Many here seem to be up on your digital, but it would seem that you are behind on analog distortion mechanisms.
40 years ago, I was working on the TIM standard test (sine-square) which used multi-tone IM to achieve a consistent test for TIM. It was convenient to use a bandwidth limited square wave as the rate-of-change medium, but it was known that most of the IM products will fall back on the test frequencies themselves, and measurement from the original test waveform and the output waveform from a piece of audio equipment was almost impossible, so a deliberately chosen frequency was added that was not related to the square wave or its harmonics. At this time 15KHz was chosen and the fundamental square wave frequency was chosen to be 3.18KHz, so that the IM artifacts would stand out between the test tones. I also found that you could use a bandwidth limited triangle wave instead of a square wave, (just the turn of a switch on the square wave generator) and get a close approximation of SMPTE IM for comparison, thereby separating TIM from normally measured IM distortion. This showed the problem with slower amps more clearly. For the record however, while multi-tones are sometimes used, you only need 3 tones to get pretty much all you need, (2 tones will not give you everything).
40 years ago, I was working on the TIM standard test (sine-square) which used multi-tone IM to achieve a consistent test for TIM. It was convenient to use a bandwidth limited square wave as the rate-of-change medium, but it was known that most of the IM products will fall back on the test frequencies themselves, and measurement from the original test waveform and the output waveform from a piece of audio equipment was almost impossible, so a deliberately chosen frequency was added that was not related to the square wave or its harmonics. At this time 15KHz was chosen and the fundamental square wave frequency was chosen to be 3.18KHz, so that the IM artifacts would stand out between the test tones. I also found that you could use a bandwidth limited triangle wave instead of a square wave, (just the turn of a switch on the square wave generator) and get a close approximation of SMPTE IM for comparison, thereby separating TIM from normally measured IM distortion. This showed the problem with slower amps more clearly. For the record however, while multi-tones are sometimes used, you only need 3 tones to get pretty much all you need, (2 tones will not give you everything).
There is no magic to directivity effects except in the design of the waveguides, perhaps. 🙂
Either you sit in the near field and up close to the monitor to reduce room influence on the sound or you can effectively extend the near field further by narrowing the speaker dispersion.
The low freqs are not made directional but then they almost cannot be made so. At least not with this direct radiator design. However those freqs are more in control of room modes and lowered detect-ability of direction in bass... so there is nlittle need to try to make freqs below the x-over as directional as the mid-higher freqs.
THx-RNMarsh
Last edited:
3 tone IM dates back to a Wireless World article. There was a construction article - once upon a time -- for a 3 tone T&M. Such 3 tone IM is discussed here also -- http://www.cordellaudio.com/papers/multitone_test.pdf
Or... https://books.google.com/books?id=-...#v=onepage&q=3 tone IM test for audio&f=false
and - https://books.google.com/books?id=-...#v=onepage&q=3 tone IM test for audio&f=false
THx-RNMarsh
Last edited:
The cone shape and construction of the M2's 15 inch woofer leads to a controlled directivity at the crossover frequency not much wider ( maybe even a little less) than the HF horn's and then there is a gradual widening of the directivity as the system goes down in frequency once the crossover to the woofer has occurred. The M2 HF horn's directivity is pretty constant over it's operating band.
Attachments
Last edited:
Jan,
The problem remains that if you use those frequencies at 44,100 and 96,000 sampling rate they can not be exact bins for both. The list of frequencies and the write-ups have insufficient information about the rest of the details to get unambiguous results in all cases.
At 96k the frequencies are integer inverse multiples of 1.041666e-05 and at 44.1k 2.2675737e-05. None of the frequencies in that Excel spread sheet are exact bins in either. Let me see if I can find the ISO spec and sort this out. I still have difficulty believing all possible intermods of any order do not fall into the same bins eventually, I suspect there are some simplifying assumptions underlying this.
Last edited:
So some comments 40 years ago define ALL BBC loudspeakers, despite Harbeth later in life denying the dip existed other than to get around cone resonance? Given that the design reports for all the classic BBC monitors are available online you can even read them.
E.G. http://downloads.bbc.co.uk/rd/pubs/reports/1979-22.pdf which is the LS5/8 design and measurements of the R&D prototype that was Harwood's last design before he retired. It's flat.
Likewise The design of the prototype LS5/9 studio monitoring loudspeaker. - BBC R&D
Any response anomalies are due to companies such as Rogers not building to spec, certainly the 'famous dip' was not designed in by BBC R&D and a recessed midrange is not a sound favoured by UK audiophiles.
What interests me is how clever do we have to be to get IM from the DAC and ADC out of the equation. With single tone HD measurements there are hundreds of pages of angst on here about source phase noise and deep notches. In the multi-tone case this would seem to be even more of a problem, at least having not run the numbers...
I can't see why it shouldn't be equation driven, based on FFT length and sampling rate, since that's all you need to know bin width. (This is all to agree with you)
Just a note about the dispersion angle of the JBL M2 horn is. As stated by JBL it is 120 degrees and not 60 degrees, it is 60 off axis so 120 total. Reason that they recommend the speaker is to be toed in. So to say the room is not affecting the high frequency response with those horns if they are not toed in considerably and that meaning pointed at the listening position throws that notion out the window.
You win on that one Jan, the room is a significant factor with those JBL's. The other reality is that even with that so called advanced shape change they still left the horns as a very sharply truncated flare with little to no radius on the end of the horn so I would suspect there is some serious edge diffraction going on there. This is just a seriously dsp tuned two way horn/ dynamic driver system with lots of electronic correction. It would be interesting to hear them without the corrections. How many of us grew up with similar horn and cone systems in the 60's and 70's, I still have my Altec's sitting in the living room.
You win on that one Jan, the room is a significant factor with those JBL's. The other reality is that even with that so called advanced shape change they still left the horns as a very sharply truncated flare with little to no radius on the end of the horn so I would suspect there is some serious edge diffraction going on there. This is just a seriously dsp tuned two way horn/ dynamic driver system with lots of electronic correction. It would be interesting to hear them without the corrections. How many of us grew up with similar horn and cone systems in the 60's and 70's, I still have my Altec's sitting in the living room.
Hope you are feeling better now!
I think to those of us brought up on the 'horns are shouty, bad and narrow bandwidth' the work of Toole et al at JBL and Earl Geddes who have come up with fundamentally the same compromise and produced 2 ways of what are reported as superlative performance is an eye opener. CD seems to be a step in the right direction in the google of ways you can make a speaker.
I am also (as many others are) following those who are extending this along the lines of the Danley tapped horns with what appear to be fine results. They may be onto something! Certainly interested in your views on this.
Yes but for a given set of frequencies there is one combo (or integer multiples of it) that works 44.1k and 96k can't both work on the same frequencies except at a ridiculously huge common multiple IIRC. I had this problem trying to make some multi-tone files for folks, they couldn't get their heads around "these work only on 65536 point FFT's", period.
Last edited:
Good horns have existed for decades. Good implementation, maybe not so common. 😉
And goals have perhaps shifted.
One thing I did, at Charles's suggestion, is to walk around the room to see how well the M2 holds up with off axis imaging. It does really well. Way better than anything I have heard before in the 2ch world.
You guys are trying to use specs to tear down the M2 performance. Speaker envy maybe? Very childish.
You guys are trying to use specs to tear down the M2 performance. Speaker envy maybe? Very childish.
Bill,
One of the major differences between what we are seeing today and in the past is in the compression drivers themselves. Even without the dual diaphragm design many newer compression drivers are much shorter than the older Alnico designed required in the now distant past. So the throat length is much shorter and can have a much greater initial angle than you could possible fit into the older long throat designs, the major cause of that horn shout in the first place. That goes along with Earl's concept of matching the exit angle of the compression driver to the horn, there isn't the mismatch that was so common back in the day. I've only been making my horns that way since the mid 80's to match the pancake type drivers so I guess I must be late to the game!
This is one of the reasons I think that all these guys chasing large older style 2" exit compression drivers are just missing the boat on what is really going on. You just can't overcome that horn shout with a combination of large slow opening compression driver/horn combinations, not going to happen.
One of the major differences between what we are seeing today and in the past is in the compression drivers themselves. Even without the dual diaphragm design many newer compression drivers are much shorter than the older Alnico designed required in the now distant past. So the throat length is much shorter and can have a much greater initial angle than you could possible fit into the older long throat designs, the major cause of that horn shout in the first place. That goes along with Earl's concept of matching the exit angle of the compression driver to the horn, there isn't the mismatch that was so common back in the day. I've only been making my horns that way since the mid 80's to match the pancake type drivers so I guess I must be late to the game!
This is one of the reasons I think that all these guys chasing large older style 2" exit compression drivers are just missing the boat on what is really going on. You just can't overcome that horn shout with a combination of large slow opening compression driver/horn combinations, not going to happen.
I'm not reading that at all. It's simply a wide dispersion (ergo room-sensitive) speaker design. That's all.
morinix,
I think the point I was trying to make is that the M2 doesn't really qualify as a narrow dispersion horn design just the opposite is what I would say, 120 degrees is actually just the opposite and I would call it a very wide dispersion design. The point being that as with any wide dispersion speaker system you just can't say the room is not a factor in what is going on. I am not trying to knock the design in any way, just bring some reason to the conversation. If we had had the possibilities of what dsp could do back in the 70's I think there would still be many more horn based systems around today.
I think the point I was trying to make is that the M2 doesn't really qualify as a narrow dispersion horn design just the opposite is what I would say, 120 degrees is actually just the opposite and I would call it a very wide dispersion design. The point being that as with any wide dispersion speaker system you just can't say the room is not a factor in what is going on. I am not trying to knock the design in any way, just bring some reason to the conversation. If we had had the possibilities of what dsp could do back in the 70's I think there would still be many more horn based systems around today.
I use 4 of the 1970's JBL yard wide 2395 lens with 2441 Alnicos in our 4 way shop PA - rather like these in black plywood with twin 18s in each
4 of them do 3000sq feet quite well
Brown Sabbath anyone ? https://youtu.be/nTd5ghvMkvI
4 of them do 3000sq feet quite well
Brown Sabbath anyone ? https://youtu.be/nTd5ghvMkvI
Attachments
Last edited:
- Status
- Not open for further replies.