Desirability of flat FR in speaker design???

[terry j]

hi everybody,

on another forum a discussion has started on 'accuracy vs personal taste'. It came about because I disagreed with someones statement that 'it's how it sounds to you that matters', I only really disagreed to the degree that a statement like that opens the door for scoundrels and wacky theories, both of which abound in hi fi.

I stated that personal taste is definately a factor, two different systems with the 'same' FR can sound totally different, yet surely as a starting point a flat speaker is necessary in a high end system?

I must say I was stunned, it has almost become 'forget measurements, trust your ears'. It is probably not that simple, but it would certyainly be true to say that very few agree that a flat FR from the speaker is desirable, or at least needed.

For the sake of simplicity I didn't get into house curves etc etc, but it got me wondering. What I want to know is, do 'audiophiles' ( what a snobby sounding word) believe that a flat FR is desirable in a speaker? In other words, I made the assumption that, like me, people into stereo would believe that a flat ( also read even ) FR is the hallmark of a well designed speaker, and a starting point ( in whatever price bracket you can afford ) for auditioning. Perhaps I was stunned because in fact very few agree with me and I am in the minority.

The attitude almost seems to be that a 'flat' FR would be detrimental to their sound! I just can't agree with that, but am I an 'only one'?

I'm not trying to defend my territory or convert others or even say the other side is wrong, I'm almost doing a survey if you will. Of course if you want to add your feelings on the qyestion then by all means do so.

As I say, I'm leaving out the complexities of house curves etc etc

lots of love

terry

[Otto]

Personally, I would agree that a reasonably flat frequency response is a good starting point. If there are big peaks and valleys that accentuate and detract from the original recording, there's something wrong. Of course, and it seems like you're stating this, a flat FR doesn't tell the whole story. There are so many other factors that go into getting a sound that's pleasing to you, that you can't rely solely on FR.

You might ask them what their "limits" on FR deviations would be. Would they accept flat +/- 3d B? +/-10 dB? 20 dB?

[geekwithfamily]

Everyone that's willing to spend more than $500 on a pair of speakers better consider all of the measurements that they might publish in Stereophile in a loudspeaker review. Stereophile reviews are a great balance between the subjective impressions of the main reviews and John Atkinson's measurement sidebar, where he comments, "I can see why the reviewer mentioned this coloration because of this peak or long decay or whatever…"

I never really understood measurements until I subscribed to the Audio Perfectionist Journal by Richard Hardesty and Shane Buettner. Mr. Hardesty explains all the measurements and their varying importance but also how to listen critically and hear beyond the measurements for each component in your system.

Flat or on target measurements can tell you whether you are fooling yourself at a dealer by enjoying a short term coloration at a demo that may be fatiguing when you get the kit home.

Last month I attended a talk by Richard Vandersteen at a dealer and he explained his philosophy of a good speaker. Your loudspeaker needs to be time and phase correct in addition to a flat frequency response to pass an accurate waveform from your amp. A time/phase coherent speaker is defined by: acoustically centered drivers, one driver per frequency range and first order crossovers connected to the drivers in phase (meaning when the tweeter pushes the woofer pushes). Any deviation from these parameters and you have a lesser speaker. Get ready for colorations, smearing and tilted timbre that will not do justice to the original recording.

Only three current manufacturers produce time coherent loudspeakers: Vandersteen, Thiel and Quad ESLs. I have Vandersteen 1c's so I may be biased because I've spent money, but I think they're the best speaker for under $1000. I've never enjoyed my music more. The speakers' measurement make me feel secure and let me enjoy the music.

Mr. Vandersteen also taked about the disappearance of tone controls on our amplification systems and that loudspeakers and cabling have taken their place. Many speakers that can seduce a buyer at the dealer have a frequency response similar in shape to the inverted arc I set my boom box's 5 band equalizer to when I was twelve. The boost in the lows and highs would make badly recorded (too loud and midrange only) albums sound better.

It's hilarious that audiophiles will buy the most accurate and expensive sources and amplification only to have the signal mangled by a colored loudspeaker at the end. This is where your golden ear decides for you.

So yeah, flat frequency response is paramount along with spectral decays and step responses.

"Whatever sounds good to you" is a cop-out. Any experts on these forums should give their expert advice to raise our signal above the noise.

[terry j]

thanks guys

I wasn't going to reply until a few more responded, but looks like quite a few views with not many responses! Early days yet tho.

Otto, thanks for your input, have enjoyed your tips on using REW to measure speakers. Thats right up my alley too.

Enjoyed your writeup Geeky, it is exactly what I've always thought. Yeah, that phrase is a cop out, hides a multitude of sins doesn't it.

The speaker manufacturer has absolutely no idea of where the unit may ultimately end up, so has got no choice other than to make it flat. That doesn't mean we can't have house curves, or that flat in the home is enjoyable or not enjoyable or whatever. None of that is the manufacturers responsibility. If we can't set it right in the home then hopefully you've bought from a dealer willing to help and knows what they are doing.

What gets me shaking my head in disbelief is that these guys will spend thousands on a DAC for crying out loud, 'because it's important to extract every last bit of info from the disc' and then have an attitude like that about speakers??? huhhh???

Imagine buying an amp with a 10 db hump in the midrange! And thinking thats OK!

What also has me in a state of disbelief is that I'm arguing for a flat well designed speaker on a hi fi forum - go figure.

I'm tinkering with a DEQX, amazing. It's a bit like the TAct if you've never heard of it. I'd always assumed that in the hi fi crowd, the only barrier on the desirability of a unit like that was the cost. But it looks like to most it wouldn't even be considered. Yet, for say 80% of the cost of a DEQX some people will buy a preamp that doesn't have one fiftieth of the capability of the deqx.

Guess it takes all types.

see ya

[drf]

I might be a little simplistic when it comes to this sort of thing, but my opinion is that an eq is only there to help with the FR and to enhance the sound, not to be a substitute for bad speaker response. E.G say your speakers physically can't reproduce 2KHz due to bad crossovers and or bad allignment, then no amount of Eq'ing will bring that freq' back. Therefore I assume that a flat reponse is essential, I also say an eq is essential because sometimes the engineer who works on a cd has "different" taste in tone than the home listener.

Cheers,

Dr F

[terry j]

Mr Geek, quick question. I've looked at the two 'free' issues of the audio perfectionist, and liked very much what I read.

You've said you subscribe, in your opinion is it an opngoing worthwhile magazine??

DRF, yeah, but we'd hope a speaker designer at least got the basics right?? Maybe not in every case tho..ouch.

[geekwithfamily]

Terry, you can call me Kyle. Mr. Geek is so formal.

I subscribed to the first twelve issues of the Journal. Its advice and education led me to the system I have now (see my sig for the affordable hi-fi). Each issue focuses on a different aspect of a high quality hi-fi and home theater system.

The journal got me back into vinyl and turntables, made me finally realize the benefits of separates, explained all the measurement charts and how to equate them to speaker performance, how to enjoy the best of home theater and hi-fi in the same system, and made me brave enough to save a ton of money buying used components. The journal is really responsible for my deep interest in this hobby.

The only reason I didn't renew my subscription for the next run was because it wasn't in my tight home budget anymore. I hope to subscribe again soon and catch up on the back issues I'm missing.

The only problem I ever had with the journal is the need to buy a new component after the APJ has covered the category. The loudspeaker issues came out, I bought my Vandersteens, pre-amp issue = Rogue 66 then 99, amp issue = Anthem MCA-2 stereo amp. So really a financial problem but I have a system I love and am very satisfied with.

[SteveCallas]

It comes down to accuracy vs personal taste. Do you want the most accurate reproduction or the one that sounds best to you? Then what do you say about a case where speaker A has a flatter FR than B, but B has less distortion, better transient response, better imaging, better off axis performance, and a bigger soundstage.....which is the more accurate speaker now? It becomes gray area. Rather than argue about such things, I say just get the ones that sound best.

[terry j]

thanks Steve for jumping in,
looks like this issue is not that impoortant, 90 or so views, 1/2 dozen answers only.

Funny, this being in the forum which has room eq I would have thought that an accurate response was more important here than elsewhere but, just goes to show, never assume!

Steve, I understand what you're saying, my only respose would be that a 'better' designer (ie one who could haVE GOTTEN A FLATTER RESPONSE FROM THIS HYPOTHETICAL CHOICE OF DRIVERS) sorry about that, hate it when that happens!, a better designer can design a better network - in this case flatter - will give better imaging, better off axis performance,and a bigger soundstage.

All I was saying was surely, the mark of a good designer is one that can get as flat a response as possible.

So looks like this thread's interest has petered out, but thats ok cause as I said it was kind of a survey. Must say that I'm suprised at the relatively low importance attached to it

lots of love

terry

[JCD]

I'll jump in with my opinion..

I think a flat frequency response (ffr) is a desirable attribute. It's one of many and one of the more important ones imo. To me, it's (among others) what the artist/recording engineer meant for us to hear.

I think Steve makes an excellent point in that if a ffr is at the expense of extra distortion, poor imaging, etc, it may not be worth it.

However, all that being said, taste is king. If someone likes a fat bass signal, more power to them. It's not my preference, but who am I to tell someone what they like or don't like.

JCD

[Sonnie]

I wouldn't assume that just because members don't discuss a topic in a forum that it's not important to them. It may be that it's just not something they care to discuss... or there are other equally important topics they prefer to discuss.

Personally I have never heard what a perfectly flat (or even +/-3db) speaker sounds like. I've heard quite a few speakers in my lifetime, just never had the opportunity to listen to any in an anechoic chamber... or in a room where the response could be flattened. So I wonder, if the response is not perfect, would we really notice it over listening to the speaker in a room where it was unequalized and/or could not be equlized to flat?

I would think that it would make sense though, to design a speaker with as flat of a response as possible. It certainly appears to me that this is what the majority shoot for in their designs, so it must be desirable.

As far as "how it sounds to you" or "let your ears be the judge"... I don't think those comments are generally what designers are saying... at least that's not what I read. I would suggest those are comments coming from people who are choosing speakers for their own personal use.

[terry j]

fair call sonnie

not meaning to disparage anyone by my comment re views/reply.

Was always leaving out house curves in this discussion, if you want fat bass, and know how to do it ( rather than a boom box with q of 1.5) then that's OK. Rather I'm talking about covering up a poor design with appalling FR etc etc with "if it sounds good to you then it is good". That address is for backyarders who have no desire to improve their skills, or charlatans.

All I was trying to get across on the other forum, and here of course, is that given the same set of drivers in a box, two different designers, the one with the flatter response will be the better. That designer has better mastered the art of the speaker network than the other.

Of course going active is a whole new ballgame, and could very well be for another thread!!

Geez Sonnie, how do you keep up with all the posts, don't you have a real life???ha ha

lots of love

terry

[Otto]

Wow, I'm on vacation in Oregon, and I wrote out all the below last night about midnight PST. I went to "go advanced" to preview my post, and I got the message "We are temporarily unavailable while maintenance is performed. We will be back online shortly." Or something like that. I thought my post was lost, but I was able to salvage it! Anyway, here you go...

Hi guys,

First, let me say, terry, that this is a good thread, and I'll at least try to throw some more opinion to it. I'm not a measurement pro or spec-freak, but I do appreciate the information garnered from such metrics.

So here's a question -- would it be possible to design a speaker that has terribly accurate and flat FR, but which sounds like ? I'm asking here, I don't know the answer.

At some point, Kyle has said:

Kyle wrote:

So yeah, flat frequency response is paramount along with spectral decays and step responses.

"Whatever sounds good to you" is a cop-out. Any experts on these forums should give their expert advice to raise our signal above the noise.

I think I know what you're getting at... But I'm going to infer from your statements that you have implied that it would be possible to definitively measure a good speaker. To some extent, I agree -- I would imagine that if I had a speaker with FR "flat" from 20 to 20 kHz +/-30 dB, it wouldn't sound too hot (only one metric considered). In addition to FR, you have included spectral decay and step response (is that the same as impluse response?). There are, of course, other measurements that we haven't even brought up (and with which I don't pretend to be any type of expert). My question here is: are you saying that we can measure one speaker as better than another? In general, I think the answer is "yes", especially if one speaker has a horrendous metric (e.g., by +/-30 dB example above).

I listened to the Vandersteen 3A Signature today (30-22 kHz +/-1.5 dB)*. It was nice. Very nice. To me, it smoked the Magnepan 1.6QR (40-24 kHz +/-3.0 dB) and the Paradigm Studio 100 (44-22kHz +/-2.0 dB). No question in my mind (my wife also noticed a distinct difference, and even said the right words -- it was amazing). Was this because of the Vandersteens' flat FR? Or was it because of the Vandersteens' patented magnet mechanism on their midranges? Or was is because of the Vandersteens' phase aligned arrangement?

As you see, there's some slight variations in their FR. I do believe that the Vandersteen bass was better, more solid and more present than the others, as is reflected to some degree in their stated FR. But there's not a ton of variation in the FR between the Vandersteen and the Paradigm, or even the Maggie for that matter. So what was it that made, to me, the huge difference? (Granted, they were in different rooms, and were using different (but all very nice) electronics)... Was it some other measurable metric? If so, what? Back to my one of my previous questions -- can we measure one speaker as better than another? If so, we can presumably measure one speaker as better than [u]all other[/] speakers, right?

Or is there some other magic involved? Or have we just not yet discovered or employed that other metric?

That's it for now... Have a great day!

*Vandersteen FR info from official 3A Signature literature. Magnepan and Paradigm FR info obtained from their respective websites.

[terry j]

Otto, those questions were EXACTLY the type of questions I was hoping would be answered here - so why didn't you!!!

I asked the guys at DEQX the other day something very similar. I know, for example, that by analysing the impulse response, ( the whoop noise ) and performing an FFT on it ( evidently that means a Fast Fourier Transform, a mathematical piece of legerdamain ) enables the FR, group delay, phase response and time alignment to be extracted. ( these are all corrected by the DEQX system ).

My question to them was, did correcting the FR automatically lead to the correction of the other factors?? Unfortunately, the answer was no.The other factors are corrected by seperate algorithms within the program.

Those factors are, I suppose, at least the start of an answer to your question, and are other metrics that are of importance to the actual performance of the speaker.

I would imagine that technical queries such as these would be right up John's alley, so fingers crossed he gets some time to give some input here! I'd love to get his feelings on a good FR vs poor FR , all other factors equal.

By using the DEQX, the anechoic response of the speakers can be better than plus
or minus 1/2 db, with correct phasing, delay and timing. That tends to help the imaging! The most obvious immediate difference is the sudden 'seperateness' of the instruments and space that now permeates the sound. And this is in it's most basic guise, using the pre-existing passive network!!

Go active with it, and in addition to the above we can use linear phase x-over slopes of up to 300 db/octave!

Not wanting to get into a 'praise deqx' thread, but just by way of background this was how this topic started on the other forum, and I'm making veeery slow progress on being able to convince some others that the type of abilities mentioned above just mmaaaaay be of a tiny sliggghhhttt benefit, it's exhausting!

I've even been hit with the proposition that using something like the DEQX could detract from the sound.

I've simply come to the conclusion that these attitudes are firmly rooted in the past ( keep the signal path pure) and perhaps based on poorish results from using graphic equalizers in the seventies ( say ), and leading from that it's just pure ignorance of the capabilities of these type of systems.

Of course, now you have the 'background' ( which I originally didn't want to get into) that it came about from the idea of utilising a DEQX into an existing system, you can see why I've always said 'all things staying the same' when discussing the difference between a flat response and not. Cause the system does stay the same in this example, the only changes are the changes mentioned above made to the speakers. And for the life of me I couldn't understand there being ANY questions about the desirability of those changes, hence this thread.

The AFFORDABILITY of those changes are another question, yet when someone looks at buying a preamp for 80% of the cost of ( say ) the DEQX, without at least checking the DEQX out, is OLD thinking to me.

Be interested on others feelings regarding 'the way of the future', just to make mine perfectly clear ( and possibly stir up a hornets nest ! ) is not only my thoughts mentioned above, but that passive networks are also dinosaur!!

For heavens sake, a good quality digital xover can often be gotten for about the same cost as an audiophile passive network!! There is no comparison. The only people who would argue against that are maybe those speaker manufacturers that still use passive networks ha ha. There you go, got daring and was being deliberately provocative.

So it all comes down perhaps to the extra cost of the amplifiers, and once China gets into that in a big way then it's hard to see a future for passive networks.

Anyway, that's enough for now

lots of love

terry

[terry j]

Somehow couldn't edot my post so another quick one.

Otto, you asked can a speaker with a goo FR sound bad? I've heard the Wilson Watt Sophias over here. I'm assuming they are flat for the sake of this discusion.

I was totally and completely underwhelmed. NO attack, NO bite. It truly was like getting flogged by a cat'o'nine tails made of limp spaghetti. About the only thing that stunned me with them was the thought " Twenty thousand for THAT????"

So, I guess the answer, at least in my experience, has to be yes.

But, my counter would be that they would sound even worse if they had a poor response.

lots of love

terry

[JCD]

I've even been hit with the proposition that using something like the DEQX could detract from the sound.

I've simply come to the conclusion that these attitudes are firmly rooted in the past ( keep the signal path pure) and perhaps based on poorish results from using graphic equalizers in the seventies ( say ), and leading from that it's just pure ignorance of the capabilities of these type of systems.

Ok, that's another bowl of wax completely. I'm with you on this one. I'm firm believer in an active system. I can almost guarantee that my next DIY project will be a fully active, 3 driver plus sub system.

There are some that will/have argue(d) that they want to keep the signal analog and "pure" all the way through the system, but I think going digital would not be detrimental. I don't know if I'm in the majority (unlikely) or minority (likely) about that issue, but I do believe it falls into the "new is bad, old is better" attitude I've encountered. It mirrors the vinyl vs. CD and, to some extent, tube vs solid state arguements that crop up in these forums.

I've only heard two systems that were fully active. One was a 5.1 system (main l/r were mtm's with a bass box and then a sub) that used a Behringer 2496 active crossover. The owner, who shares/influenced my views, admitted that there is a noise floor issue with this unit, but after tweeking, he was able to eliminate the hiss and still be able to take advantage of ALL the benefits of the piece of equipment. He even uses it for some MINOR eq'ing to flatten is FR. Too much of anything is a bad thing, but a little nudge hear and there I think can squeeze some extra performance out any system.

The other fully active system I've heard was a pair of Mackie 824's. The one listed above is the best I've ever heard. In addition to a great set of speakers, it was also in a awesomely treated room. The 824's were also amazingly good. Especially for the price.

I've heard and read about that DEQX unit. A boatload of power in that piece of equipment. More than I would even know what to do with to be honest! The big thing is the cost. Man that thing costs a pretty penny. Probably means I stick with something far simpler/cheaper when I go active. Something like a Rane AC-23, although, maybe the Behringer 2496 might be more fun to play with..

Of course, the other issue with an active network is all the amps you have to buy. For example, for my planned 2 channel 4 way system (MTM, woofer and IB sub) I'll need 8 channels of amplification. It just adds up so quick! Being the cheap-skate that I am, I'll probably cheat a little bit in this area and go with Beheringer A500's. Since I've already got a 2 channel amp and a pro-amp for my planned IB, I'll only(!) need an additional 4 channels of additional power, but that's still gonna be another $400 for the system package. And forget about putting it in the living room. My wife has a cow about the system as is already! Anyway, I digress.

So, to sum up:

1. Go Active -- yes
2. Digital ok -- yes
3. Flat response good -- yes
4. Other "parameters" needed for a good system -- yes
5. Taste is still king -- yes

JCD

[F1 fan]

JCD wrote:

I can almost guarantee that my next DIY project will be a fully active, 3 driver plus sub system.

JCD

I don't think you will be disapointed if you go this route.


terry j, these measurements may show why you were not blown away by the Wilsons.They were done at Canada's NRC .http://www.soundstagemagazine.com/me...on_wattpuppy7/
Not very flattering for a $20,000 price tag.

For comparison here are some measurements of a very expensive active speaker. http://www.soundstagenetwork.com/mea...s_active_300b/

[terry j]

thanks F1

tho please don't get the idea I consider that I have got 'golden ears' and could easily tell it had a bad response, I don't and didn't. No, they lost me on sheer dullnes ( ie lack of attack), but you're right, you'd perhaps expect a better graph in an anechoic environment. Maybe we can't really expect better from passive???

JCD, the DEQX is expensive, but all power to you in going active. I too have been using the DCX 2496 for my setup the last few years, any 'disadvantages' in using behringer gear are far outweighed by the advantages in going active IMO.

I'm with you, nothing at all wrong with using Behringer amplification either, great bang for the buck. Sound quality??? well, if I'm in any camp on that one I must admit I'm more in the Richard Clark camp than not, so Behringer wins on that score as well for me. If I'm wrong, and some amps sound better than others ( an experiment I'd love to be part of) 1) is the better sound worth the extra money?? and 2) active with '' Behringer gear will I'd wager murder a passive set up with exotic uber amps..reckon the DEQX is expensive, what's a Halcro worth over there???!!!!

Looks like a little consensus is creeping in here.

lots of love

terry