Frequency Response

[Mr. Widget]

Now that you have all diverged and headed into the "room curve" direction, I'll say that I too have found that a straight line that falls approximately 10dB from 20Hz to 20KHz is what I have found to sound "most natural". That said, we use software (prerecorded music) that was most likely designed for just such a curve. So I am not sure that we really do prefer a falling response.

In any case we are discussing several loosely related topics.

Thanks Rob for posting the info on pink vs. white noise. I meant to get back to Steve's post but gosh darn it I keep having to get back to work.


Widget

[Steve Schell]

Rob, thanks for your informative post. So then it seems that a flat line is what we are looking for after all when measuring pink noise with an RTA.

Mike, I agree that a gently falling response sounds best, at least to me. I have heard various theories over the years that an ideal response falls at 2 or 3dB per octave above 2kHz. or some such.

Zilch, that curve you posted falls at about the same rate that I am seeing in my room at present, when measured with pink noise. I have tried straightening it out with the parametric EQ function of the DEQX, but it sounds too bright to me.

This probably all relates also to our earlier discussion of diaphragm resonances. I once asked Dr. Bruce Edgar about all this, and he said "Flat response would sound fine if it was clean." Problem is that most drivers do not reproduce the top octave cleanly, without artifacts that are grating to the ear.

Widget, when all is said and done we might have to conclude that the response curve that sounds best IS best, which I think is in harmony with your first post!

[Titanium Dome]

Included in this very positive and illuminating review of the K2 S9800

http://www.ultraaudio.com/equipment/jbl_k2_s9800.htm

is a bit on the topic at hand:

My in-room measurements were very promising indeed. Placed well clear of walls, the K2 delivered unusually smooth and flat far-field traces, with a close-to-ideal, gently down-tilted overall characteristic. There is a little too much output around 500Hz, and some unevenness above 7kHz, but the result is very impressive overall, helped because the main horn, running 800Hz to 10kHz here, avoids the usual 2kHz-to-4kHz presence-zone discontinuity.

Also, it seems that Paul Messenger knows a bit about the Lansing legacy.

[scott fitlin]

Mike, I agree that a gently falling response sounds best, at least to me. I have heard various theories over the years that an ideal response falls at 2 or 3dB per octave above 2kHz. or some such.



This probably all relates also to our earlier discussion of diaphragm resonances. I once asked Dr. Bruce Edgar about all this, and he said "Flat response would sound fine if it was clean." Problem is that most drivers do not reproduce the top octave cleanly, without artifacts that are grating to the ear.

OK, days work is done, now talk talk!

Years ago, we used to have people come in a noise the room, walk it with a mic, and use the RTA.

Flat, as in a flat line, everything@0db was not what they did. More like the gradually downward sloping straight line, the graudual slope beginning around 2K as has already been mentioned! We also use a slight rolloff at the very extreme top end.

They called this a big room house curve. Then, after the room was curved, we would spend another hour listening to familiar recordings, and putting the final touches on by ear!

[Ian Mackenzie]

There used to be special threatre room curves and the like however I don't know what they do now with THX.

Back on track with FR, I find the balance of overall response in terms of level of one region relative another crucial to voicing and its lot hard to get right than than tweaking out small dips and peaks.

For example if the 50-250 region is just 1/2 a db down on the 250-1000 region it will effect the tonality of the bass and lower mid. If the 1000-8000 region is down 1/2 or up 1/2 a db it will sound either recessed or forward..Try seeing 1/2 a db on your curves, you probably won't but it will either sound nice or crap......

Small random variations tend to be of little consiquence audibly compared to large regular variations. If I see a small dip or peak I don't loose sleep over it in other words. The area from 1000-3000 however appears more sensitive to variations than some others and is often a source of irritation if it is peaky or rough.

The reality is inexperienced people spend a heap of time equ ing out their home system and often make it worse, although they are convinced they made it better until a friend comes over and says that sound wrong.

Now, early reflections which cause a blurring of the system time response to impulse transient tones are a sigificiant issue in terms of imaging and resolution but you won't see those on a regular Frequency chart. Waveguides are becoming common is small systems, this is because thet are effective in the control early relfections.

If you can't control these early reflections at least make sure they are symmetrical for both left and right loudspeakers. The ear and brain is very good at differentiating variations in the left and right ear and this is why well balanced and set up systems are often referred to as Hi End. This is because the ear/ brain has very little to complain about. The whole response might not be necessarily dead flat but what the brain hear's is the same in both ears.

None of this is rocket science (at least these days) or audio puffery, but is based on research and common good practise.

Ian

[Mike Caldwell]

Hello
Human hearing is at it's most sensitive in the are from about 800hz through the 6000hz area. So a system does not need as much output in those areas for it to sound correct to our brains....in theory.
If I take out the SMAART system to a show I'll give the system a quick check with it and fine tune with my voice using the vocal mic of choice for the given show. Generally I just use the voice through mic and then will tweak a little during sound check and the show. Some guys will bury their heads in screen of their lap top watching the display and forget to listen how things actually sound. Like Scott said you have to spend to time listening.


Mike Caldwell

[boputnam]

Human hearing is at it's most sensitive in the are from about 800hz through the 6000hz area. So a system does not need as much output in those areas for it to sound correct to our brains....

...and for best intelligibility.

Great post, Mike.

Once the house EQ is Smaart'ed, and the gain structure is right, there is generally very little that is needed of the faders. You can bring out the lead vocals at times of increasing band volume just by a (very) slight tweak in the 2kHz to 3.5kHz region. Very subtle changes to EQ in the range you mention have big impacts to it sounding "correctly". These are great "tricks" to keep the SPL from getting away from you. I hate to see even 105 dB anymore...

[boputnam]

Zilch, that curve you posted falls at about the same rate that I am seeing in my room at present...

Me too, but not while running Pink noise.

Steve - I run Pink noise and get the response flat as possible (FFT mode, Smaart). That curve Zilch posted is basically the curve I then get on an RTA. There is less energy needed with increasing frequency to sound "correct".

[Mike Caldwell]

For me a live show be it music or spoken word is all about the vocal intelligibility, if you achieve that everything seems to fall into place. Some people think that a podium mic and a lapel mic gig would be easy but when you may have 1500 or more people hanging on every word someone is saying they don't care what the response cure looks like on your lap top! I used my most recent speech only show as an example, earlier this week I did a big welcome home ceremony for a returning Olympic athlete.
Just to make this more off topic than I have already taken it my favorite podium mic is the AKG c535.


Mike Caldwell

[pde2000]

you are all totally wrong (and right). enjoyment of music is so subjective. we remember the experience of a piece of music bringing tears to our eyes (and the best is live, accoustic preferably) and want to recreate the same every time we put needle to groove. no chance. ive heard extremely expensive stuff and been left cold. one listening room at the penta (uk show) was for cheap little 2 way boxes (audax drivers $100pair) and the front end was an oracle thru krell, but the music was susan vega - awesome, best room in the show. maybe there was a magician behind the curtains casting a spell. i thought for a while that the mains electricity was affecting my hifi because sometimes it gave me no pleasure. then i went to stay in a forest for a few weeks with no mod cons and lots of fresh air and silence (original version of a floatation tank) and when i got back the magic was back. All, and i mean all, hifi reviewers are full of fertilizer and probably take a bung for their opinions to be favourable.

the actual pleasure is in your head. i dont take drugs myself but i understand they can make all the difference (i mean beer of course).

You know Walker from accoustical (Quad) never listened to any of his designs, except sweeps etc., and now the esl57 and quadII are reference. he built all his stuff purely on technical measurement and got a high degree of accuracy which translated to near perfection in music for some customers. But if you compare the esl57 to the esl63 tho, the later is technically superior yet gets no praise.

when you think you know, you dont.

[JBL 4645]

Seems like fairly common issue regarding “it sounds too bright” now what is the common frequency that will have ether one of us grumbling and moaning.

I seem to recall it’s within (2 KHz to 4 or 5 KHz) can’t remember what thread I read this on?

Now, what are the easiest tones that normal hearing can pick out, in the higher frequency register?

KHz 1 2 3 4 5 6 7 8 9 (10 11 12 13 14 15 16 there about)

Also there are tones that seem so easy to hear they sound higher in level over the smaller tones above it because there’s 4 or 8 KHz blocking what is ahead of it, like 10 KHz or maybe 16 KHz.

Would it not be easier for total flat response in the high range to iron out all the peaks so its virtually flat and easier to hear the common tones with less harshness and the more uncommon exclusive tones above 10KHz to 16 KHz or tiny fraction higher thou that would be wishful.

[1audiohack]

Hi Ashly;

Score another one for Bell Labs, the curves you are looking for were created by Harvey Fletcher and W.A Munson. In 1933 they did an equal loudness study and created the Fletcher/Munson curve chart and released a paper on it in 1937.

I don't have a copy of the chart handy but if you go here, lindos.co.uk and open the articals tab, toward the bottom is an artical called Equal-Loudness Contours.

Cheers.

[boputnam]

...the curves you are looking for were created by Harvey Fletcher and W.A Munson. In 1933 they did an equal loudness study and created the Fletcher/Munson curve chart and released a paper on it in 1937.

Here it is, with the supplemental info from 2003 study. You can see the abruptly increased sensitivity between 2kHz and 5kHz.

[JBL 4645]

Hi Ashly;

Score another one for Bell Labs, the curves you are looking for were created by Harvey Fletcher and W.A Munson. In 1933 they did an equal loudness study and created the Fletcher/Munson curve chart and released a paper on it in 1937.

I don't have a copy of the chart handy but if you go here, lindos.co.uk and open the articals tab, toward the bottom is an artical called Equal-Loudness Contours.

Cheers.

Yeah I see bop has posted below or above this post.

Is there a winder stretched out version of the same graph, or is that it?

Okay the way I see it is the red has bit of loudness in or around 1 KHz peak.

The Blue has it softened down reduced followed with some gain reduction around 4 to 9 then slowly rising though the rest of the tones that would otherwise appear very faint due to wavelength and therefore the need to be bit louder while remain tones behind it are reduced because they are easy to hear or would show up on an RTA/SPL db meter.

I was thinking about this a few weeks back because some high frequency tones in films modern Dolby films tend to have this, exclusive high pitched tone that would appear faint (but kinder there).

I don’t have enough bands to deal with all the tones in the high range and this would mess-up the high frequency transparency from LCR to surrounds because the tonal characteristic would change.

Still its worth another go at it. Also I need to boast the amplifier power up gently for the HF for the Control 5.

I don’t like the 1/3 TrueRTA too much. It dances around too much and I’ve tried all three settings to slow it down.

Guess I could run it with (Peak Hold) just have to remain quiet when doing it, the slightest sound and it bumps up a few db!

Still I’d like specification sheet for the HF drivers for the control 5 not the pdf manual it doesn’t show what I need to know, the frequency response for the driver as for sensitivity and ohm/power handing.

[herki the cat]

Flat response with titanium-diaphragm compression drivers is too "bright" for me, With aluminum, damped, and mylar diaphragms, I find that I prefer flat.

herki;
I :dont-know Mylar is a tough durable material with a slight price to pay in that it presents a tizzy artifact around 6,000 Hz which is obvious in certant Electrostatic speakers. Perhaps compression drivers have better damping to control mylar behavior.

Phenolic membranes win every time for damping. Don't be surprised that poorly damped soft titanium diaphragm domes present intensified high frequency artifacts in the spectrum above the mass breakup point activated in circumferential stretching modes at the voice coil junction to the surround suspension.

Note: the 2440 wins with an excellent "half roll' surround. LE-175 & LE-85 & 2420 with identical aluminum diaphragms with tangential surrounds are also magnificent especially with flux density in the 20,000 gauss range.

Intensified high frequency loudness has to do with that fact that perceived loudness is a function of both sound pressure level as well "duration of the sound" which a presents in poorly damped membranes or oscillating bodies in the same manner that a good auditorium intensifies sound duration, especially at the lower frequencies where the true warmth of music prevails.

Conversely Putting the magnificent Philadelphia Orchestra in a cow pasture produces a very anemic sound compared to the beautiful loudness of the concert hall auditorium extended reverberation sounds.