Pavel,
Thanks for the reference numbers. One clarification, is there a difference between using a pair of test tones to determine audibility and using music with reference amplifiers.
My experience has been listening to computer generated DSP files of music containing precise amounts of distortion, but not under the best listening conditions.
Scott,
I was not quite precise, I actually am selling something on this thread. One is Audioxpress as it astounds me how many folks here seem to be ignorant of material it has covered so well for so long. At $25 a year it is one of the best bargains around. The second is Jan's Linear Audio. A lot of folks have been looking at the graphs of resistor distortion and I think the full story is scheduled for the next issue. The first issue of course contained much useful material and a few pages wasted on my stuff.
I though I had sent you an advance copy of the resistor data, as I thanked you for your help. (Or am I spam filtered out?)
You also reminded me of an important issue no one seems to pay any attention to:
Open for everyone: "What is the most common speed of an induction motor in the U.S.?" "Why is this important for sound system design?"
John,
Western Electric made the first "talkies" sound systems. They used all of the back of the screen depth for the low frequency horn. It was molded into the plaster walls. The "receiver" was used to drive a separate high frequency horn. The mercury vapor rectifier case was placed outside the projection room in a well ventilated box to reduce noise. The amplifier had a current meter and adjustment knob on the front so the projectionist could keep the output tube properly biased during the show. The owner's manual that came with the system started off by mentioning several tips on running the show such as "Always start the show at the same time every day." "Be sure to collect the money as the audience enters."
Jan,
Can I send you the bill for my drinking also?
ES
Last edited:
There was an era when little or no electronics were involved on either end of reproduction. There are probably better folks to discuss this. Things like electro-mechanical pianos where you can actually hear Mahler playing his own music.
Ed yes I got it, I thought I made a few suggestions. We have different approaches to this. I would make a model with all non-idealities included and then try to separate them, the magnitude and phase vs frequency of the thirds is a good start to separate the thermal effects from voltage ones. With enough care you could even make a good guess at the thermal capacity and time constant. Since no one would use carbon comp resistors anyway, showing their badness is mostly curiosity. Sorry about not following your logic, but the experiment of using a 5W resistor with a bad voltage coefficient does not extrapolate to using a 2 or 5W metal film feedback resistor in a 400W power amp just to make sure there are no thermal effects. If a 1/4W Dale has a few ppm at 50V rms I'm sure a 5W one would be better.
Last edited:
I am not here to give an entire course in horn theory or its history, but it is important for audiophiles to understand some of the concepts, AND where they might have gone wrong, in loudspeaker design. This is why I have deliberately left out a number of loudspeaker companies from MY discussion, because they don't add anything important to the points that I am trying to bring out. I am NOT writing a book, just a short chapter, here, if I get a chance to do so.
Ed,
I meant that a listening test (using orchestral music) showed audibility of CCIF IMD 19+20kHz measuring some 0.03%-0.05%, vs. some 0.003%. The higher distortion made sound subjectively 'bigger' and 'smoother', BUT with troubles to clearly identify similarly sounding instruments, like oboe/english horn, cello playing lowest string notes (confusion with double bass playing same notes).
P.S.: the distortion profile was low order, with mainly 2nd and 3rd. 4th and higher with fast decay.
Last edited:
Yes indeed, Linear Audio Vol 1 (that's the second issue of course, Vol 0 being the first issue
) will have Ed's complete measurements results and conclusions. With a few unexpected twists. And of course Scott's contribution, surely to cause a lot of discussions.
And a class-A amp with 67% efficiency. And lots more.
Hey, you want to bankrupt me??
jan didden
Moving on, back to the mid 1960's and beyond.
Horns were, and often still are most often used in PA systems. Improvements have been made over the years. It is now possible to have 'time alignment' to a degree, even with an all horn, or a direct radiator, horn combination. This was overlooked by Paul Klipsch, because he was taught, when he was in college, very specifically, that the ear was monaural phase deaf, and there were tests that proved it. In other words, if you could not hear a tap going through a speaker system as 2 or more separate events, the spacing did not matter. This was called: Ohm's Law of Acoustics.
However, my associates and I found, over time, that a 2 foot, for example, path difference WAS audible, and annoying even, especially if you knew the person's voice personally, and this happened with me listening and working with Jerry Garcia, and then listening to my mastering tapes with his voice on the K-horns. It is enough to drive you crazy, once you become sensitive to this. That is WHY I sold my K-horns, 30 years ago, and went back to direct radiators, with all their attendant problems. more later
Horns were, and often still are most often used in PA systems. Improvements have been made over the years. It is now possible to have 'time alignment' to a degree, even with an all horn, or a direct radiator, horn combination. This was overlooked by Paul Klipsch, because he was taught, when he was in college, very specifically, that the ear was monaural phase deaf, and there were tests that proved it. In other words, if you could not hear a tap going through a speaker system as 2 or more separate events, the spacing did not matter. This was called: Ohm's Law of Acoustics.
However, my associates and I found, over time, that a 2 foot, for example, path difference WAS audible, and annoying even, especially if you knew the person's voice personally, and this happened with me listening and working with Jerry Garcia, and then listening to my mastering tapes with his voice on the K-horns. It is enough to drive you crazy, once you become sensitive to this. That is WHY I sold my K-horns, 30 years ago, and went back to direct radiators, with all their attendant problems. more later
PMA have you seen a spectrum of 1/3 octave noise at 15kHz off of vinyl.
Yes.
Regarding noise, narrow-band noise is the signal where you may check distortion audibility quite successfully. In a controlled test
.Regards,
Agreed as long as it is really made by the same process.
As it takes about an hour to run each test, I probably will not take it much farther. Besides I think we also agree once distortion is lower than 24 bits or 147db, we probably don't need any better.
But it does seem the real distortion in a circuit is 12 db higher that shown on my graphs. My math says it should be 6, so one of these days I may recheck it.
PMA have you seen a spectrum of 1/3 octave noise at 15kHz off of vinyl.
... and once again, vinyl distorts horribly, which I have already written here, and started an avalanche of disagreement
PMA, while your data seems OK, I think that you are over-stressing the amount of 3'rd harmonic that can be present and not even noticed, for the most part. Loudspeakers certainly generate it, as well as analog magnetic tape, where virtually all music recorded before 1980 comes from, except for perhaps direct disc and its problems as you are well aware. Would you throw away the 'baby' with the 'bathwater'?
Hi John,
I just think you are assuming too much about what Paul Klipsch was all about and his decisions. K-Horns have their pluses and minuses and you either really love them, or hate them. They aren't my taste either, but Klipsch did make some crossover improvements later on that addressed some issues. They even learned to raise some x-over frequencies to reduce the honk a little more. Their latest horns that I have heard (Tractrix ? sp) are far more natural, but still not my cuppa. I could listen to a system that used these now.
The length of path issues are something not easily overcome in horn designs, and it's a specific problem that is well known. Since you're covering this material, why not cover it as the issue that it is, rather than as an example of a design "mistake" that it really isn't? Horn designs can be examined without the idea that these issues are an optional design problem - which they are not. It's the nature of the beast and to gain some performance, trade-offs are made in other areas. It's nothing more than that and Paul Klipsch was not a bad designer. Nor was he blind to these design issues. What would you expect him to do?
Analog tape vs digital recording is another example of design trade-offs. Recording in the digital realm has a completely different mind set from analog tape recording. For example, when you run out of 1's and 0's, nasty things happen in digital. Analog tape handles overloaded signals in a more benign manner. Today's recording engineers have learned, but in the early days, some truly evil things happened with digital gear. Low level signals had their own devils, like reduced resolution in bits instead of simply increased background noise. Back then, a Studer equipped with Dolby SR had a wider signal to noise ratio than the 16 bit A/Ds and D/As did. There was no contest between digital in the 80s and 90s to a good analog setup.
It took several tricks to allow digital technology to compete successfully with good analog systems. Back then, it was just simply ugly. I spent some time at "The Metalworks" recording studio in the 90s where they were playing with new digital technology.
Everyone simply did their best with the tools they had at hand. Amazing considering the NS-10 near fields and JBL studio mains everyone suffered with. Nothing even close to high fidelity, or even flat, was in common use in most recording environments. Consider too, most systems being driven with Crown (Amcron in Canada) DC300's or (gasp) Bryston 4B's. Sound quality in a recording studio didn't stand a chance!
Those were the bad old days for sure!
-Chris
-Chris
I just think you are assuming too much about what Paul Klipsch was all about and his decisions. K-Horns have their pluses and minuses and you either really love them, or hate them. They aren't my taste either, but Klipsch did make some crossover improvements later on that addressed some issues. They even learned to raise some x-over frequencies to reduce the honk a little more. Their latest horns that I have heard (Tractrix ? sp) are far more natural, but still not my cuppa. I could listen to a system that used these now.
The length of path issues are something not easily overcome in horn designs, and it's a specific problem that is well known. Since you're covering this material, why not cover it as the issue that it is, rather than as an example of a design "mistake" that it really isn't? Horn designs can be examined without the idea that these issues are an optional design problem - which they are not. It's the nature of the beast and to gain some performance, trade-offs are made in other areas. It's nothing more than that and Paul Klipsch was not a bad designer. Nor was he blind to these design issues. What would you expect him to do?
Analog tape vs digital recording is another example of design trade-offs. Recording in the digital realm has a completely different mind set from analog tape recording. For example, when you run out of 1's and 0's, nasty things happen in digital. Analog tape handles overloaded signals in a more benign manner. Today's recording engineers have learned, but in the early days, some truly evil things happened with digital gear. Low level signals had their own devils, like reduced resolution in bits instead of simply increased background noise. Back then, a Studer equipped with Dolby SR had a wider signal to noise ratio than the 16 bit A/Ds and D/As did. There was no contest between digital in the 80s and 90s to a good analog setup.
It took several tricks to allow digital technology to compete successfully with good analog systems. Back then, it was just simply ugly. I spent some time at "The Metalworks" recording studio in the 90s where they were playing with new digital technology.
Everyone simply did their best with the tools they had at hand. Amazing considering the NS-10 near fields and JBL studio mains everyone suffered with. Nothing even close to high fidelity, or even flat, was in common use in most recording environments. Consider too, most systems being driven with Crown (Amcron in Canada) DC300's or (gasp) Bryston 4B's. Sound quality in a recording studio didn't stand a chance!
Those were the bad old days for sure!
-Chris
-Chris
Anatech, I first met Paul Klipsch in 1965 over lunch, I sold Klipsch, owned Klipsch, and drank whiskey with Paul Klipsch (his bottle) with Richard Heyser on several occasions at AES. I worked with John Meyer, (remember him?) in sorting out the +'s and -'s of the K-horn, both corner, and LaScala. I also worked with John Meyer in HORN loudspeaker design for 1.5 years at our lab in Switzerland. This is AFTER working for the GD on their sound system for several years and Kelsey and Morris in London for a summer.
Please, just let me teach so interested people something they might not have learned.
Please, just let me teach so interested people something they might not have learned.
Moving on, if I am still allowed:
The next real problem, besides path length, in horns, is the generation of horn throat distortion. This distortion is primarily second harmonic, and can be computed by the horn cut-off frequency, test frequency, and initial horn throat diameter, usually 1/2", 1", 2" maybe more, today in some examples, and input power to the horn. The smaller the throat, the larger the distortion. However, if you use a multiple horn, like the K-horn, and don't try for more than living-room filling sound, horn throat distortion is fairly low, probably better than an equivalent direct radiator doing the same output volume.
However, for PA, this causes a BIG problem, especially today. The only REAL, or obvious, improvement in horns, today, is the power handling capacity of their drivers. In the old days, when many of the horns were designed, a 10W or 20W driver might be used. Today, maybe 100W or 200W drivers might be attached to the SAME basic horn design. See the problem? more later
The next real problem, besides path length, in horns, is the generation of horn throat distortion. This distortion is primarily second harmonic, and can be computed by the horn cut-off frequency, test frequency, and initial horn throat diameter, usually 1/2", 1", 2" maybe more, today in some examples, and input power to the horn. The smaller the throat, the larger the distortion. However, if you use a multiple horn, like the K-horn, and don't try for more than living-room filling sound, horn throat distortion is fairly low, probably better than an equivalent direct radiator doing the same output volume.
However, for PA, this causes a BIG problem, especially today. The only REAL, or obvious, improvement in horns, today, is the power handling capacity of their drivers. In the old days, when many of the horns were designed, a 10W or 20W driver might be used. Today, maybe 100W or 200W drivers might be attached to the SAME basic horn design. See the problem? more later
What's wrong with furrin cars?
An externally hosted image should be here but it was not working when we last tested it.
se
I'm going to have to admit that Paul Klipsch, unfortunately, never had the opportunity to meet me.
Best Regards,
TerryO
- Status
- Not open for further replies.