I have a question about horn on compression driver. I managed to get best tone so far with: 1. Sub is 240 cm 12" tapped horn. 2. 15" long throw car speakers as mid basses. 3. Dayton PC83 as mids 4. pair of 44T30H8 cheap compression drivers stolen from Behringer passive PA speakers. My amps are Behringer NU4 for sub and midbasses. Samson Servo 200 for mids and finally ebay mono blocks for tweeters. I cut the system at 40hz(sub) , 40-235hz(midbasses) and 7200hz (mid/tweet) all done digitally in PC DAW with 42db depths. I match mids by ear which turns out to be very obvious. Different sources need little bit different volumes for mids in ranging in around 6db range. My question is what is the meaning of possible horn on compression tweeters. I feel the tone is even better without horns. Does it have any meaning when it is cut so high?
Considering the circumstances I'm inclined to suggest you go with what sounds best to you. At 7kHz much of the waveform will have developed before the exit of the driver. This is presuming you're not interested in measuring and analysing to work out what is happening.. It's plausible that no horn could give a reasonabe result, and could be better than some random horn.
Thank you for your kind answers! The goal is to be able to push the volume up to live level without loosing definition or hurting ears. Not many sources are suitable to find that edge. I have found couple though. In fact I'm very interested in also measuring what is happening. I have a measuring mic at least to start with. What I really would aim for is to be able to find meaningful simple constructions and also hear some differences depending on construction and deduce the same thing also with measurement method. So it is tricky because meaningful method is ropedancing. I have done driver juggling quite a long time and biggest advance was to start using those little full ranges as mids. It allows very much higher volumes as 6.5 incher (due to now missing 4200hz cone breakup) used before. This can be measured easily. I have tried many things and it seems to be so that less is more as long as the bass end is having maxiumum capacity to rattle the windows without driver hitting the rims and as a bonus same midbass can be used up to 235hz enabling 4 incher to step in. 4 inchers' cone breakup is shifted to over 10k so that is the reason for great advance in volume without any EQing. I cannot afford coaxial Genelec so I try to get there with the small full ranges. Then only one PC83 is needed as mid for a side and only one compression tweeter. I have tried Seas 27TBFC/G and Dayton AMTpro4 but the Behringer compression driver just play symbals' upper part cleanly very loud. Before using PC83s I was not able to pinpoint the problem lurking in the forementioned tweeters.
Horns I have are Dayton Audio H10RW 10" Round Waveguides. So how would I go about measuring that horn thing? I guess first thing to find out would be able to tell diffraction and aberations in frequency response? Then as second step one should be also to be able to hear the problem. Or somehow produce the same effect as much stronger and then gradually be able to make it come in and out and learn to listen it.
Horns I have are Dayton Audio H10RW 10" Round Waveguides. So how would I go about measuring that horn thing? I guess first thing to find out would be able to tell diffraction and aberations in frequency response? Then as second step one should be also to be able to hear the problem. Or somehow produce the same effect as much stronger and then gradually be able to make it come in and out and learn to listen it.
Cone breakup is not about the bump. Cone breakup happens with fullrange drivers too.
If you want loud then you should reduce diffraction.. it sets a limit to how loud you can go. For this you should find a good waveguide. Remember that the fullranger beams at higher frequencies and that's what you want to continue.
About cone breakup. I continued...: "4 inchers' cone breakup is shifted to over 10k so that is the reason for great advance in volume without any EQing." That was just my repeatable observation. I take frequency response at phoneme range (1300hz-4300hz) matters because brain is there most prone to autocorrect things. The shift of the cone breakup can be seen in measurement and manufacturer's published FR. I just found these PC83 full ranges do not need any low cut. They behave so extremely well that they perfectly cut out themselves. The cone breakup is meaningless. Then I need to add just a little compression drivers to mix (starting at 7000hz or so) and picture keeps very detailed. Currently I can play so loud that some things in room are just starting to resonate. But it is only at listening position. I'm basically seriously listening only at listening position.
What frequency range is most vulnerable for diffraction? So I take you mean I should be able to keep the directivity same with full range's highs and tweeter's low at the frequency where the tweeter crossover kicks in? 7000hz is equal to 4.9cm. What kind of consideration and method you suggest to start with to get the current picture?
What frequency range is most vulnerable for diffraction? So I take you mean I should be able to keep the directivity same with full range's highs and tweeter's low at the frequency where the tweeter crossover kicks in? 7000hz is equal to 4.9cm. What kind of consideration and method you suggest to start with to get the current picture?
From a DAW track I can send any kind of sweep and noise and noise ranges or impulse to amp->driver with or without crossover turned on in DAW track. Then I'm able to mike it back to another DAW track and then redirect the result to a FFT analyzer. Maybe the recording distance should be quite short because it must be done in a room full of objects. Is there some unwanted effect if the miking distance is very short like 10cm?
See, the usual way to a good loudspeaker is to make a construction that has a defined, reasonable frequency response AND dispersion, in a free space. Manufacturers, speaker designers and scientists even use an-echoic rooms for this task, which have no reflections and resonance. They only use chassis best for a special task, carefully matched to each other. Then they design sense full things around this speaker, that enable the consumer to match the well dispersing, linear speaker to his individual room. If the room is good for listening, there will be very few things to adjust, if the room is bad, even the best sophisticated sound matching magic will fail.
Now your case. You start to build a speaker from components not made to match each other, like PA, car audio and HIFI chassis. You try to do this in a room not suited for listening, too small and full of things blocking and reflecting the sound randomly.
The result is predicable and will be quite disappointing. Not measuring does not hurt in such cases, as it is impossible to make any useful measurements anyway. Garbage in = garbage out. Non existing cone break up is your last problem.
So maybe return the car woofer to the car it came from and think about your situation, then sell all the mixed stuff you have and start again.
Maybe with something someone with a clue in loudspeaker design, like Behringer, has done. You may also realize, that no PA speaker is made to perform in a small, cramped room. There are all kinds of proven HIFI speaker designs around, finished and in Kit from, that suit such a situation. If this is not loud enough for you, get a head phone and a body shaker.
A wise guy once said: You are not going to bring your home HIFI speakers to a football arena when a rock band plays, so why should you take a rock bands PA system to your living room. Cows are cows and goats are goats.
Then, a universal tip, measure while you design and before you run into problems. A failed design will not get better when, as a last option, you start measure the catastrophe you have just build.
Now your case. You start to build a speaker from components not made to match each other, like PA, car audio and HIFI chassis. You try to do this in a room not suited for listening, too small and full of things blocking and reflecting the sound randomly.
The result is predicable and will be quite disappointing. Not measuring does not hurt in such cases, as it is impossible to make any useful measurements anyway. Garbage in = garbage out. Non existing cone break up is your last problem.
So maybe return the car woofer to the car it came from and think about your situation, then sell all the mixed stuff you have and start again.
Maybe with something someone with a clue in loudspeaker design, like Behringer, has done. You may also realize, that no PA speaker is made to perform in a small, cramped room. There are all kinds of proven HIFI speaker designs around, finished and in Kit from, that suit such a situation. If this is not loud enough for you, get a head phone and a body shaker.
A wise guy once said: You are not going to bring your home HIFI speakers to a football arena when a rock band plays, so why should you take a rock bands PA system to your living room. Cows are cows and goats are goats.
Then, a universal tip, measure while you design and before you run into problems. A failed design will not get better when, as a last option, you start measure the catastrophe you have just build.
If you are interested in measuring what is going on, you will need seriously good setup. Just consider that at your crossover frequency, 7.2kHz, second harmonic is 14.4kHz, third is 21.6kHz...and so on. How are you going to measure 5th or 7th? Especially those are offensive sounding. Is your microphone calibrated till 48kHz? Even capable? How good is soundcard?
Without good measurements and data, all you wrote so far Valeri is just mumbo jumbo.
Without good measurements and data, all you wrote so far Valeri is just mumbo jumbo.
It’s easy to get a horn tweeter level too high. I’ve crossed horn tweeters at ~7K and when the level was right the tweeters were barely noticeable.
Would a 6” waveguide be more appropriate?
Would a 6” waveguide be more appropriate?
My "design" is best for me so far and it makes sense to me. I really don't enjoy cone breakup in phoneme range 1300-4300hz (indicated in Fletcher-Munson curve) because it exites my tinnitus. No crossover or construction can remove cone breakup of larger mids. I don't need to measure it because Zaph has measured lot's of drivers and his measurements show that there is no such mid woofer driver that had not that cone breakup. With my realtime crossovers I can easily verify that thing to be listenable. The cone breakup will always be present also in measured noise of 6.5 inch woofer however the crossover has been set up. My speaker management system allows me to individually match each drivers volume for the line level signal from PC screen so I don't need to rebuild the construction. That would be total madness(in my case). This leads me to the method were physical world iteration is possible. What comes to listening volumes my observation is that every listening volume needs different implementation due to Fletcher-Munson curve. This is my most advanced goal. It is possible with forementioned DAW based speaker management system.
It looks like you are interested in distortion components.
about those 5th and 7th of 7000hz. I cannot hear things that high you mentioned. Usually human hearing can go up to 20k but most people over 50y can hear only up to 12k. Bats can hear those 5th and 7th of 7000 you mentioned. But being limited to 12k is no problem. Older people can enjoy high tones too because even the most beautiful cymbals do enough of their brightest musical thing at just under 10000hz. I prefer to keep it simple meaning focusing on generic human hearing range. Zaph has made great measurements in his site: http://www.zaphaudio.com/6.5test/compare.html In addition to frequency response there is tens of those distortion partials(harmonics) revealed and how to read them means to understand that 5th harmonic of 3000hz is 15000hz which has no more meaningful musical content. On the other hand let's say the fundamental of some string instrument (or woofer cone) is 100hz. Then you got 100 of them up to 10k all being meaningful. That is crazy (and great) amount of meaningfull harmonics.
Could you imagine some method to be able to exaggerate the baffle phenomena so that it can be listened as isolated?
This is exactly what has happened to me when I started to cut them so high. I only use the high part of the compression driver's response. It can be measured to be flat over 7000hz up 10k. Not that it means anything.
To clarify my situation I attach the shot about how I match the volumes of the drivers in the digtal domain before the signal goes to a crossover channel.
DBs have no meaning because the amps are different.
I just realized that how ever bad your room might be you still can make improvements in that particular environment. It is full apples to apples comparison in your own room. BTW: If you consider this Eric Kennedy drum solo being hifi recording then we are on the same page:
So ability to be able to taylor the nearfield hifi system for a particular setup is very good added value. Such a thought is very controversial and heretic of course. I see if somebody feels offended if they have not such setup and possibilities. But what is the perspective of a manufacturer? It can measure in unechoic rooms and free fields and make product logical in that way. Then many times the same product is sounding not so phenomenal in a customer's room. Every room is different. My approach to make hifi possible for poor(or not rich) people is to allow myself tune all those basic aspects which are normally kept to manufacturer like matching driver sensitivity and crossover points and steepness. Little by little I have tumbled in the way to get some serious hifi out of my poor man's system. Much of it happened because there is no volume limit in my house. So 2012 I started to experiment with tapped horn. It plays the first octave of spectrum. It was very hard to accept that the subwoofer and bass needs a lot of power and xmax and mids and tweets do not need it. Then I noticed that for midbass I cannot find budget 12inch speaker but some 200$ 15inchers are having enough xmax. Checking out tens of published measurements also suggest 15 incher is better midbass overall. One idea which big 15inch woofer enables is the ability to set the midbass/mids xo frequency so that mids themselves do not need to move much thus keeping the distortion of mids minimal. If you want oldschool 70's tone make it 320hz and for modern hifi it will be 238hz because at 320 15 inchers start to bring their cone rumble in which is nice effect for some sources. Mids being mini cannot be expected to move much. In a first place I tried full ranges as mids because tweeters could not reach low enough without being sordino like. Then only after those full ranges arrived I found out that cone breakup is not for me. The whole system is having super pleasant nearfield sweet spot. The tone works to other room too though.
Now I have made new effort and this is the compression driver screwed on flat baffle above the full range. Only topmost full range is connected. The combination still sounds really good and there is a sound stage. Mids and tweeters are stitched currently at 7350/42db. I wonder what might be scientific differences between a flat baffle and proper waveguide. Could I hear it as an average but very interested listener?
I would like to try Danley Synergy Horn type construction so the compression driver would come closer to mids.
But as three sides instead of four.
What could be expected from that?