fscarpa58 said:
Hmmmmm..... so it is, it should be n^2/4,
leaving 2nd unchanged and weighting the rest.
Olsons work suggested a linear weighting, i.e. x n/2.
The square rating comes from Shorters work, i.e. n^2/4.
(in 1950 ! he seems to have lost the measuring debate)
www.zainea.com/liviu/Comparison_of_Non-linear_Distortion_Measurement_Methods.pdf
See references at the end of the article for Shorters paper.
sreten.
lumanauw said:
All musical instruments have harmonics. Different instruments
have different spectra which is one of the main factors why
different types of instruments sound different. I am not an expert
on this, but if I remember correctly brass has a lot of rather strong
high-order harmonics which gives the somewhat harsh sound
typical of brass, strings I think have predominantely lower-order
harmonics, while the organ has a very heavy 2nd order harmonic.
instrument experts, please corerct me if I am wrong. Then we
have the human voice which is a much more complex instrument.
So, in a sense there are a lot of harmonics which we perceive as
pleasant, or at least desired, when they occur in the right context.
So, while we may not like an amplifier to produce high-order
harmonics, we do want't the brass section of the orchestra to
retain its harsh spectrum so we recognize it as brass. We don't
want the amplifier to add a lot of high-order harmonics to the
violins since they might end up sounding very harsh.
Re: Re: Re: Re: Re: What harmonic does to sound?
Yes...
I am guessing you got your frequency from a real instrument or similar. A "perfect fifth" is exactly 3/2's the frequency. So by going up a fifth, you reach a consonant note, using induction you should see that a fifth above this note should be consonant with it. Indeed this is so. As you continue this process, after 12 new notes you arrive at a note that is almost the same as the fundamental you started with but 7 octaves higher. If all 12 of these new notes are transposed down by octaves to the original interval you'll have the 12 notes of the modern scale. However, the math isn't perfect because the exponentials aren't the same. In one path you're multiplying by 3/2, in the other you're multiplying by 2. So what used to happen is that all instruments were tuned to a certain key (just temperment) which was derived from using this process from a given fundamental note. This disallowed creation of pieces which changed keys. So musicians created equal temperment which distributed the difference between the 12th fifth and the 7th octave throughout the entire range. Equal temperment allows an instrument to be played in any key without retuning.
The interval between two consecutive notes is divided into 100 cents, basically percentages. Equal tempered fifths are tuned off by about 2 cents but major 3rds are off by 14 cents. 5 cents is the limit of human perceptability.
Christer said:
Yes...
Christer said:
I am guessing you got your frequency from a real instrument or similar. A "perfect fifth" is exactly 3/2's the frequency. So by going up a fifth, you reach a consonant note, using induction you should see that a fifth above this note should be consonant with it. Indeed this is so. As you continue this process, after 12 new notes you arrive at a note that is almost the same as the fundamental you started with but 7 octaves higher. If all 12 of these new notes are transposed down by octaves to the original interval you'll have the 12 notes of the modern scale. However, the math isn't perfect because the exponentials aren't the same. In one path you're multiplying by 3/2, in the other you're multiplying by 2. So what used to happen is that all instruments were tuned to a certain key (just temperment) which was derived from using this process from a given fundamental note. This disallowed creation of pieces which changed keys. So musicians created equal temperment which distributed the difference between the 12th fifth and the 7th octave throughout the entire range. Equal temperment allows an instrument to be played in any key without retuning.
The interval between two consecutive notes is divided into 100 cents, basically percentages. Equal tempered fifths are tuned off by about 2 cents but major 3rds are off by 14 cents. 5 cents is the limit of human perceptability.
Re: Re: Re: Re: Re: Re: What harmonic does to sound?
No, I just computed it, but I just found out I could just have looked
it up in one of my books, which apparently has a table. However,
I did not assume a perfect 5th, but an equal tempered one, since
that seems more relevant for real music, I think. Maybe that is
a wrong assumption.
I don't know where you get the figure from that an equal
tempered 5th is off by 2%, though. I thought I had misunderstood
equal temperement, which is why I just checked with my book,
which says that an equal tempered 5th is a ratio of 1.498, which
is the same as I calculated (well, I said 0.1% off but recalculating
it I see now I rounded it off in the wrong direction, so 0.2% is
more correct).
azira said:
No, I just computed it, but I just found out I could just have looked
it up in one of my books, which apparently has a table. However,
I did not assume a perfect 5th, but an equal tempered one, since
that seems more relevant for real music, I think. Maybe that is
a wrong assumption.
I don't know where you get the figure from that an equal
tempered 5th is off by 2%, though. I thought I had misunderstood
equal temperement, which is why I just checked with my book,
which says that an equal tempered 5th is a ratio of 1.498, which
is the same as I calculated (well, I said 0.1% off but recalculating
it I see now I rounded it off in the wrong direction, so 0.2% is
more correct).
" I certainly likes the sound of Santana's guitar and it must be distorted purposely."
Fine, but not germain to distortion in reproducing a performance. Or to put it another way adding that effect after the fact to a Segovia, Almeida, Assad brothers or John Williams guitar recording would likely sound more than a little unpleasant!
Fine, but not germain to distortion in reproducing a performance. Or to put it another way adding that effect after the fact to a Segovia, Almeida, Assad brothers or John Williams guitar recording would likely sound more than a little unpleasant!
Interesting passing mention of the 6DJ8 'sound', which I'm not very fond of btw, in your link. Btw, there exist small dual triodes with nearly as high transconductance that don't have the 6dj8 'sound' which would seem to give them the sonic advantage. The only one I have close experience with is the 6bk7 which I think sounds greatly superior to the 6dj8, but the little talked about 12av7, among others, would seem to have some potential in 6dj8 type audio applications.
It's hardly a new revelation that different relative distributions of harmonics at audible levels should result in different percieved sonic characters. Doing a listening test in a high-end shop full of different model tube amps is like tasting ice cream flavors - will the variety never end? However, insert one or two leading zeroes in the THD figure and will it really matter any more?
In both the DIY and store-bought arena extremely low THD figure can be had, yet the assertion that there are audible differences among them persists. Is the perception valid? Is it valid with the proviso that clipping (or near clipping) is avoided? If differences are appearent only at or near clipping, what characteristics lead to which sonic "flavors"?
Relating measurement to perception is not done enough. The tests described above are a fascinating read and the result of much diligent effort, it kind of covered ground already trodden. Results for specific circuit variations is useful I'm not something to surprise me. I wish someone with such good equipment would shock and amaze me with some measuremnts that relate questions I possed which so far as I know have not been dug into in much detail.
In both the DIY and store-bought arena extremely low THD figure can be had, yet the assertion that there are audible differences among them persists. Is the perception valid? Is it valid with the proviso that clipping (or near clipping) is avoided? If differences are appearent only at or near clipping, what characteristics lead to which sonic "flavors"?
Relating measurement to perception is not done enough. The tests described above are a fascinating read and the result of much diligent effort, it kind of covered ground already trodden. Results for specific circuit variations is useful I'm not something to surprise me. I wish someone with such good equipment would shock and amaze me with some measuremnts that relate questions I possed which so far as I know have not been dug into in much detail.
It certainly can and often does. The assumption that it cannot is what misleads many. If the fiftieth odd harmonic of a 100db spl 25hz component of the sound is 80db down, it will definitely be perceptible, not to mention the slew of other attendant harmonic and intermodulation byproducts that will be produced along with it.
In the past, I've likened good tube sound to watching an event through a window that is optically clear but perhaps is slightly nonflat at the edges. With transistors, it may be flatter at the edges, but it also is more or less frosted and has micro inclusions of contaminating substances.
Oh I 2nd, 3rd, 4th and nth this comment! No kidding. The number of publications and books and so on that go into detail of what can be seen on the display of measuring equipment are bountiful. But which of these dares relate them to how something actually sounds? How many have the courage to do this? How many have the courage to subject their cossetted invention to 3rd party comparison in their local hifi store and then publish the results?
I was lucky in a way. I started out building and listening to circuits before I learned about how to use measurement equipment. So I am predisposed to blame the measurements if something doesn't sound right. And then I adjust the measurements and my understanding of what the circuit is doing to correlate with the sound and not the other way around.
Hmm,
In post 9 of this thread I already posted a link to a 25 years old article by Henning Moller of Bruel & Kjaer about correlating objective measurements and subjective perception. It seems not many took notice of it and also it seems there is not much progress of it during the last 25 years.
Cheers
In post 9 of this thread I already posted a link to a 25 years old article by Henning Moller of Bruel & Kjaer about correlating objective measurements and subjective perception. It seems not many took notice of it and also it seems there is not much progress of it during the last 25 years.
Cheers
"If the fiftieth odd harmonic of a 100db spl 25hz component of the sound is 80db down, it will definitely be perceptible"
Hmmmm.
I've got a table from OSHA in front of me. According to this table 100db spl is like being on a subway platform as the train goes by while 20db is like a quiet conversation in a speech studio.
I'm trying to remeber what is was like last time I stood on the platform while the train (S-bahn actually)went byand imagine what other sounds I could hear. Plus if the 80db down figure is to correspond to THD, individual harmonics would be even less. In any case, if I take "subjectivist" advice and trust my own ears and experience I've got to conclusde that while have quite an interesting argument . . .
Sorry, no cigar.
Hmmmm.
I've got a table from OSHA in front of me. According to this table 100db spl is like being on a subway platform as the train goes by while 20db is like a quiet conversation in a speech studio. I'm trying to remeber what is was like last time I stood on the platform while the train (S-bahn actually)went byand imagine what other sounds I could hear. Plus if the 80db down figure is to correspond to THD, individual harmonics would be even less. In any case, if I take "subjectivist" advice and trust my own ears and experience I've got to conclusde that while have quite an interesting argument . . .
Sorry, no cigar.
sam9 said:"If the fiftieth odd harmonic of a 100db spl 25hz component of the sound is 80db down, it will definitely be perceptible"
Hmmmm.
I'm trying to remeber what is was like last time I stood on the platform while the train (S-bahn actually)went byand imagine what other sounds I could hear. Plus if the 80db down figure is to correspond to THD, individual harmonics would be even less. In any case, if I take "subjectivist" advice and trust my own ears and experience I've got to conclusde that while have quite an interesting argument . . .
Sorry, no cigar.
You are misunderstanding the implications of having a 15th
harmonic 80db down. By definition this cannot be the only
harmonic distortion, it implies a series at least to the 15th,
and probably much further - 81st harmonics have been
identified in Crowhursts work.
And for any complex signal IM distortion is far more important,
though this is related to the level of harmonic distortion produced
when driven with test tones.
sreten.Posted Earlier:
"I want to hear the instruments and their harmonics as recorded. I am confident that they are perfectly fertile. So an amp that changes nothing about them is the one for me. I don't want any modification or extra harmonics thrown in, nor should it be necessary to do so."
I agreed. If I don't like the source of the music, I change it. The musician/recording artists produce the sound. I just want to re-produce the sound. Of course, re-production depends on more than just the power amp. CD/player or vinyl/turntable, etc., preamp, power amp, speakers, entertainment room acoustics, and noise polution to name some, have a combined effect.
A power amp with .001 percent THD will likely have a not so desirable transient response. Bandwidth of 20 Hz to 20 KHz is for sinewaves. Music is complex and can have some fast attack/fast rise time content. Sinewave bandwidth may need to be in the 100s of KHz or even 1 MHz. At least one high end audio manufacturer's company president says he can hear the difference when he switches between 1 MHz and 2 MHz bandwidth on his amps. The proof is in the hearing, but with good techniques using modern test equipment, one should be able to measure some quality differences between amps.
"I want to hear the instruments and their harmonics as recorded. I am confident that they are perfectly fertile. So an amp that changes nothing about them is the one for me. I don't want any modification or extra harmonics thrown in, nor should it be necessary to do so."
I agreed. If I don't like the source of the music, I change it. The musician/recording artists produce the sound. I just want to re-produce the sound. Of course, re-production depends on more than just the power amp. CD/player or vinyl/turntable, etc., preamp, power amp, speakers, entertainment room acoustics, and noise polution to name some, have a combined effect.
A power amp with .001 percent THD will likely have a not so desirable transient response. Bandwidth of 20 Hz to 20 KHz is for sinewaves. Music is complex and can have some fast attack/fast rise time content. Sinewave bandwidth may need to be in the 100s of KHz or even 1 MHz. At least one high end audio manufacturer's company president says he can hear the difference when he switches between 1 MHz and 2 MHz bandwidth on his amps. The proof is in the hearing, but with good techniques using modern test equipment, one should be able to measure some quality differences between amps.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.