Should Speaker Manufactures provide a Frequency Response Graph?
Eric at Tekton Designs has been battling two different reviewers who have posted measurements without his permission, using Klippel devices for their respective measurements.
It seems to me that if manufactures provide a simple smoothed out graph, consumers can see how much a speaker is editorializing with a frequency response that deviates from neutral.
MAYBE some people may like the sound of the speaker regardless of the graph. MAYBE it does sound good even with a not so perfect graph. MAYBE it's thinking outside the box and creating a speaker that has it's own unique signature.
ThomasAndStereo does NOT post any negative reviews. If it doesn't sound good to him he just sends it back.
This way the company won't get the benefit of a positive review and they also won't get slammed negatively in an open forum that could have serious impact on their business. Could even put them out of business on the opinion of ONE bad review.
"If you don't have anything good to say, don't say anything at all" because YOU might be completely wrong!
Posting a graph wouldn't be bad idea. I'm sure some would resist, especially those with poor graphs. I think if we had them all, Tekton would fall toward the top. Eric has always been proud of his flat frequency resonse.
They already post numbers like "40-23000 Hz +/-2 dB" so show us the graph.
I agree, buy speakers based on how they sound but as a physicist, I would likely use graphs to choose which ones to check out. I certainly would have a problem with one that had a big dip in middle C, for example.
I think those who are having fun criticizing Tekton would not really want to see a graph of their own speakers.
The amplitude response of a speaker is a moving target. It will change depending on the room. There are also other factors that play a significant role in audio fidelity like phase behavior and group delays. I do not pay attention to specs. I review the design of the speaker and make the best determination I can that I will be able to make the speaker perform to my expectations.
So, the most important marketing tool should be measuring frequency response in an anechoic chamber.. This means, once they are done designing a speaker to meet your measurement requirements, they are done with their design work. What could go wrong?
I think they have a lot more to publish before the frequency response graph. Honestly it's a mess what they don't publish. Power handling, sensitivity, impedance, crossover frequency, prices, where to buy, decent photos, company info, contact there is so much that's missing
Should Speaker Manufactures provide a Frequency Response Graph?
Yes, my speakers came with graphs! For power handling, sensitivity, impedance, crossover frequency, efficiency, directivity, dispersion, power linearity, distortion, ect..😎
Meh. As a person with 3 measurement mic's in his house, I wouldn't trust the speaker maker's measurements to begin with, and don't think most would interpret them correctly, and what good is a quasi anechoic (or fully anechoic) measurement when it comes to my room?
I'd much rather see off-axis frequency plots as done by Stereophile and others, as well as dynamic range plots.
Eric at Tekton Designs has been battling two different reviewers who have posted measurements without his permission, using Klippel devices for their respective measurements.
So what? They posted measurements without Eric’s blessing/approval? Oh the horror! It’s a free country dude so comes with the territory. Other “manufacturers” (ehem) get by just fine without threatening litigation except for, uh, Bose. Go figure.
I’d really like to see an impedance graph along with phase angles. A frequency response graph doesn’t mean squat if my amp can’t adequately drive the speakers. Nominal impedance??? C’mon man.
The waterfall plot displays both the frequency response of a loudspeaker (in various dimensions) and it’s behaviour in time. It is very revealing of any "ringing" in the speaker’s response, ringing being a major cause of the lack of transparency a loudspeaker exhibits.
I agree with @bdp24. After looking at waterfall plots for years, I feel like I can get a good idea of a speaker's amplitude response even though of course in the end, you have to listen to it.
I also agree with @bdp24. A waterfall plot provides the most revealing graphical representation of a speakers frequency response. However, unless procedures are standardized, comparisons between manufacturers/models will have latent errors. Also, on and off quasi-anechoic measurements do not represent real world response. While critical during the manufacturer’s design phase, quasi-anechoic response graphs are useful for to attempt to make some preliminary determinations if a speaker may meet our listening preferences. Audition in a real world environment, and if possible your own home and system is necessary.
The full measurements that the kipples spit out are super valuable. At $100k-ish they don’t cost much for a business to own. It is a way for smaller (compared to Harman, Kef, etc) companies to do measurements and make better speakers. It is much cheaper than anechoic chambers which can be upwards of a million dollars. When companies want to sell speakers for $30k+ I expect them to invents money and time into design equipment.
"Nope. I buy speakers based on my hearing, not a graph made in a testing lab."
This! I agree completely.
"The full measurements that the kipples spit out are super valuable. " In what way? Do these measurements tell me I will like what I hear? No! I am not an ASR minion that needs to be told whether a speaker sounds good or not.
The only figures I look are is the efficiency and the nominal impedance as they will give me an idea if my amplification will drive a speaker without any issues, then I just listen.
If graphs are important to you, check out Ascend Acoustics. They provided them and even included a copy of my specific pair of raal towers in the shipping box. That was more than 10 years ago. I even discussed it with the owner. They were just slightly off the one published, but even better. I no longer own these but have only good memories of them.
Graphs and measurements can be important - but they are not the be all and end all. I remember reading a long review on Audio Science Review Forum that gave an exhaustive analysis of a myriad of measurements and then made conclusions and recommendations based on those measurements. Nowhere in the review did it mention how it sounded. It left me scratching my head in wonderment.
Imagine a restaurant review that tells you how the food was cooked and what it looked like, but never mentions how it tastes.
No. But they should include a warning label stating that their sound can be highly addictive, thus straining relationships, creating an additional financial burden, and many hours of sitting in the "sweet spot" can be detrimental to your health.
I agree that one measurement (out of a possible 7?) does tell the whole story. Horsepower and torque specs when selecting a automobile can be "useful". But, we did reject a vehicle because we didn't like the cupholders.
If you go to the drag strip and test run your car and you post your 0-60 and 0-100 run, I wonder if you will get in trouble with the manufacturers. And thus, the conversation.
The most important graph for me is the Impedance & Phase vs frequency graph. I want to know if my amp needs to drive 2 ohm loads with a difficult phase angle. I believe it to be the most important data when choosing a power amp. next up is sensitivity. That too dictates my amp selection. Then I can vet by sound quality to my ears.
I do ears and brain, not graphs and numbers. I don't know the measurements of the gear I've got here and I don't much care, I just know it sounds awesome.... But if people want them, they should probably be available, with the usual caveats....
“In what way? Do these measurements tell me I will like what I hear?”
Yes, they do tell you what you hear and estimate in room too. Each room is a bit different though so there is some variance which is mostly 200hz (not to mention some of us have HF hearing loss) and below. once you hear a few speakers that have full measurements and decide which ones you like better it is pretty easy to look at other measurements and know which ones you will like better and seek a demo based on those measurements. You can even look at the dispersion and understand what the soundstage is going to do.
I don’t read or post on ASR either, as I don’t like the absolutist attitude and whiny commentary on their but there are a bunch of knowledgeable people there just as there are here (many of the same people…)
Lack of measurement understanding does not make measurements untrue. Some things are very subtle in the measurements and it takes time to understand them.
Some things are still lacking in the measurements, I feel transients and detail is still missing but measure and use the data you have and demo for the other things.
I have owned two speakers that have both had full spins. The Revel 228be and the JBL 4367s. I did same room/system direct A/B comparisons. Listening to them back to back and looking at the measurements is almost 1:1 how they actually sounded in my room.
I bought the revel after demoing and bought the JBL 100% blind based on the measurement…. I kept the JBLs and sold the Revels after direct A/B in my room.
As engineer for 20+ years I can assure you 99% of the high tech things you use (or even keep you alive, I work in the biopharmaceutical industy) were build or designed using models.
I agree with erik_squires. A frequency response chart made in an anechoic room has little to do with how that speaker will sound in your room with your amplifier. Besides, the impedance of the speaker is changing with frequency which will affect how your amplifier/speaker cables sound coupled to those speakers.
Back in the early 1990's I built my first dedicated listening room and I added sound dampening to make my room have a flat frequency response. The room was very close to anechoic. The music sounded dead and lifeless. I had to remove much of the dampening material.
That's my point, I guess. If you want flat frequency response then you will need a hemianechoic room like the room where they test speakers. It will not sound so good. Think about live music. No band ever plays in an anechoic room. Life is messy and so is music.
Most garbage speakers I have seen (and heard) had 3 things in common: Cheap connectors, a little tab to pull off the velcro attached front grille, and a frequency response graph showing flat response from 30Hz to 22kHz.
Some of the responses here display the anti-science attitude which I find very frustrating in this forum. The vast majority of speaker manufacturers will measure the frequency response of their products, as well as a host of other parameters. An estimated in-room response (based on a large number of on and off axis frequency response measurements), such as that measured by Klippel, actually will do a pretty good job of indicating how the speaker will sound in an "average" room. It doesn't tell you everything, of course...distortion, directivity, impedance, sensitivity...all are important. I wouldn't buy any speakers based on data without listening, but I would certainly rule some out based on data.
"As engineer for 20+ years I can assure you 99% of the high tech things you use (or even keep you alive, I work in the biopharmaceutical industy) (sic) were build or designed using models."
I do not doubt this, however I do not need to know and am not interested in measurements; I am interested in how they sound to me and whether I like that sound.
"Some of the responses here display the anti-science attitude which I find very frustrating in this forum." - see my comment above.
A major consideration where run of the mill measurements completely miss the mark is the materials used in the drivers themselves. For example, Kevlar, polyprop, etc used in drivers that measure perfectly...you might as well throw a blanket on your speaker before you started listening.
This is one of the reasons i don’t like the low effort/resources brands who buy mass market drivers from someone else and put it in a box. It takes some guy who has put serious thought into the materials used in drivers and developed his own from scratch.
We could go on about crossovers, etc... barrel bottom quality components in crossovers can still measure perfectly. With sim software, you can very quickly design great crossovers, But, as you move up in quality of components used, the sound totally changes, i.e. the measurements say little about "quality" (whatever that word means to you).
Last, but not the least, humans perceive way too much in transients (spatial nuance, etc). The configuration (Drivers+crossovers+box) that gets botched w.r.t the latter, no matter how great it measures with traditional measurements will sound like a ..."meh".
The above mentioned perhaps falls into the "esoterica" (dismissed) category for the ASR type of guy, another reason i despise that loathsome forum.
Manufacturers should include specs and explanation of how measurements were made for reference. It adds credibility for their design philosophy, then have a listen. Why would anyone want to look a Porsche 911 specs or its track performance? They are just measurements, my family vehicle feels faster.
@erik_squireswrote: "I'd much rather see off-axis frequency plots as done by Stereophile and others, as well as dynamic range plots."
Yes! The off-axis information in particular is imo extremely useful.
I WISH John Atkinson did not "normalize" his off-axis curves, as that makes it harder to see the overall picture of what's REALLY going on. I think that for one or maybe two issues he posted off-axis curves that were NOT normalized to the on-axis curve, but then he went back to posting normalized curves.
I couldn't care less, simply because their measurements aren't in my room which is always vastly different. I also don't put much stock into the tech specs which tend to be off as well.
So if they or ASR wants to post frequency curves that's fine, neither will tell me what they measure in my room.
I have never been able to discern the voicing of a speaker from performance specs.
From manufacturer specs? That makes sense.
However, it is not difficult to do with thorough third-party measurements. It merely requires an understanding of how the measurements correlate to perceived sound quality. For example, take any two tweeters and provide a CSD/waterfall graph for each. Based on the graphs, I could tell you with certainty which one will sound cleaner and more refined.
Another example: a speaker with a wider horizontal dispersion over a large bandwidth will produce a bigger soundstage than one with a narrower dispersion over the same bandwidth.
“It merely requires an understanding of how the measurements correlate to perceived should quality“
Good luck with that. A panel of listeners will rarely, if ever unanimously agree on the sound quality of a speaker or system. On top of that, any speaker will sound different based on the room, the setup, the amps, the cables, the source, condition of the power coming into the system, etc.
“…take any two tweeters and provide a CSD/waterfall graph for each. Based on the graphs, I could tell you with certainty…“. No, not with certainty. You might guess right some of the time.
“… a speaker with a wider horizontal dispersion over a large bandwidth will produce a bigger soundstage than one with a narrower dispersion over the same bandwidth.“. Not really. Careful, you are using lots of fancy terms here without context.
”The amplitude response of a speaker is a moving target. It will change depending on the room. There are also other factors that play a significant role in audio fidelity like phase behavior and group delays. I do not pay attention to specs. I review the design of the speaker and make the best determination I can that I will be able to make the speaker perform to my expectations.”
“I might also add that measuring your system in place is much more informative.”
This is how I roll. However, when I was much younger, I was a technical specification person. If the device had low THD/IMD and depicted a ruler flat response, to me, it must be a good thing. As we grow and mature in this hobby, we now know through experience that that methodology is not entirely accurate as to what we actually perceive as musicality in the actual operating environment which has numerous variables.
There is one caveat. When auditioning very large speakers, multiple listening samplings at various locations / venues must be conducted in order to grasp the persona of that speaker. When comparing and contrasting a speaker you have owned for many years to something different, it will be readily obvious that what you hear is either good or bad in a fairly short amount of time. That’s why when I go to audio shows and walk into a room, if it sounds bad to me, I am out in either a few seconds or a minute. My comparative reference is different than anyone else and what sounds good to me, may not necessarily sound good to someone else.
What should start a new thread call titling it “What have you learned” or something like that.
For what reason? Are you going to buy because of the chart?
Great question @bigkidz....The chart is more likely to cause me to avoid a particular speaker. If for example the speaker has a rising response due to lack of baffle step compensation....that's not for me. I know that speaker will sound too forward and possibly even shouty.
So if they or ASR wants to post frequency curves that's fine, neither will tell me what they measure in my room.
@sunshdw Your speakers will dominate the measurements until the room starts to take over. If you have a larger room, the room will start to dominate the response at around 500hz....for a small room it's around 250hz...
However, if you have a very small room or speakers very close to sidewalls for example, you're going to end up with phase cancellations or peaks at much higher frequencies.
Lets say a speaker has a 5dB frequency response error due to a rising response...and then your small room is reinforcing that peak another 5dB...at that point a 10dB peak in the mids or higher frequencies will make the speaker unlistenable...this is a core argument for mitigating frequency response errors at the factory.
@seanheis1It is rare now for speakers to have a 5 dB anechoic amplitude error except at the very ends of the audio spectrum. On the other hand early room reflections can cause errors over 10 dB. I had a window adjacent to the right channel speaker I had to remove because of a 12 dB response error it caused above 10 kHz.
Amplitude errors do not just occur in the bass. They are usually more pronounced in the bass, but if anything in the room is resonating, like my window, strange things can and do happen. Using "room control" only in the bass is inadequate. It has to cover the entire audio band. When a processor can perform room control an added benefit is that it can also do EQ given the user the ability to voice their system the way they want it.
I want to also add that the knowledge of what an audio system is or should be capable of is gained only by experience. You have to have heard a system performing at that level. Live performances can be useful, but only if acoustic in nature. Once you add amplification, particularly in large venues, all bets are off. The sound quality at large stadium concerts is typically hideous and in mono. A good system with a good live recording has no problem outperforming that circumstance. IMHE, as an audio insider in the late 70's and 80s, experienced listeners always know when they are in the presence of an exceptional system. It only takes 30 seconds of a familiar recording, you can see their eyes widen, and exceptional systems will do this to every honest experience listener that hears them. Not that these systems have no faults, their faults can be glaringly obvious, usually in the low bass, but their performance extremely convincing none the less.
@seanheis1- I am not sure what the graph will tell you. You may know more than I do in that regard. BUT in manufacturing equipment, I guess I can tailor components to produce the sound I prefer. I am not convinced that you can tell from a graph what the overall sound will produce. Maybe an emphasis but not the overall sound. Magico speakers use computer modeling I think and they changed the way that model works recently so who knows. Vandersteen uses more of his ears to design and finalize what those speakers sound like. I can only go by our experience on what we have heard, modified and repaired which has been a lot over the years.
@bigkidzif you manufacturer hi-fi components...you would notice if the response of an active component was +/- 3dB. Having said that, I have never heard of a component that creates such large frequency response errors....maybe a tube amp with a very low damping factor trying to drive Magnepans would create some big response errors. ;-)
You must have a verified phone number and physical address in order to post in the Audiogon Forums. Please return to Audiogon.com and complete this step. If you have any questions please contact Support.
