DSP: EQ’ing Full Range or Not – Testing the Ears

[Erik Arisholm]

But they are interrelated. Adjusting delays affects the frequency response.

Yes, of course, they are very much related. I agree wrt blending subs with mains, but I am not really talking about sub/main integration here, just the overall response of the speaker(s) vs room ;-)

 

[Erik Arisholm]

The bass hump has nothing at all to do with compensating for poor alignment... it is simply what most people prefer to compensate for the hearing loss down low (Fletcher Munson curve).

Yes I understand but this is where I tend to disagree, based on my experience ;-) I think, once you get the timing right (and that includes control of reflections), you will end up wanting less tilt below 200Hz or so... What most people miss is the bass impact, not the overall bass energy and there is a difference...the impact is the slope, the energy is the integral... (and it does not matter if subs are involved or not)

 

[Erik Arisholm]

Sorry I have to correct myself, I do agree that we have to compensate for less hearing sensitivity at lower frequencies at lower SPL. But that was not really my point.

 

[teyhyrh4r]

Achieving a smooth frequency response is the result of managing delays and equalization and is talked about all the time with regard to subwoofers

most methods you fid in the internet result in a phase alined sub at crossover, but a cyle too late. in reality the mains need delay, not the sub

 

[teyhyrh4r]

loudness compensation shouldn't be done with arbatrary bass boosts. we have a lot of solutions using ISO 226 for this allready. Equalizer Apo has a module for this for example

 

[Eric SVL]

With my new 20" ASC Isothermal tubetraps I achieved significant improvements down to about 40Hz or so in terms of decay times. And exceptional results around 80Hz where it really matters. Not so with my earlier "bass traps" from GIK etc:-)

So yours are tuned and not broadband? That changes things.

 

[Eric SVL]

Yes, of course, they are very much related. I agree wrt blending subs with mains, but I am not really talking about sub/main integration here, just the overall response of the speaker(s) vs room ;-)

I am not, either. I'm talking aligning subs to each other.

 

[teyhyrh4r]

the deepest mode has to be treated at the backwall, and with thick material.

 

[Eric SVL]

most methods you fid in the internet result in a phase alined sub at crossover, but a cyle too late. in reality the mains need delay, not the sub

In 10+ years I have never heard this. Do you have a link?

 

[teyhyrh4r]

In 10+ years I have never heard this. Do you have a link?

the problem is that subwoofer intergration was pioneered by video guys (home theatre). and there is a practical reason for the bad solution: when you delay the mains, you also have to somehow delay the video signal.
just look at excess group delay and you will see it.

 

[Eric SVL]

the problem is that subwoofer intergration was pioneered by video guys (home theatre). and there is a practical reason for the bad solution: when you delay the mains, you also have to somehow delay the video signal.
just look at excess group delay and you will see it.

I have a feeling you don't own modern HT equipment, is that right?

When we set up subs and speakers, we don't take into account the video at all. We are only aligning the sound from all speakers and subs to our seats.

When a delay is needed for the video, which is typically only with projectors, a global delay is added to the entire system. It does not change relative speaker and sub delays.

 

[teyhyrh4r]

I have a feeling you don't own modern HT equipment, is that right?

correct. I don't even own a reviever/amp

We are only aligning the sound from all speakers and subs to our seats

my point is that phase alignment is not enough.
if your excess group delay looks like this, your bass is too late:

I'm not saying that you hear the delay...you don't.
but you will lose some "punchyness"

 

[Sonnie Parker]

I want more than just "punchyness" with my bass, especially with movies, but also with music. I want it to be accurately reproduced, which is what I believe I have with the target curve I have. I also keep another preset curve with a little more bass for fun sometimes. But if I don't compensate for the hearing loss, punchyness ain't gonna fix that. Now... if you've listened for years like you listen, you have grown to the sound, even though it may be inaccurate reproducing what the recording engineer intends for you to hear. This happens quite frequently to most of us. We like what we hear, we keep listening, even though it may be a little off of intended, we listen to it long enough and we adjust to it and like it.

 

[teyhyrh4r]

how is "accurately reproducing" defined?
A falling FR after 800-1000Hz-ish at LP is psychoacousticly explained, the bass boost isn't. it is just something we got used too cause it will happen in every untreated room.

Now, if you want to get close to what the recording engineer heard....his room wont produce a strong bass boost like a normal room, cause he will work in a heavily treated room.
He also wont have a sub that is too late

 

[Michael Boeker]

most methods you fid in the internet result in a phase aligned sub at crossover, but a cycle too late. in reality the mains need delay, not the sub

I agree. One problem is that Denon and Marantz receivers only allow for a maximum of 20ms of delay. In a regular sized home theater, this often isn't enough to properly align the subwoofer. I also see a lot of people start with the actual measured distance from the seating position to the subwoofer and then use frequency response to fine tune the delay of the subs so the crossover region is smooth. What they don't realize is that you can delay the sub a cycle and the frequency response looks exactly the same.

the problem is that subwoofer integration was pioneered by video guys (home theatre). and there is a practical reason for the bad solution: when you delay the mains, you also have to somehow delay the video signal.
just look at excess group delay and you will see it.

Unless you are in game mode, the lowest input latency for a TV is 17.5 ms. A Sony OLED can have 168.6 ms. A projector can have 30 ms to over 100ms, depending on the processing. I use a Sync-One-2 for for accurately setting lip sync. There is never an instance where the video needs to be delayed vs the audio. The audio always has to be delayed - even after time, phase, and polarity alignment of the mains and sub.

 

[Sonnie Parker]

how is "accurately reproducing" defined?
A falling FR after 800-1000Hz-ish at LP is psychoacousticly explained, the bass boost isn't. it is just something we got used too cause it will happen in every untreated room.

Now, if you want to get close to what the recording engineer heard....his room wont produce a strong bass boost like a normal room, cause he will work in a heavily treated room.
He also wont have a sub that is too late

Accurately reproduced is what the recording engineer intends for us to hear. I defined it in my previous post... you must have missed it.

I don't have a sub that is too late, but I still need to compensate for the lower frequencies to hear what the engineer intends me to hear. No amount of treatment is going to fix what I am missing and what I should be hearing... unless I just get used to not hearing it and live with it, become accustomed to hearing it.

 

[Erik Arisholm]

Accurately reproduced is what the recording engineer intends for us to hear. I defined it in my previous post... you must have missed it.

I don't have a sub that is too late, but I still need to compensate for the lower frequencies to hear what the engineer intends me to hear. No amount of treatment is going to fix what I am missing and what I should be hearing... unless I just get used to not hearing it and live with it, become accustomed to hearing it.

It is true that at low SPL, the ear is less sensitive to the lowest and highest frequencies than the mids, but this difference diminishes at higher SPL. I.e., the so-called equal loudness contour of Fletcher Munson. And at very low SPL, the lowest bass will even fall below your threshold of hearing (and typically also the noise floor of your room, but that is another matter). But the amount of boost needed depends on the level. Hence the (in)famous loudness button in the old amps. Thus, a static house curve with a 1dB-2dB per octave tilt in SPL below 200Hz regardless of volume will not give you what the recording engineer intended, even if you may prefer it that way in your room and for your music preferences (and I have no objections to that;-).

I am sure you know all this but that is why I get a little confused by your statement in bold. Loudness control is still a good idea I think and can now be implmented better with DSP assuming that you have a calibrated reference level, as for example Lyngdorf Room Perfect does in the "leveling" stage of the room correction procedure (so that -20dBFS will results in 85dB SPL at a "reference" volume setting). But at THX reference levels there should be no need for adjustments above an in-room linear frequency response, as that would be about what the recording engineer heard, at least if you, like the recording engineer, have reasonable control of decay times at all frequencies in your room.

Still, I can certainly respect that many people prefer a bass boost or the various famous tilts. Heck, I even just invested in new stereo subs (4x18" servo, 6,4kW amps) with my Westminsters and active crossovers with Merging HAPI multichannel processor and Audiolense linear phase XO (so it will be perfectly time aligned). So I am a bass junkie and it has to go plenty deep, but it must kick at the right time and decay at a "decent" rate: the bloated bass rumble I hear almost everywhere where people throw a sub in the mix and crank up the volume is the worst of all, IMHO :-) It is by no means perfect in my room either, not yet at least, but the goal I strive for is linear and punchy (so properly time aligned) bass at my "typical" listening levels. I do however prefer a slight downward tilt above 1kHz.

 

[markus]

how is "accurately reproducing" defined?
A falling FR after 800-1000Hz-ish at LP is psychoacousticly explained, the bass boost isn't. it is just something we got used too cause it will happen in every untreated room.

Now, if you want to get close to what the recording engineer heard....his room wont produce a strong bass boost like a normal room, cause he will work in a heavily treated room.
He also wont have a sub that is too late

Unfortunately professional control rooms aren't as similar as one would hope. Here's some real world data:

(Source: http://seanolive.blogspot.com/2009/10/audios-circle-of-confusion.html)

It can also be observed that recordings from different decades or genres follow similar spectral patterns. Additionally there's personal preference, radiation pattern of speakers, room acoustics and equal loudness. All in all a multitude of factors which ultimately prevents us from hearing what the recording/mixing/mastering engineer intended us to hear.

P.S. Don't shoot the messenger

 

[Sonnie Parker]

I suspect the engineers are not listening to music at reference levels... maybe movies, but I doubt it on music.

I also believe on movies there is the +10dB in the LFE channel already.

 

[ddude003]

It is called art... Fortunately professional control rooms aren't similar and don't sound similar... Music and sound tracks would get really boring... Imagine only one singer, one piano, one guitar, one drum kit, etc... Only one mic, only one equalizer, one mixing board... One speaker... Well, maybe you get my point... Just think of all the creators of fine musical instruments and all the creators of amplifiers, speakers etc... Choices are good... Part of the skill of being an artist is knowing and picking the various parts and pieces that bring a signature sound together... Vive la difference...

 

[Eric SVL]

how is "accurately reproducing" defined?
A falling FR after 800-1000Hz-ish at LP is psychoacousticly explained

It is not psychoacoustics. It is the result of measuring a room with an omnidirectional microphone as directivity of tweeters narrows and loads the room less and less.

 

[Eric SVL]

He also wont have a sub that is too late

I can do some impulse measurements. How much resolution is needed to look for this effect? What frequency are we defining how long one "cycle" is? And are we looking at the start of the impulse, or the peak? Remember sound starts before the peak.

100 Hz = 10 ms
90 Hz = 11 ms
80 Hz = 12.5 ms
70 Hz = 14 ms
60 Hz = 17 ms
50 Hz = 20 ms
40 Hz = 25 ms
30 Hz = 33 ms
20 Hz = 50 ms

After measuring, the next step would be to establish audibility. Just because it's measurable, doesn't mean it's audible - humans don't have the same resolution. There are thresholds in time and level.

 

[teyhyrh4r]

I still need to compensate for the lower frequencies to hear what the engineer intends me to hear

again:

loudness compensation shouldn't be done with arbatrary bass boosts. we have a lot of solutions using ISO 226 for this allready. Equalizer Apo has a module for this for example

Unfortunately professional control rooms aren't as similar as one would hope. Here's some real world data:

View attachment 43944

(Source: http://seanolive.blogspot.com/2009/10/audios-circle-of-confusion.html)

It can also be observed that recordings from different decades or genres follow similar spectral patterns. Additionally there's personal preference, radiation pattern of speakers, room acoustics and equal loudness. All in all a multitude of factors which ultimately prevents us from hearing what the recording/mixing/mastering engineer intended us to hear.

P.S. Don't shoot the messenger

while this problem obviously exists, I have two problems with the conclusion:

1) is the survey realy concluding what Olive wants to show? first thing that got my atention was "Surround Audio". also "Measured rooms have been equipped with factory calibrated three way monitors acoustically calibrated in-situ after installation with a standardized apparatus". It seams the study didn't realy meassure what the engeniers are working with.

2) who realy does decide the overall spectral balance is the mastering engenier, and they don't work in monitoring rooms

good post, though. I think someone would have to realy read the study to understand the relevance for our discussion: https://www.aes.org/e-lib/browse.cfm?elib=10075

I suspect the engineers are not listening to music at reference levels

they don't. there is no standard at all. some mastering(!) engeniers will have realatively quiet levels, others crazy loud ones.
but it doesn't matter too much in therms of equal loudness contour because they calibrate their ears with existing material.
on a sidenote: that's why I personaly think the best "reference level" for your system is calibrated per genre and using ears only

It is not psychoacoustics. It is the result of measuring a room with an omnidirectional microphone as directivity of tweeters narrows and loads the room less and less.

while you are correct, if there were no psychoacoustics involved we still would EQ to flat on LP. why is this too bright? because around 1000Hz we start to hear direct sound sperated form reflected sound. around 1500Hz only direct sound will dictate. https://en.wikipedia.org/wiki/Sound_localization#ITD_and_IID
we can say: up to 1000Hz our inroom meassurements show what we hear. obove that we have to start observing the direct sound

 

[markus]

It is called art... Fortunately professional control rooms aren't similar and don't sound similar... Music and sound tracks would get really boring... Imagine only one singer, one piano, one guitar, one drum kit, etc... Only one mic, only one equalizer, one mixing board... One speaker... Well, maybe you get my point... Just think of all the creators of fine musical instruments and all the creators of amplifiers, speakers etc... Choices are good... Part of the skill of being an artist is knowing and picking the various parts and pieces that bring a signature sound together... Vive la difference...

I beg to differ. Differences in sound reproduction (amp, speaker, room acoustics) should not become part of the art. It creates two issues:
1. You can never know what the artist wanted you to hear because you don't know and therefore can't recreate the acoustical environment in which he made his artistic decisions.
2. You can never easily adjust for your personal preference. One recording might need different adjustments than the other because it was created in a different acoustical environment. Again you don't know what to adjust for. The endless tweaking for "best sound quality" begins...

In my mind better reproduction standards would let you appreciate the art even more. Right now it just creates confusion and some of the art is even lost.

 

[markus]

I think someone would have to realy read the study to understand the relevance for our discussion: https://www.aes.org/e-lib/browse.cfm?elib=10075

Well, I did