FR Design Goal? Which is the "best" curve?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This is the only research I know of on the subject Gedlee: Cookies Required

In any case, the original question was about the room. What Toole found was that a flat on axis or listening window(anechoic measurement) was preferred and a response with a slight rise down to the bottom octaves below there. So in room, something like the B&K curve should be pretty close. Audyssey's top notch software should get you fairly close.
 
It would seem that even though some players have their own defenition of the perfect curve they share a common general view.
Mainly flat with a little rise in the LF and a little slope in the HF to get the full but not bright sound that's enjoyable.

To me this would translate into building a hybrid speaker that is set up to be easily EQ'ed to the preferred house curve. Am I getting close?

There is one small issue that I would like to address as well. What sort of peak level should the speaker be able to handle?
My personal rule of thumb has been a minimum of 95dB average listening level + 15dB headroom for peaks -> at least 110dB in the speaker passband.
This is loud but having the woofer bottoming out during a party would be kind of bad.
What's your take on this?
 
Administrator
Joined 2004
Paid Member
That a very good goal, but in truth most people don't listen at 95dB average in typical domestic rooms. 85dB is more realistic. Well recorded music will have peaks at 22dB or more above average, so that still gets you to 110dB with dynamic music.

Not everyone listens to dynamic music, tho.
 
Yeah, 85dB is much more the common/realistic listening level for me as well.
I picked 95dB to have some margin.
To be perfectly honest I think 75dB is probably enough for most of my listening.

I guess I could probably get away with 105dB as well?
But hey, it's always good to plan for the worst case scenario and then make informed compromises when called for. :)

Would it be ok to say that 110 is a good aim and 105 is a fair compromise?
 
Well, I think Pano's 22 dB headroom over 85 dB is not really what is required with most recordings, even for classical music. Unless you have a free standing house or a very well isolated place, you would drive your neighbours nuts. 110 dB is really incredibly loud for a residential dwelling.
 
Administrator
Joined 2004
Paid Member
Yes, that was more a "loudest case scenario" :) Most folks don't listen that loud at home.
I can, even if I don't, because I have a dedicated room and the neighbors can't hear it. My casual listening tends to be in the upper 70s, average. That seems to be typical of most people.
 
That sounds great, I'm mainly crossing the t's and dotting the i's.
If the model/simulation is somewhat accurate the driver is hitting Xmax at 97dB/20Hz and 110dB/40Hz.
It's all done with some creative tweaking but knowing the limits is always nice.
Most likely my amps will clip before hitting Xmax so it's a purely academic issue. Lol.
 
Well, I think Pano's 22 dB headroom over 85 dB is not really what is required with most recordings, even for classical music. Unless you have a free standing house or a very well isolated place, you would drive your neighbours nuts. 110 dB is really incredibly loud for a residential dwelling.

When I am monitoring and mixing a recording, the average level is about 75db. Most often my recordings have about 20db of dynamic range. So 95db(and the caveat is the room has to be very quiet because there is micro and macro detail below 75db) is just the minimum when the room has a very high ambient level.

It is not all about the recording, the ambient level of your room plays a huge role in how loud you should play it back.
 
Are we talking about A weighting here? You have to be careful because this will give you a false impression of what you need to design for. You need C weighted measurements if you are going to use them to determine SPL level goals from your loudspeakers.

C levels in theaters can easily reach 115-120 dB no problem. I've measured this myself.
 
Are we talking about A weighting here? You have to be careful because this will give you a false impression of what you need to design for. You need C weighted measurements if you are going to use them to determine SPL level goals from your loudspeakers.

C levels in theaters can easily reach 115-120 dB no problem. I've measured this myself.

Dr. Gedlee, at what frequecy did you get a 115-120db reading, and how far were you from the front of the theater?

I ask this because in a properly aligned theater, the maxiumum output you should experience from the main speakers is 105db from 2/3 of the way back in the theater - and from 40hz up. From the LFE it's 115db from 80 hz down. Most theater sound system meet this criteria, but 120db is far beyond comfort for most theater customers no matter the frequency. It is also approaching the limits of most theatrical sound systems out there.

Most theater owners would not playback any levels approaching 115db from any channel. We have had in the past begged theaters owners and managers to keep all levels at THX or Dolby approved levels so the soundtracks would playback correctly in their theaters.
 
Administrator
Joined 2004
Paid Member
C levels in theaters can easily reach 115-120 dB no problem. I've measured this myself.
In the bass, maybe. I doubt you'll find that in the midrange. The C weighting might have been leading you astray, there. Although I have been to some cinemas, in Montreal for example, where 115dB might have been close. When I complained I was told "The distributor makes us play it that loud, we have no choice." Yeah, right. :rolleye:

I've also been the projectionist at film premiers where the director is in the house. They usually want it so loud the speakers bleed - it's nutz. SPL to overcome nervousness, I suppose.

But most cinemas aren't that bad. In fact they seem to be getting better.
 
And it seems that they usually try to sustain that level during the trailers for the new movies.
I now bring earplugs along so my ears are not shot by the time the feature comes on.

Trailers are different from the movie soundtrack itself. Trailers are for advertising, and though SMPTE and THX recommend certain levels, nobody follows them that much. Much like popular music, a lot of limiting is applied, and there is not much dynamic range in trailers. It is not uncommon for trailers to bump and stay at a consistent 95-105db level during its entire playback time. For the movie, it would be fatiging and audibly overwhelming to listen to it with sustain long term levels of 105db.
 
Guys, you don't understand "weighting". Its an integral across the audio bandwidth, more a power measurement that SPL. You cannot say that 100 dB "C" means 100 dB at any particular frequency. only if the source is wide band pink noise can you relate the weighted measurement to the SPL at any given frequency. For real, complex signals there is no correlation.

We can say that if we get a "C" weighted level of 100 dB that it is possible that at some frequency there could be a level of 100 dB as long as that was the only signal being played. That why I like to use C weighting - it gives you a worst case level. But 100 dB for many tones would be a "C" weighting of more than 100 dB.
 
Even though I have read about measurements and weighting it's hard to get a real feeling for what the numbers represent in real life.
This chart should make it clear, dB SPL is flat, the different weighting curves vary from flat in a variety of ways, "A, B,C,& D" scale all are reduced in HF sensitivity above 10K.
As you can see, "A" scale rolls off below 1000 Hz, around 35 dB down at 40 Hz, while "C" is only about 3 dB down there.
 

Attachments

  • Weighting curves.jpg
    Weighting curves.jpg
    53.7 KB · Views: 106
Last edited:
Guys, you don't understand "weighting". Its an integral across the audio bandwidth, more a power measurement that SPL. You cannot say that 100 dB "C" means 100 dB at any particular frequency. only if the source is wide band pink noise can you relate the weighted measurement to the SPL at any given frequency. For real, complex signals there is no correlation.

We can say that if we get a "C" weighted level of 100 dB that it is possible that at some frequency there could be a level of 100 dB as long as that was the only signal being played. That why I like to use C weighting - it gives you a worst case level. But 100 dB for many tones would be a "C" weighting of more than 100 dB.

Much of what you are saying here may apply to two channel stereo, but it does not apply to theatrical sound systems.

When properly aligning, a theater sound system has to follow ISO 2969/SMPTE 2202 alignment curve. That means only a 1.5db deviation from 40-2khz, and a 3db per octave roll off above 2khz for the mains, and 80hz to 2khz with the same roll off for the surrounds. That is C weighted. The LFE must be flat down to 35hz, with a strong roll off below that point(most theater sub woofers are equalized to 25hz) to prevent overload to the speaker. There are different SPL requirements for the LFE, the front speakers, and the surrounds. The LFE subs must be able to reproduce 115db 2/3 into the auditorium with low distortion(they don't state how low that is). The front mains each must be able to reproduce 105db's 2/3 of the way into the auditorium with low distortion(once again unstated). Each wall of surrounds(left/right wall, and Left/right back wall) must be able to match the front speakers 105db's when combined output is factored in.

Based on these standards, it is not likely you will be hearing 115-120db from the mains or the surrounds at any frequency even when combined. However, it is possible from the LFE channel as theaters(especially the newer more high end ones) tend to run their sub woofers 2-4db hotter than the 115db's spec for added impact and balance. With the addition of overhead speakers combined with the existing speaker arraignment, running the LFE channel a little hot is critical in achieving the proper SPL levels and frequency balance between each channel.

This is why one must measure the SPL levels based on the frequency not just overall.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.