During my working career I had the great pleasure of hearing some of the great audiophile equipment around the world and in a few cases in “perfect” rooms such as the listening room at Sony in Japan. I can tell you that I don’t always like a flat response curve best, even if the gear is very good, and especially as my 64 year old rock damaged ears get worse and especially once your set up in a given room. I don’t know about you guys but I do this in part because I like to learn technical things, I like to save money and I can make exactly what I like. I use a technique suggested here when building guitar amps. I put an active multi band equalizer in the chain during breadboarding to help me find problems and tune the final design. Sometimes my preferences surprise me. Some signal chains just struggle with the IM generated with some dissonant chords more than others for instance. Another thing to consider is the predominant source material of the listener. I built a value ported metal 2 way from Zaph and was blown away while listening to some very well recorded acoustic jazz/vocals. This $300 pair (I do my own woodwork) is a crowd favorite around my house. With some of my favorite rock that is generally much lower overall production quality, not so much. There are elements to the sound character and our brains interpretation that are not in graphs and charts or are very difficult to model well.
Another example for me is that I almost never met a ported speaker that I like. The bass extension out of the smallish box is great but I just don’t like the character of the bass as much as sealed boxes and more brute force. There is probably a ported design out there somewhere that nails the transient response, cone dampening, or whatever else it is that my brain doesn’t like but in the meantime I’m back to working with sealed boxes and good subs. I sold my Krell amps on eBay, just don’t like them, they always sounded sterile and a little harsh to me. Now the Mark Levinson monoblocs I heard in Japan were great but I don’t have $30k per channel so in the meantime an old used Aragon 8008 is one of my favorites. I’ll probably never completely know why but I am probably tweaking my speakers to sound right with that amp.
There are so many factors that deliver the ultimate sound experience to our brains, some of which are not in charts. We all live in a crazy rabbit hole here together, I hope we can find our way out. In the meantime, if a chart that is not flat sounds good to you then by all means, build that one.
In the meantime I will go back to my crossover designer and keep trying to get the flattest possible curve with the least amount of components like I’m locked in an online video game battle for the world title. I think speaker design is giving me OCD. LOL.
George
Another example for me is that I almost never met a ported speaker that I like. The bass extension out of the smallish box is great but I just don’t like the character of the bass as much as sealed boxes and more brute force. There is probably a ported design out there somewhere that nails the transient response, cone dampening, or whatever else it is that my brain doesn’t like but in the meantime I’m back to working with sealed boxes and good subs. I sold my Krell amps on eBay, just don’t like them, they always sounded sterile and a little harsh to me. Now the Mark Levinson monoblocs I heard in Japan were great but I don’t have $30k per channel so in the meantime an old used Aragon 8008 is one of my favorites. I’ll probably never completely know why but I am probably tweaking my speakers to sound right with that amp.
There are so many factors that deliver the ultimate sound experience to our brains, some of which are not in charts. We all live in a crazy rabbit hole here together, I hope we can find our way out. In the meantime, if a chart that is not flat sounds good to you then by all means, build that one.
In the meantime I will go back to my crossover designer and keep trying to get the flattest possible curve with the least amount of components like I’m locked in an online video game battle for the world title. I think speaker design is giving me OCD. LOL.
George
Last edited:
Dr. Floyd Toole curve is a really good reference, but my current DSP setting is a little off from that. Even if you are not 100% satisfied with this curve, I think it's a good idea to start voicing from this, and it will always be a good reference point.
I think this is called the "circle of confusion". Siegfried Linkwitz and Toole have talked about this. It introduces a huge amount of subjectivity and in some instances wrecks the sound of the final mix. It's one reason why mixing should be done with flat and neutral monitors that meet some kind of in-room response standard. By the same token it is better to use measurements to design the speaker and aim for an anechoic flat-on-axis response AND controlled directivity if what you are seeking is something that can sound good in a variety of listening spaces.
There are lots of ways to get your favorite recording to sound good in your room with your own loudspeaker set up at one listening position. The results might not be so good when you start to move away from that particular combination of variables (room, position, music track). But DIY audio is for your own enjoyment, and how you find that is all up to you.
That's a real good gedank-experiment but not for the reasons you think.
Sound in an anechoic chamber is awful and it isn't the FR*. Of course, I have nothing but my long-ago experience in once world's largest.
B.
*I suppose if you recorded anechoicaly and then played in normal rooms it might be OK - as is actually the case with many pop and vocalist recordings folks think are really great after pseudo-echo is synthetically added.
Hold-yer-horses right there, CharlieLaub. I'm the guy forever posting, "Where are your measurements". You mistake high-school fantasies of "engineering" for proper rigorous analysis. Whether you like to admit it or not, humans are in the loop.
Without moving this thread too far, I'll just say one obvious thing: we want to assess the listener's perception of correct FR in their room, not the near-field mic FR.
Pity the new acoustics forum has been hidden away elsewhere in this website. It should be part of loudspeakers.
B.
Hmmm, OK, I seem to have misunderstood your previous post. I think.
I'm not sure who said that a nearfield measurement is "the" metric (in this thread)... Unless that is where you place your head when you listen it is not representative of the FR at the listening position.
But aside from quibbling about measurements vs human perception of sound, do you agree that it is clear that certain aspects of the loudspeaker's radiation are tied to the design and cannot be corrected/adjusted later (e.g. by the crossover or EQ)? Furthermore, with not all (types of designs) being equal, there must be some designs that are better in terms of how they relate to perceived sound quality, and in fact that listener-based (e.g. human preference) and measurement-based research into sound quality has been conducted at Harmon and the conclusions made public?
I am referring to this (I am sure you are aware of it, Ben, this is for other readers):
https://www.amazon.com/Sound-Reprod...-Loudspeakers-Engineering-ebook/dp/B074CHY128
Highly Recommended Viewing - Lecture from Toole:
YouTube
The above video includes the "circle of confusion" I mentioned earlier.
I see this argument often. I think it has been blurred by the use of DSP for active crossovers. If the DSP is providing the crossover AND the room corrections then it becomes hard to separate room correction from crossover in the discussion.
If someone is NOT using DSP heaven forbid a passive crossover 😉 Then the argument is very different! I do not believe that a passive crossover (or an analog active crossover for that matter) should be tuned to the room. IMO If you want to adjust the speaker to the room you should be using some other form of EQ quite separate to the speaker.
If the speaker has been well engineered, then this EQ'ing should be much easier without resorting to heroics that may do more damage than good.
FWIW I just live with my room effects, and have minimal room treatments. I have speakers that measure pretty flat on axis semi anechoic. The one issue I have is that they have a bit of a hump around 500 Hz in room due to room gain (I went a bit too flat). But it has never been enough to make me redesign the crossover (yet). I have an active analog crossover to my dual subs and passive on the MTM's.
Tony.
If someone is NOT using DSP heaven forbid a passive crossover 😉 Then the argument is very different! I do not believe that a passive crossover (or an analog active crossover for that matter) should be tuned to the room. IMO If you want to adjust the speaker to the room you should be using some other form of EQ quite separate to the speaker.
If the speaker has been well engineered, then this EQ'ing should be much easier without resorting to heroics that may do more damage than good.
FWIW I just live with my room effects, and have minimal room treatments. I have speakers that measure pretty flat on axis semi anechoic. The one issue I have is that they have a bit of a hump around 500 Hz in room due to room gain (I went a bit too flat). But it has never been enough to make me redesign the crossover (yet). I have an active analog crossover to my dual subs and passive on the MTM's.
Tony.
Guess you haven't crafted a system using DSP?
To start with, passive crossovers are a crap-shoot and in my opinion can't be designed to work too well even in theory given the phase motions, variable impedance, little resonances, and - ready to hear this - the room unless you hand-tune the crossover; except that's why fancy brands use 18 elements in their crossovers which they've hand-tuned to their market. You can do the same (with enormous testing effort) or you can use DSP.
With DSP, you run your drivers with no crossover (carefully) and see what you are getting at your seat (using a mic to speed things up). You'll soon enough see what they can do. Last time I tried it (vacation home) I had just bought some silk-dome tweeters. My goal was to give as much of the sound compass to the domes as they could handle. No trustworthy way to sim the answer to this question.
With the crossover region chosen, you can set up loudness levels in your DSP. Then you try different slopes and you can even use different slopes and different crossover points for bass and treble. Model that in your sim.
We all just love our systems - even the ones always shown in this forum with hard surfaces, few carpets (never wall-to-wall with true horsehair felt underlay) and, I bet, inevitably hideous reflections* that everybody gets to love. Called "the burnt soup principle" (like mother's soup) or among psychologists, "adaptation level".
B.
* good reason to favour large dipoles, planar panels, or ESLs
Last edited:
Many people seem to blindly believe acoustical correction (speaker design) should be better than electrical correction (EQ), but I'm somewhat skeptical about this idea after correcting horrible frequency response vintage high efficiency speaker with a digital EQ for flatter response.
In this digital age, We DIYers who can access to digital tools may want to rethink if the whole concept of the modern flat response speakers to be really a right direction.
In this digital age, We DIYers who can access to digital tools may want to rethink if the whole concept of the modern flat response speakers to be really a right direction.
Hi Ben, you are correct, I have not used DSP in any systems. You will note I did not rubbish their use though. Just pointed out that there are crossovers and there are room corrections, and these things should not be considered as the same thing. If you take your speaker to a new room (say you moved house). Do you design the crossover again from scratch, or do you just tweak the room corrections?
I would be so bold as to guess that you have never designed a passive crossover using the tools available to do it right!
You are welcome to your opinion, but I'm sure that there are a very large number of both amateurs and professionals alike who would disagree with you!
Tony.
I would be so bold as to guess that you have never designed a passive crossover using the tools available to do it right!
You are welcome to your opinion, but I'm sure that there are a very large number of both amateurs and professionals alike who would disagree with you!
Tony.
Do the math: I've been making or modifying speakers for 60 years and DSP has been in my system for maybe 10. I think you can answer your own question now. Yes, it is pretty bold (?) to say you didn't think I ever designed a passive crossover.*
Tools to do it right? Since quality manufacturers often go to like 18 elements in their two-way crossovers (starting, I believe, with the famous BBC little box), can you show me any tool to "do it right" that guided them? BTW, the "theory" of crossovers is trivial and can be done in your head; and anybody who knows the significance of the numbers "6.3" and "160" will understand.
But to your core point, if the new room is different enough from the old that what worked in the old no longer sounds right in the new, then YES, you need to fix your system whether cobbled together your way or the simple, easy to modify way with a mic and DSP.
There is a confusion running through this thread that it is inherently nobler to get each piece's FR perfect than to simply make the final result optimum. You do have to design some pieces right with respect to some of their properties. But FR will be so monumentally affected by the room and by how you hear tone colour, it makes no sense to finesse the little pieces.
BTW, after you've designed the theoretically simulated perfect crossover, when you open the parts box, you might discover the values of the parts are all over the map. Or you might not discover that if you don't have the tools to measure the parts before you wire them in permanently.
B.
*well.... OK... but not for many years since I've been multi-amped with my own active crossovers for quite a while, even in my vacation condo.
Last edited:
Ben the core point was "using the tools available to do it right". I did not say you had never made a passive crossover. (perhaps I should have said using the tools that are available that will help you to get it right)
Your comments make me think you are stuck in the world of 1970's passive crossover design, where you do some crude calculations, put something together and start to tweak things because it does not sound right.
The really ironic thing is that the many of the same tools and techniques are needed to do active properly too whether it be analog or dsp.
Tony.
Your comments make me think you are stuck in the world of 1970's passive crossover design, where you do some crude calculations, put something together and start to tweak things because it does not sound right.
The really ironic thing is that the many of the same tools and techniques are needed to do active properly too whether it be analog or dsp.
Tony.
Last edited:
Hello Bentoronto
LEAP it's quite an amazing piece of crossover design software! All you need is a good reliable measurement system to go with it. I use CLIO and LEAP and it's a really good combo. LEAP comes with an awesome set of tutorials to get you going and after that you learn from doing.
Rob 🙂
Last edited:
What do you guys think about LEAP vs SoundEasy. I’m thinking of upgrading. I currently have:
DATS
REW
Xsim
Jeff Bagby excel
Various DIY calculators
Thanks,
George
DATS
REW
Xsim
Jeff Bagby excel
Various DIY calculators
Thanks,
George
You are too "kind" to me. Your put-down could have said I was stuck in the late 60's when I made my first electronic crossover using parts from the fabled stockroom of my employer*.
Instead of posting personal insults, you could post substantive evidence like REW measurements of how great your results have been.
I've spent more time than I'd ever want to admit playing (and measuring REW) with DSP crossover choices. I certainly couldn't reliably tell Linkwitz-Riley from Butterworth and barely can A-B 12dB from steeper slopes.
On the other hand, there are monumental changes introduced by the room once you install your speakers.
LEAP, on first examination, seems to take loosey-goosy inputs and calculates parts values to three decimal places.
B.
*Bell Labs, Murray Hill, see "The Idea Factory" book; staff called the stockroom "the gift shop" (don't repeat that)
Last edited:
The only thing loosey-goosy about LEAP is if your measurements are not accurate. Other than that it's predicted curves are as good as your measurements. LEAP was one of the tools Greg Timbers was using at JBL.
Bi-amping the JBL Model 1400 Array
And a couple of screen shoots
Rob🙂
