Frequency Response/Contour EQ in full range systems.

Phil Graham

Honorary PhD
Mar 10, 2011
651
1
18
Atlanta, GA
Re: Frequency Response/Contour EQ in full range systems.

I keep seeing a discrepancy between what the sytems measures and what the ear hears in this discussion. They are two different things.
They can be different, or not so much. See my reply above on SMAART 7 vs. Systune.

It doesn't make sense to me to claim that because the ear does not hear flat, you shouldn't tune the system flat.
Not only do ears not hear flat, they hear completely differently at different volume levels. No one here is suggesting that we create system curves that are the inverse of the F-M or ISO226:2003 curves, and even if we did this, what average level would we equalize to?

Mixing boards are limited devices. Limited in the amount of filtering, and limited in the degree of skill with which operators can apply those filters. To me it seems only reasonable to "build in" certain aspects to the response that are likely to be needed, regardless.

Hypothetically, lets say that all humans had extreme pain at 8kHz because of our ear physiology. In this case would it be reasonable to say "I'll tune the system flat, and the operator can just remove 8kHz with an EQ." To me, this would be a waste of an eq band, and foolish resistance against the hypothetical human anatomical bent. What I, and the other non-flat tuners in this thread, are suggesting is merely a softer form of this logic.

I think there are a few aspects to the human listening experience we can all agree upon:
  1. High frequencies are absorbed by air, we expect some amount of this. A system dead flat to 20kHz at a given point in space is preternaturally bright for most listeners. We all innately expect some VHF rolloff. Even a loudspeaker that is flat to 20kHz quickly has some built in VHF rolloff applied as sound makes it way to the listener.
  2. Real life objects, other than loudspeakers, have very different power response at low frequencies than high frequencies. More energy is radiated out in all direction at mid and low frequencies than at higher frequencies from every objects and instruments.
  3. The lack of directivity control at lower frequencies also extends to everday objects, so the low and mid energy arrivals from those objects interact more with the environment, and get more strongly included into our picture of the sound.
  4. At low and low-mid frequencies our brains develop to include more of this radiated energy in our hearing perception, therefore the total energy our brains "integrate" over is longer in the low and midrange that at high frequencies.
Herein, to me at least, lies the innate pull for some degree of tapering of the energy the system puts on the audience from the low frequencies to the high. Real objects dump out more low frequency energy in more directions, HF is absorbed somewhat, and this is what our entire existence on the planet is conditioned towards. It is no wonder then that some aspect of this "sloping" sounds "right" to many people who listen critically.

The L-Acoustics posted response gets some people riled up because it looks like it has extra mid, low, and low-mid boost. Really this is because of the location of the y-axis zero line. Move the line up a little bit, so the HF instead looks "softer" and fewer people would question the curve. I personally don't try to match a curve that looks like this one, but their assertions on the overall response taper have definite merit.

I also think it is helpful if the SE can build in flexibility to the stimulus curve so that it can be adjusted in reasonable amounts in either direction. A system that is tuned to the extremes of its capabilities lacks that flexibility.
Agreed.
 

Langston Holland

Sophomore
Jan 13, 2011
222
0
16
Pensacola
Re: Frequency Response/Contour EQ in full range systems.

This thread has been as enlightening and fun as the best the old PSW forum had to offer - thanks to all who've participated. :)

To summarize, Helge asked what he probably thought was an innocent question about biasing a PA's frequency spectrum to something other than flat. My take on the implication is that once biased, the majority of musical sources would require a minimum amount of channel EQ from the FOH console to achieve the best subjective result. Benefits would include a quicker/easier FOH soundcheck as well as a better end result from time and/or skill challenged operators.

Measurement systems and their various time and frequency windows got honorable mentions as needed - how can we objectively communicate or create repeatable results without measurements?

An Observation:

A guy named John Murray posted on the SAC list in June 2004 that Japan's TOA engineers taught him how to flatten the phase response of a loudspeaker using all-pass filters. His post was a mistake because I stalked him with emails and phone calls to tell me how to do it until he relented. In one of my conversations with him he mentioned how he was able to EQ a PA so that it sounded good with TEF (generally excludes later arrivals from the room), but was never as happy with the results he got when using Smaart (generally includes later arrivals from the room). I was quite happy with the results I got with Smaart but wasn't able to do consistently as well with my other system at the time - MLSSA.

A Thought:

It took me about 5 years and thousands of measurements to figure this out, but your brain can largely compensate for the weaknesses of your measurement platform! You look, you listen, you learn.

I now use three measurement systems regularly in their specific areas of strength. I know how to get the same results across all platforms in most situations, but I've found that near-anechoic responses at mid and high frequencies with near-RTA responses below that are most helpful in system tuning. The polar patterns of things that make sound, ensuing environmental effects and the human perception of that soup nicely explained by Phil led to the development of the decimation and/or multiple windowing techniques available in Smaart, SysTune, ARTA, SIM, MLSSA, Praxis and probably others.
 
Last edited:

Nick Hickman

Freshman
Apr 21, 2011
31
0
0
100dB.com
Re: DFT, energy arrivals, and when to use SMAART vs. Systune

Hi Phil,

Nick: For anyone interested, my summary of the topic is that the time window inherent in the DFTs in Smaart will not exclude late-arriving energy (reflections and reverb) from the transfer function magnitude display.
This is not entirely true. The energy is "excluded" by the nature of the averaging of uncorrelated signals across multiple frames. The shortest TC for Smaart 7 is approximately 7ms, according to Jamie. The more averages one does on the uncorrelated energy, the greater the SNR between the first frame arrival and the later frames. Within the TC of each frame, the uncorrelated late energy will increasingly be averaged to the background.
My first thought was that, to the extent that the uncorrelated energy varies between frames, yes, averaging will help, but that the same averaging can also be done in an RTA mode where the results of the DFT frames are being shown directly without doing the TF division by the DFT of the reference signal.

However, I realise the TF is not literally calculated as B/A where A is DFT of input and B is DFT of output but that what's actually calculated is (A*/A*)(B/A) or (B*/B*)(B/A), where A* and B* are the conjugates, giving either cross-spectrum/power-spectrum-A, or power-spectrum-B/cross-spectrum*. I understand, in the presence of noise, the two results will be different (and it's the difference between them that's the source for the coherence result). The paper I just checked suggests that the former estimator is better than the latter in the presence of output noise. Unfortunately, I don't have a good enough grasp of this to get a clear view of the improvement rendered relative to averaging a single DFT result. I shall attempt to study further!

I have to say, though, that in the practical examples I've attempted to set up, any immunity to late energy in the averaged TF has been subtle (relative to a similarly-averaged RTA view). My earlier example, for instance, used bucketloads of averages.

Nick: People often castigate the humble real-time analyser for being "time blind", but Smaart's transfer function magnitude is equally time blind unless you specifically enable a time window (and then it will perform an IDFT to get the system's impulse response, truncate it at the specified time, and perform a DFT back to give the windowed magnitude response).
This is utterly untrue, at least by the conventional definition of time-blind in the context of system measurement. An RTA integrates the energy of arrival over a given amount of time, but completely discards the phase information within each arrival bandwidth. You cannot capture the phase, and use it for your advantages, on an RTA. Because SMAART calculates a true H(f,t) by dividing against the reference signal.

If you have a late reflection in SMAART, its energy will manifest itself it both the phase and coherence traces, and those skilled in the art can quickly categorize the reflection as such. Similarly, if you have spectral contamination in the space from spurious noise, this can be readily ascertained. SMAART is not always the best tool for maximum noise immunity, but to say that is time blind is fallacious.
Hang on, I didn't say that Smaart was time blind: obviously, there's a phase display (and phase is time), coherence (that will reflect the presence of uncorrelated energy), and impulse response (showing system response entirely in the time domain), none of which exist in an RTA, and I agree with what you say about them. I said that the transfer function magnitude display is time blind if you don't enable an explicit time window (what you're calling a "secondary" window). Subject to your caveat about averaging, I still believe this to be the case.

I wasn't at all seeking to challenge either the cleverness or the usefulness of the dual-channel FT analysis method (or of Smaart), just what seems to be the widespread misunderstanding that the FT time window will inherently serve to exclude energy arriving after the length of the window from the measurement.

PS, Jamie assures me this "secondary" time windowing of the impulse response is most certainly in the works for SMAART 7, with no inherent limitations to implement it.
Not having used it, I didn't know that this wasn't yet in Smaart 7. Definitely worth implementing.

I've no disagreement with the rest of your post: good insights.

Cheers!

Nick
 
Mar 1, 2011
47
0
0
Vegas
Re: DFT, energy arrivals, and when to use SMAART vs. Systune

hey Nick,
you mentioned a paper in your reply, could you cite it please?
cheers
ferrit
 
Mar 1, 2011
47
0
0
Vegas
Re: DFT, energy arrivals, and when to use SMAART vs. Systune

hey Nick,
Thanks very familiar with that paper, Have you read any of Farina's papers? log chirps, silence sweep.
Cheers
ferrit
 

Phil Graham

Honorary PhD
Mar 10, 2011
651
1
18
Atlanta, GA
Re: DFT, energy arrivals, and when to use SMAART vs. Systune

However, I realise the TF is not literally calculated as B/A where A is DFT of input and B is DFT of output but that what's actually calculated is (A*/A*)(B/A) or (B*/B*)(B/A), where A* and B* are the conjugates, giving either cross-spectrum/power-spectrum-A, or power-spectrum-B/cross-spectrum*. I understand, in the presence of noise, the two results will be different (and it's the difference between them that's the source for the coherence result). The paper I just checked suggests that the former estimator is better than the latter in the presence of output noise. Unfortunately, I don't have a good enough grasp of this to get a clear view of the improvement rendered relative to averaging a single DFT result. I shall attempt to study further!
Nick, I don't know this either, at least off the top of my head, though I feel like I should. If Dave Gunness was around I'm sure he could clear it up succinctly.

Hang on, I didn't say that Smaart was time blind: obviously, there's a phase display (and phase is time), coherence (that will reflect the presence of uncorrelated energy), and impulse response (showing system response entirely in the time domain), none of which exist in an RTA, and I agree with what you say about them. I said that the transfer function magnitude display is time blind if you don't enable an explicit time window (what you're calling a "secondary" window).
My mistake, I totally glossed over the word "magnitude." And you're right about that, barring any trickery that we are both unaware of.

I wasn't at all seeking to challenge either the cleverness or the usefulness of the dual-channel FT analysis method (or of Smaart), just what seems to be the widespread misunderstanding that the FT time window will inherently serve to exclude energy arriving after the length of the window from the measurement.
I think this is a common misconception, yes, but I doubt anyone who groks how DFT on an on-going signal is set up would be confused. Certainly that group is only a small subset of measurement system users, though.
 
Mar 1, 2011
47
0
0
Vegas
Re: contour EQ

Hey Guys,
So as not to hi-jack the original post...
Has anyone noticed that some engineers are putting a "notch" in the system response around 125Hz?
I've got a couple of theories on this :-
1) this is where the typical lengths of arena arrays are losing pattern control so energy is being projected into the nether-regions
2) It's at these freqs that typical room effects are (also see 1)
3) that they've been listening to poorly setup domestic satellite/subwoofers, where the subs seem/are a seperate thing, and they are trying to emulate that (my favourite) :twisted:

And finally in my travels I've noticed a cultural bias in curves
In the UK/Western Europe sub bass is the trouser-flappy 40Hz stuff
In the US Sub bass is more chest thumpy (80Hz)
That in the Pac rim they seem to want more high end, Japanese, Cantonese and Mandarin languages are ripe in sibilants.

great topic, very interesting

cheers,
ferrit
 
Mar 1, 2011
47
0
0
Vegas
Re: contour EQ

Hey Bennet,
yep it could be that, but it seemed like they were going for seperation between the two bands ???
And it wasn't just on Aux sub but also on standard 4 way systems too.

ferrit (seperation anxiety)
 

Tim McCulloch

Graduate Student
Jan 11, 2011
2,920
30
48
Wichita KS USA
Re: contour EQ

And finally in my travels I've noticed a cultural bias in curves
In the UK/Western Europe sub bass is the trouser-flappy 40Hz stuff
In the US Sub bass is more chest thumpy (80Hz)
That in the Pac rim they seem to want more high end, Japanese, Cantonese and Mandarin languages are ripe in sibilants.

great topic, very interesting

cheers,
ferrit
My experience with Asian program material is that there is never enough 1) HF and; 2) reverb.
 

Jay Barracato

Graduate Student
Jan 11, 2011
1,528
2
38
Solomons MD
Re: contour EQ

Hmmm,

Many systems I run into have this same notch. I am now wondering whether it is actually set in the system DSP. I think I have always attributed it to poor crossover setup and cancelation between subs and tops.
 
Re: contour EQ

Philip sent up the bat signal via facebook ;-)

Wow - what a thread! This is one for the archives.

Now - how to keep this short? Not possible. There are too many topics worth commenting on.

Smaart

The key to understanding Smaart's "hybrid" windowing is its use of a continuous signal (typically pink noise) as the stimulus.

Consider a Smaart setup with the offset delay properly set. The first data point in the measurement record is entirely a response to stimuli that happened before the beginning of the reference record; so any correlation with the stimulus is accidental. Now consider the last point in the stimulus record: the entire response to it will happen after the end of the measurement record; once again, no correlation.

However, if you consider the middle point in the measurement record, the stimuli that produced it are mostly captured in the reference record; so nearly all of the direct response will be highly correlated. So, the further in time a response is from the stimulus that produced it, the lower the correlation of the two: the early part of the impulse response is highly correlated and the late part of the impulse response is progressively less correlated.

In essence, the Smaart display combines a windowed direct response with an RTA of the reverberant field. The relative weighting of the two depends on the length of the record and the direct-to-reverberant ratio. At low frequencies, where there is a lot of reverberation, the room has a correspondingly larger effect on the result; whereas at high frequencies, the direct response dominates. That's one reason why the displayed HF level normally increases after you set the offset delay.

... which provides a nice segue into my tuning comments:

Tuning

If you put a flat speaker into a reverberant room, its measured response out in the room (using Smaart) will not be flat. Because reverb generally falls with increasing frequency, the measured response will usually ramp up toward the low frequencies - just like the curves posted earlier in the thread. If you shelve the low frequencies down to make the curve flat, you don't just reduce the reverberation, you reduce the direct sound in that range; resulting in a system that sounds harsh; low-mid heavy instruments (especially transient ones) that sound small instead of powerful; and a midrange that sounds disconnected from the subs.

So my philosophy in that range is to let the system/room keep most of the room's overall character, and mostly use EQ to mitigate the effect of problem frequencies. Outdoors, there is of course very little reverberation; so the natural response of a flat system is pretty flat. A little low-mid emphasis (shelving up below 1000 Hz) will sound better on recorded music, make individual instruments sound a little bigger, and reduce the negative effect of having the highs blown around in the wind.

The appropriate amount of subwoofer boost depends on the style of music. Some sources like upright bass, piano, and bass vocal need their 80 Hz to be balanced with their 200 Hz to sound right. At the other extreme: in house music and most of hip hop, what happens in the subwoofer range is essentially a separate event from what is happening in the mains; so the subs can be radically louder than the mains and the two don't even have to be in the same place for it to work.

Esoteric Transfer Function Stuff


The reason for using cross-correlation and power spectral densities to obtain the transfer function, rather than just dividing the measurement by the reference, is that they're also the inputs for the coherence function. If you keep running averages of all three, you can get both the transfer function and the coherence function from the same set of averages. I'm not aware of any noise advantage to this approach, so I'd be interested to read the paper Nick referred to about that.

Cheers,

David Gunness
 
Re: contour EQ

Hey Guys,
And finally in my travels I've noticed a cultural bias in curves
In the UK/Western Europe sub bass is the trouser-flappy 40Hz stuff
In the US Sub bass is more chest thumpy (80Hz)
That in the Pac rim they seem to want more high end, Japanese, Cantonese and Mandarin languages are ripe in sibilants.

great topic, very interesting

cheers,
ferrit
Many years ago I heard a statement that summarized this. This was in the context of typical 2-way studio monitors or home hi-fi loudspeakers. "In the West a 3-way system is one which has a subwoofer. In the Far-East a 3-way system has a super tweeter."