for a room. The end result is not likely to match up with what the majority of non-color blind folks might choose.

Agreed. 
Along with this, knowing how to use Smaart/Systune allows you to tell why you like or dislike a particular voicing.
Title: Re: What should a Sound System frequency response look like?
Post by: brian maddox on November 13, 2019, 11:01:03 PM
So here's the thing...

Sound Sources [musical instruments, etc] vary WILDLY in their sonic signature depending on their construction, their design, how they are played, how loudly they are played, the direction they are facing when played, etc.  To capture this widely variable source we use...

Microphones, which vary WILDLY in their sonic signature depending on their construction, their design, what direction they are facing, how loud the source they are capturing is, and the electrical characteristics of the system they are plugged into. 

We then run that signal through a system of filters, attenuators, gain stages, various pieces of copper, aluminum, silicon, etc. all of which have an effect on the sound that is then produced by our ...

Loud Speakers, which vary WILDLY in their sonic signature depending on their construction, their design, what direction they are facing, and how loudly they are reproducing the signals given them.  Which then excites...

A room full of Air that will vary WILDLY in it's sonic signature depending in it's temperature and humidity, the room's size and construction, it's design, the relationship of the room to the sound sources [which may be a mixture of reinforced and un-reinforced]. 

This Milieu is then captured by...

Ears, which vary WILDLY in their frequency response depending on Age, Size, Shape and that signal is then shipped off to

Brains, that vary WILDLY in how they perceive all of this information depending on biases, relationship to visual information, and a dizzying array of experiences that are literally different for every single human...

Yes, i'm stating the obvious [and ironically over simplifying].  But it bears remembering.  What we are attempting to do is 90 percent science, 90 percent art and another 70 percent Black Magic.  There is no definitive answer to the question "What should a Sound System frequency response look like?"  That doesn't make it a bad question.  In fact, it's a GREAT question.  But the closest we're gonna get ultimately to an answer is...

it depends
Title: Re: What should a Sound System frequency response look like?
Post by: Uwe Riemer2 on November 14, 2019, 04:46:37 AM
The array processing target from d&b is a good start.
Applied with reason and said technique from said manufacturer will give you this in more than one place.
Title: Re: What should a Sound System frequency response look like?
Post by: Randy Pence on November 14, 2019, 06:57:40 AM
Using just YOUR ears is akin in my opinion to letting a color blind person design a light show or choose paint colors for a room. The end result is not likely to match up with what the majority of non-color blind folks might choose.

I once worked at a club whose owner was colorblind. I only found out as I was doing some troubleshooting on the phone for him and he of course could not see any red lights in the rack, lol

He also did the interior design and architectural lighting. It was actually quite tasteful. Lots of brown woods, orange, and reds. Very little blue. I have no idea what he actually saw.
Title: Re: What should a Sound System frequency response look like?
Post by: Peter Morris on November 14, 2019, 07:47:30 AM
The array processing target from d&b is a good start.
Applied with reason and said technique from said manufacturer will give you this in more than one place.

These ....
Title: Re: What should a Sound System frequency response look like?
Post by: Douglas R. Allen on November 14, 2019, 09:15:36 AM
These ....

    Not sure if the graph will follow over or not in the quote.  I've seen many post through the years on this. Some wanting a bass hump. Some wanting a more or less flat response through the response range of the system.  When it comes right down to it doesn't what is to be coming out of the system really determine what is needed for a pre emphasis? Putting a big hump on the end may "sound" fine with recorded music but we are "Mixing" aren't we? Also as the Fletcher Munson curve changes with spl so at what spl do you want to see this curve?  You put in bass drum or lower toms or bass guitar into the board etc. Eq those items to make the PA sound pleasing ( we hope ) in the low end for the majority of the patron's in the room.  Does it make sense to add this low end then have to Low Cut everything that doesn't need the low end boost, even if its to remove the harmonics that this boost in the lows Main PA gives in channels that don't need it? Another way of looking at it is you have some inputs that don't have a lot of high end in them so you put a big shelving boost on the high end of the PA then have to High Cut many of your channels because now they are too bright.  ( Maybe setup Aux Fed Highs so these won't reach the horns....)
   I'd rather add bass to a few channels than take bass out of many channels. As we know eq causes phase issues so I guess I'd want the PA setup so I can use the least amount of eq to start in the most amount of channels for the type of inputs I was mixing. If I wanted my drums to be bass heavy I could put them on a Bus. Eq the Bus Master heavy on the low end then low cut the drum channels that I don't want to ride this eq. Same as the Main Eq curve but in a different location. But does this make sense?  ??? ;)

   There may never be a correct answer. I guess what works for the mix engineer and what keeps the overall audience and the band happy and keeps the money rolling is always the real solution.

Douglas R. Allen
Title: Re: What should a Sound System frequency response look like?
Post by: Tim Steer on November 14, 2019, 12:11:38 PM
There are a couple of different schools of thought when it comes to voicing a sound system, and which one you fall into broadly depends upon where you draw the line between art and science.

One group maintain that the mix engineer should be given a 'blank canvas' upon which to paint whatever sonic picture he/she wants. The job of the system engineer then, is to tune the system to give a 'flat' frequency response, and to ensure that all parts of the venue sound the same. If the mix engineer favours a bass-heavy sound, then this is an 'artistic' decision and one that should be done by the mix engineer on the desk.

Another group of system engineers will tune a system based on any number of other desired response curves - the assumption being that flat systems generally sound bright and lacking in sub-bass, and most mix engineers will invariably boost the low end and hack out the ear-piercing high mids/top end to make the system sound more musical - so why not make the system sound nice in the first place?

Can a skillful mix engineer make a mix sound good on either system? Of course. Which is the right approach? Depends who you ask, and what the programme material is. I know respectable and successful system engineers from both schools, and I've noticed there does seem to be a difference across the Atlantic divide. Living in the UK, I find most system techs tend to tune their rock systems along the lines of the D&B curves - something like flat from 1kHz upwards, with a 3dB increase per octave below 1kHz. There is a perception amongst many UK engineers of American systems being 'harsh-sounding' and aggressive, and I suspect this might be because more engineers and system manufacturers across the pond favour a flatter frequency response. I'd be interested to know if this is true.

Back to the OP's question... how should you voice your system for dance music? Most dance music contains a lot of synth sounds with unnaturally high levels of high-order harmonics. Most dance music fans expect the music to be played loud, as the enjoyment comes from the visceral experience of being immersed in the sound and feeling it. And the music tends to be 'all or nothing' - either the music is fully pounding, or it's in a breakdown, where the beat lets up for a while and builds into the next big drop. All of these things can lead to a very fatiguing sound when the system is too bright; it's one type of music that really does sound terrible on a flat system (those high frequencies are absolutely screaming when the system's running at 110dBA). I personally find that the 'flat but with massive amounts of extra sub' approach favoured by some sounds pretty offensive to my ears too - I've always felt that people who tune EDM systems this way probably don't like or understand the music, and just think that if you give the kids lots of bass then they'll be happy. I find that more of a gradual rolloff from 100Hz upwards all the way to 20kHz works well, and this is what I do with my dance music systems; it sounds loud and the drums have plenty of punch and energy, but the highs aren't damaging peoples ears when the system's being run at high level.

Whether you make these adjustments on the desk or in the system is up to you, but I'd suggest that if it's a club system and you're only ever running DJ mixers into the console, then making the changes in the system might be the best approach.
Title: Re: What should a Sound System frequency response look like?
Post by: Helge A Bentsen on November 14, 2019, 03:15:59 PM
I'm in the process of dialing in our dance music Sound System, but it sounds way too bright if I have it EQed flat. I'm accustomed to much more sub-bass than mids and highs.

I'm wondering what you guys do when EQing a sound system with a measurement mic? Is there like a standard practice here for prerecorded / DJ music shows?

Josh Billings

Today it looks like this.
Have no idea how it measures, it’s a walk-up gig tonight.
But the lead vocal mic sounded good with that eq, so I’m keeping it for this gig.
Title: Re: What should a Sound System frequency response look like?
Post by: Scott Helmke on November 14, 2019, 03:28:35 PM
I generally go with a measured-flat system from the very top down to maybe 60-80Hz, and then let the subs be louder because that's usually what people want.

But if you really want to give yourself something to think about, take a look at the channel EQ's on the mixer before you start. If you notice that almost every channel has a similar shape to the EQ, such as all the low mids carved out, then you have a pretty good guide to what they *don't* like about the sound of that system.
Title: Re: What should a Sound System frequency response look like?
Post by: Peter Morris on November 14, 2019, 05:44:51 PM
Lots of people have said they want a blank canvas to start with. 

i.e. If you voice the system with a (flat) mic it should sound perfectly like you, the problem when I try this with a system that measures flat (amplitude response) it usually sounds thin and harsh.

The question is why … logically you would expect it to sound perfect.

What I have noticed is that some of the newer systems with better low distortion drivers and a (FIR processed) relatively flat phase responses seem to need less bass boost / HF cut to sound right.

Its still work in progress but I processed one of my boxes to have a phase response that is relatively flat to about 100Hz …. and its sounding just right when flat.  The question I have to determine is it the phase, my imagination or something else ???
Title: Re: What should a Sound System frequency response look like?
Post by: Bill Meeks on November 14, 2019, 07:57:27 PM
A free product that I started using several months ago to monitor my own mixes is the Voxengo SPAN spectrum analyzer plugin for use in DAWs (Digital Audio Workstations). The plugin is a free download. There is a pay version with some extra features, but the free one is enough for most users.  Here is a link: https://www.voxengo.com/product/span/ (https://www.voxengo.com/product/span/) (I have no affiliation with the vendor).

I use a PC with the Reaper DAW installed on it along with Dante Virtual Soundcard to record tracks from our services. I have a Dante patch configured to send the FOH audio from the mixer to a stereo Reaper track. I put the Voxengo SPAN plugin on that track within Reaper. I have the SPAN plugin set to Average mode with 1/3 octave smoothing. I can then take a glance now and then at the Reaper screen to see the overall spectral balance of the mix. You can set the smoothing to several different values, but I find 1/3 octave works well for what I am monitoring for. I look to see if some particular frequency zone is sticking out too far. I'm still training my ears, so I use the visual from SPAN to validate what I think I'm hearing.

The SPAN plugin has a default 4 dB roll off or slope that is designed to compensate for the way our ears hear high frequencies. This translates within the plugin to showing you a relatively flat frequency curve when things are "right" in terms of generally accepted mastering curves. To test, you can play what you consider to be a well recorded and mastered reference track through the SPAN plugin and watch what the spectrum curve looks like. It should look pretty flat if you set the SPAN smoothing to 1/3 octave or 1 octave.

So back to the topic, I would prefer my house frequency response curve to be as flat as possible, then I make adjustments within my mix to get a properly balanced final sound. Using something like the SPAN plugin helps you validate with an objective "ear" that your mix should be right (at least in terms of spectral balance). Your turn bright "abrasive" sound producers down or tame them with EQ, and turn up lower frequency sound producers or enhance them with EQ. That's what a recording engineer would do, and when it comes down to the basics, a live sound mix engineer is fundamentally doing the same thing as a recording studio mix engineer:  making the mix "sound good". Of course the live sound guy has more issues to deal with since he has no perfectly treated recording studio to work in.
Title: Re: What should a Sound System frequency response look like?
Post by: Tim Steer on November 15, 2019, 03:25:18 AM
Your turn bright "abrasive" sound producers down or tame them with EQ, and turn up lower frequency sound producers or enhance them with EQ. That's what a recording engineer would do, and when it comes down to the basics, a live sound mix engineer is fundamentally doing the same thing as a recording studio mix engineer:  making the mix "sound good". Of course the live sound guy has more issues to deal with since he has no perfectly treated recording studio to work in.

That's assuming the OP even has a sound engineer. It sounded like he was describing a nightclub setup. Many smaller clubs don't have a full-time desk operator, and some don't even have a sound desk if the only thing that ever gets connected to the system is the DJ mixer in the booth. I can't think of any EDM club installations that are tuned flat - they're generally tuned to sound good with dance music.
Title: Re: What should a Sound System frequency response look like?
Post by: Bill Meeks on November 15, 2019, 08:28:36 AM
That's assuming the OP even has a sound engineer. It sounded like he was describing a nightclub setup. Many smaller clubs don't have a full-time desk operator, and some don't even have a sound desk if the only thing that ever gets connected to the system is the DJ mixer in the booth. I can't think of any EDM club installations that are tuned flat - they're generally tuned to sound good with dance music.

No argument from me there. And there are plenty of folks (particularly younger ones) that love a sound that is 90% composed of sounds below 160 Hz and anything above that frequency is like -20 dB down or something compared to the bass ...  :).

Personally that has never been my idea of sounding good. But in my younger days I've gotten in a friend's car and the radio (remember when we actually used those in cars to listen to music?) would have the bass turned up to max and the treble turned down to minimum.
Title: Re: What should a Sound System frequency response look like?
Post by: Geert Friedhof on November 15, 2019, 10:56:47 AM
No argument from me there. And there are plenty of folks (particularly younger ones) that love a sound that is 90% composed of sounds below 160 Hz and anything above that frequency is like -20 -10 dB down or something compared to the bass ...  :).

Fixed that for ya.
Title: Re: What should a Sound System frequency response look like?
Post by: Bill Meeks on November 15, 2019, 06:18:38 PM
Fixed that for ya.

Yeah, the "-20 dB" number was meant to be a tongue-in-cheek exaggeration.
Title: Re: What should a Sound System frequency response look like?
Post by: Ivan Beaver on November 15, 2019, 07:21:36 PM
Consider these couple of situations.

When used for playback, often people want the HF to be rolled back a bit, a sloping downward response, and a bass hump.

But why?  Why would you want to alter what the original artist did?  They mixed it so that it sounded right to them?

But if you roll off the highs and boost the bass, you are altering the sound that the artist spent a lot of time working on.

You should want something that is a "white canvas" so that the picture the artist intended is accurately displayed.

Who are you to change what they did/wanted you to hear?

In a live situation, you have control over the tone of each and every microphone.  So if the sound system is flat (a white canvas), you can accurately "paint" the tone/sound that you want.

If the canvas is already colored, then you have to "counteract" that color/response in order to get the sound you are looking for.

I go for a flat response, and love it when the engineer says to me "WOW, I didn't have to fight the system".

To me, the sound system should "disappear" as much as possible, and not present any "sound" of its own.

But I know some people like for a speaker system to have a particular "sound".  I guess it makes it easier for them to pick out different systems.  I prefer it to not have a "sound" and to be as transparent as possible.

Maybe that is just me
Title: Re: What should a Sound System frequency response look like?
Post by: duane massey on November 16, 2019, 12:42:09 AM
Ivan, back in the 70's when we were building massive front-loaded horn systems we always tried (and mostly succeeded) in delivering a system that was accurate even at extremely high SPL's. What went in came out. So many systems, even today, just don't meet that criteria.
Title: Re: What should a Sound System frequency response look like?
Post by: Nathan Riddle on December 05, 2019, 08:31:28 AM
I haven't had a chance to read the thread.

But this is a trace from an EDM/Dance event my company provided sound & lights for.

The traces are from two times during the night. The mic is midroom on a balcony ~40ft from the mains.

Sounded really good with a +20dB sub haystack :)