Should Speakers be Designed to Have a Flat SPL?

[TheStalker]

I think it's pretty clear who is kidding themselves in this thread

It's funny how I come to this forum, share something interesting to discuss, and get voted a troll/*** behind my back. Such tough kids you are. It's also sad that most people are unwilling to think outside of the box. On such a technical forum as this, I though we would be beyond the novice idea of flat SPL. Anyone with a $50 Dayton mic and 20 minutes of time can design that. I guess this discussion was beyond a lot of you who resorted to personal attacks because they couldn't grasp a new and more complex concept.

 

[fuzz092888]

It's funny how I come to this forum, share something interesting to discuss, and get voted a troll/*** behind my back. Such tough kids you are. It's also sad that most people are unwilling to think outside of the box. On such a technical forum as this, I though we would be beyond the novice idea of flat SPL. Anyone with a $50 Dayton mic and 20 minutes of time can design that. I guess this discussion was beyond a lot of you who resorted to personal attacks because they couldn't grasp a new and more complex concept.

Is that directed at me? I assume so since you quoted me. First off, my response was only because you threw the first stone after I replied and IMO my response wasn't so much an insult, but a hint that you're ignoring sound advice from guys who know a thing or two about loudspeaker theory and design (and all that psychoacoutic stuff as well). Secondly, there was nor is anything wrong with what I suggested to the other member. Third, if you had any idea what all this stuff actually means you would know that what I suggested is quite possible since the anechoic response of a speaker is not what you're going to get in room, unless the room is heavily treated. Finally, no a $50 Dayton mic, REW and 20 minutes is not what it takes to design an accurate speaker with smooth, even off axis response.

If you'd care to learn more I'd suggest reading at minimum the "Loudspeaker Design Cookbook". At least then you'll have a small inkling as to what goes into designing and creating excellent speakers.

A person's preference is their own and whatever response he's getting from his speakers with the EQ set the way it is, is pleasing to him. I merely mused that he may not be doing what he thinks he's been doing and may actually be taming a little peak in the response of his speakers which might be caused by the room or placement of the speakers.

 

[TheStalker]

So it looks like Audyssey is doing the BBC dip automatically. They like it, they recommend it, they believe in it: https://audyssey.zendesk.com/entries/410117-Midrange-Compensation

Anyone who built the active crossover from Linkwitz kit will also have a BBC dip at 3kHz automatically.

A lot of you have probably already been listening to the BBC dip and not even realizing it.

 

[TLS Guy]

Interesting Read. Some may laugh but heres my EQ settings that i've been using for the past 20 years(Even with my 803D's).

Switching eq on and off the other night with about 8 people there and all except one liked my sound versus no eq.

Seems I deaded the very frequencies discussed here.

You don't say what your source is, but I suspect a computer since that is Winamp. If you are playing lossy codecs, then that is likely your problem. The need to set Eq like that is evidence of a problem on your reproducing chain somewhere. I can assure that when everything is all calibrated correctly and playing good sources, you do not need an Eq curve like that to get a deep and expansive sound stage.

You have trouble some place, of that I can be certain. Having to use a setting like that, is a back to the drawing board order for me.

 

[jcl]

So it looks like Audyssey is doing the BBC dip automatically. They like it, they recommend it, they believe in it: https://audyssey.zendesk.com/entries/410117-Midrange-Compensation

Anyone who built the active crossover from Linkwitz kit will also have a BBC dip at 3kHz automatically.

A lot of you have probably already been listening to the BBC dip and not even realizing it.

First you stated that the dip is to compensate for how the human ear hears diffuse sound vs. how a microphone records diffuse sound. Now you are pointing to literature that says the dip is for covering up issues in crossover design. What is your real point?

Audyssey states that this is to correct issues with the crossover from midrange to tweeter. This is fine for an after the fact way to mask the issue, that can be turned off if not needed. However, why not fix the problem in the first place with a properly designed loudspeaker?

Back to linkwitz and your original argument. I questioned your claim in the other thread and your response was to ignore the logic. Once again -
Microphones come in many different polar pickup patterns specifically designed to control the amount of direct vs. diffuse sound recorded. Further there are different miking techniques to control the amount of direct vs. diffuse sound recorded. Lastly much of the music recorded today is recorded using close miking techniques in which case there is little to no diffuse sound recorded.
So even if your idea of loudness compensation is valid, you would have to vary the amount of 'correction' for each recording. Thus, I think having it hard wired into a speaker crossover is a bad idea.

Finally Linkwitz seems to agree, suggesting - " You may choose to make the notch filter selectable with a switch for different types of recordings."

 

[TheStalker]

Maybe you should read that Audyssey thread again, in its entirety, before you star taking things out of context and putting words in my mouth. Seriously, this is getting lame. So many people out to attack because they can't, or are unwilling to understand this concept.

 

[Steve81]

So many people out to attack because they can't, or are unwilling to understand this concept.

It might help if you stuck with a single concept; you're posting equal loudness contours, arguments from Linkwitz about the difference in perceptions between the diffuse sound in a concert hall vs direct sound from a loudspeaker, and Audyssey's midrange compensation to help correct for the directivity mismatch that occurs in many loudspeakers. These aren't related to one another.

 

[AcuDefTechGuy]

It's funny how I come to this forum, share something interesting to discuss, and get voted a troll/*** behind my back. Such tough kids you are. It's also sad that most people are unwilling to think outside of the box. On such a technical forum as this, I though we would be beyond the novice idea of flat SPL. Anyone with a $50 Dayton mic and 20 minutes of time can design that. I guess this discussion was beyond a lot of you who resorted to personal attacks because they couldn't grasp a new and more complex concept.

I would never personally attack anyone just because I don't agree with him or her on ANY subject. It's cool with me if you prefer the sound of B&W. I think my 802D2 sound great as well. I also think my Salon2, KEF 201/2, Orion3, TAD 2201, Phil3, etc., sound great. So flat vs. not-so-flat? Doesn't matter to me as long as they sound great.

 

[jcl]

Maybe you should read that Audyssey thread again, in its entirety, before you star taking things out of context and putting words in my mouth. Seriously, this is getting lame. So many people out to attack because they can't, or are unwilling to understand this concept.

I'm sorry you feel that I am attacking, I merely intended to debate. I engaged you on the other thread, but you dismissed my post without the courtesy of any real response. Here once again you fail to put effort into a response.

From the Audyssey site -
"Midrange compensation is an intentional dip in the 2 kHz region where the vast majority of tweeter-to-midrange crossovers are. In that region the tweeter is at the low end of its range and the midrange at the high end of its range and the directivity of the speaker goes through major changes. We found that if that region is equalized to flat, the change in direct to reflected ratio that happens because of the directivity variations causes voices to sound harsh (among other things)."

I read nothing here regarding the recording of diffuse vs. direct sound, but rather the directivity of midrange and tweeter drivers in the speaker. Perhaps you are equating the two, but you did not make that point clear in your post.

Perhaps I was over simplifying, it's not just the crossover that is being compensated for, but the design of the speaker - The midrange chosen, the tweeter chosen, and the crossover between the two.
I would contend that a well designed loudspeaker will have a smooth transition in frequency & dispersion between drivers.

In your original post you state " I believe that speakers should not have a flat SPL, but follow more closely to how an average human ear hears. Flat speakers to me sound hard, bright, and unnatural."

That's fine, at the end of the day good sound is a matter of opinion. If you have found what you like, terrific, enjoy the music. My contention is that having a fixed equalization applied at the speaker is limiting.

 

[slipperybidness]

Has anyone read any books on audio by G Randy Sloan?

His view sums up this situation very well.

The goal of audio reproduction is not "to sound nice". Rather the goal of audio reproduction is to be as accurate as possible.

TheStalker is coming into this with pre-conceived bias based on faulty reasoning and bad assumptions.

As I said before, IF his argument were to have ANY merit, the argument MUST be that the speaker would be voiced as the INVERSE of the perception curve, not voiced following the curve.

 

[Irvrobinson]

From the Audyssey site -
"Midrange compensation is an intentional dip in the 2 kHz region where the vast majority of tweeter-to-midrange crossovers are. In that region the tweeter is at the low end of its range and the midrange at the high end of its range and the directivity of the speaker goes through major changes. We found that if that region is equalized to flat, the change in direct to reflected ratio that happens because of the directivity variations causes voices to sound harsh (among other things)."

One good thing about this thread is that it pointed me to the discussion on the Audyssey web site. Audyssey is now off my list permanently as a technology I would ever consider using, for two reasons. First, once I go through the trouble to choose speakers that sound accurate on familiar recordings I don't want optimization electronics like Audyssey making assumptions about "improvements" that were determined without knowledge of what speakers I'm using. Second, any company that purports to optimize based on in-room measurements and *then* applies a pre-determined coloration that can't be defeated is not philosophically aligned with my objectives in an audio system.

 

[ski2xblack]

Has anyone read any books on audio by G Randy Sloan?

His view sums up this situation very well.

The goal of audio reproduction is not "to sound nice". Rather the goal of audio reproduction is to be as accurate as possible.

I have not read any books by Sloan, but I have read Toole's, and it seems that there is a general preference for a gently tipped-down frequency reponse. I prefer such a 'house curve' myself. It's less "accurate" in the strictest sense, but more pleasing to me, without a doubt. I do appreciate a speaker design that provides flat on axis and smooth off axis response, for sure, it's the best baseline to work from. But then I apply a curve that "sounds nice", because to me, that is the goal.

Has anyone linked the home theater shack article on house curves? Stalker, you may find it interesting, as it touches on many of things you seem concerned with and may help your understanding. If you want to go deeper, get a copy of Toole's book.

 

[Irvrobinson]

I have not read any books by Sloan, but I have read Toole's, and it seems that there is a general preference for a gently tipped-down frequency reponse. I prefer such a 'house curve' myself. It's less "accurate" in the strictest sense, but more pleasing to me, without a doubt. I do appreciate a speaker design that provides flat on axis and smooth off axis response, for sure, it's the best baseline to work from. But then I apply a curve that "sounds nice", because to me, that is the goal.

I also use a "tipped down" in-room frequency curve, at least as measured by OmniMic, and I think the primary reason it "sounds better" is because most home rooms are "small". In a very large venue I think a truly flat curve would sound about right, but even a 3000 cf untreated room is going to have a pretty strong acoustic signature, or coloration is probably a better word, and a tipped-down source compensates for that. I didn't read Toole's book when I came to that conclusion, it was by playing back my own live recordings and determining what sounded most like the real thing. I also think most folks are discounting the level of coloration you get by playing back a recording's with the original room's acoustics in a room with its own acoustics, even if it is the same room. Nonetheless, I'm still an advocate of speaker designs that try to be as flat as possible. A speaker designer cannot know what room I'll be using their speakers in, so they cannot know what colorations they sound compensate for, so they should resist the temptation to pre-color the sound in any way in an attempt "fix" things.

 

[fuzz092888]

I have not read any books by Sloan, but I have read Toole's, and it seems that there is a general preference for a gently tipped-down frequency reponse. I prefer such a 'house curve' myself. It's less "accurate" in the strictest sense, but more pleasing to me, without a doubt. I do appreciate a speaker design that provides flat on axis and smooth off axis response, for sure, it's the best baseline to work from. But then I apply a curve that "sounds nice", because to me, that is the goal.

Has anyone linked the home theater shack article on house curves? Stalker, you may find it interesting, as it touches on many of things you seem concerned with and may help your understanding. If you want to go deeper, get a copy of Toole's book.

I have read that article along with several of Wayne's other articles. While I generally agree with him, the house curve definitely isn't for me. Two of my favorite instruments to listen to, the violin and the trumpet, are right around that BBC dip and definitely in Wayne's house curve area. My own yard stick is how close those two instruments sound during playback as opposed to when I heard them live (I used to play the trumpet, albeit poorly). For me in my room, a flatter response yields a more realistic reproduction to me with no house curve.

I'll also say that I've never exactly been blown away by movie theater sound and definitely prefer the sound in my own HT. However, whether that's due to my preference for flat, me telling my brain that I prefer flat, or the "one size fits all" X-curve that the industry uses, I have no idea.

 

[UnxpectedError]

You don't say what your source is, but I suspect a computer since that is Winamp. If you are playing lossy codecs, then that is likely your problem. The need to set Eq like that is evidence of a problem on your reproducing chain somewhere. I can assure that when everything is all calibrated correctly and playing good sources, you do not need an Eq curve like that to get a deep and expansive sound stage.

You have trouble some place, of that I can be certain. Having to use a setting like that, is a back to the drawing board order for me.

I think its just my preference. I only like these settings when I play music, and more importantly when I play loud music, which is most of the time. If I turn the EQ off the midrange just sounds too harsh to me.

My sources are fine (I think anyways). I play FLAC or 320kb mp3 output from a nice sound card with Burr Brown DAC.

I can't tell a difference between my computer source and a CD to be honest, unless i'm playing a mp3 less than 320.

 

[AcuDefTechGuy]

Has anyone read any books on audio by G Randy Sloan?

His view sums up this situation very well.

The goal of audio reproduction is not "to sound nice". Rather the goal of audio reproduction is to be as accurate as possible.

TheStalker is coming into this with pre-conceived bias based on faulty reasoning and bad assumptions.

As I said before, IF his argument were to have ANY merit, the argument MUST be that the speaker would be voiced as the INVERSE of the perception curve, not voiced following the curve.

My opinion is that SPEAKERS should be designed to be flat and accurate as possible. This is home base. The baseline. After that it is up to individual minds to modify the sound per their personal tastes. For example, artificially boost bass or treble or whatever per personal preference.

 

[surveyor]

Audyssey MultEQ XT32 is very welcome in my system!

 

[TheStalker]

I also use a "tipped down" in-room frequency curve, at least as measured by OmniMic, and I think the primary reason it "sounds better" is because most home rooms are "small". In a very large venue I think a truly flat curve would sound about right, but even a 3000 cf untreated room is going to have a pretty strong acoustic signature, or coloration is probably a better word, and a tipped-down source compensates for that. I didn't read Toole's book when I came to that conclusion, it was by playing back my own live recordings and determining what sounded most like the real thing. I also think most folks are discounting the level of coloration you get by playing back a recording's with the original room's acoustics in a room with its own acoustics, even if it is the same room. Nonetheless, I'm still an advocate of speaker designs that try to be as flat as possible. A speaker designer cannot know what room I'll be using their speakers in, so they cannot know what colorations they sound compensate for, so they should resist the temptation to pre-color the sound in any way in an attempt "fix" things.

Actually flat is what gives the colored sound. And not supporting Audyssey is a bit closed mindedness.

 

[Irvrobinson]

Actually flat is what gives the colored sound.

That's ridiculous. I've read this thread, and you are not persuasive. You are also quoting Linkwitz out of context. He tweaked the equalization of a dipole by a tiny amount, and you're using it as a general argument to justify a frequency response dip that was conceived to adjust a coloration in a non-dipolar design that isn't applicable to other speakers.

As for Audyssey, I don't think I'm being close-minded at all, I'm reading responses from one of their technical leaders.

 

[cpp]

Actually flat is what gives the colored sound. .

Flat: The faithful reproduction of an audio signal

Colored: A "colored" sound characteristic adds something not in the original sound. The coloration may be euphonically pleasant, but it is not as accurate as the original signal

To different meanings