After the above house curve test part #### I stumbled over a post by Mitch to a paper from Floyd E. Toole, retired of JBL fame: http://www.aes.org/tmpFiles/elib/20151013/17839.pdf
It included this picture of preferred house curves in general:
Here's the curve I ended up with after my last session:
Left and right channels at listening position, 1/3rd octave smoothing (predictions)
Coincidence?
It included this picture of preferred house curves in general:
Here's the curve I ended up with after my last session:
Left and right channels at listening position, 1/3rd octave smoothing (predictions)
Coincidence?
Wesayso, I have also found this to type of curve to be the most "listenable" on a wide range of music.
- 6-10db bump from 80 Hz to 40 Hz
- flat through 1 kHz
- gently sloping downwards from 1 kHz to 20 kHz.
But this is measured at the listening position. Ideally, you should measure flat anechoic (no reflections), and the combination of your speaker directivity and room absorption should give you the modified curve above at the listening position. My experience says that deviating from flat on-axis to get this response at the listening position is pleasant, but always give the feeling that something is missing.
What is a real coincidence is that most well-designed 3-way speakers in conventional rooms will measure like this at the listening position (except bass boost). I say 3-way because that's about the right number of drivers to cover the frequency range with pistonic operation (decreasing size of drivers will lead to widening directivity).
- 6-10db bump from 80 Hz to 40 Hz
- flat through 1 kHz
- gently sloping downwards from 1 kHz to 20 kHz.
But this is measured at the listening position. Ideally, you should measure flat anechoic (no reflections), and the combination of your speaker directivity and room absorption should give you the modified curve above at the listening position. My experience says that deviating from flat on-axis to get this response at the listening position is pleasant, but always give the feeling that something is missing.
What is a real coincidence is that most well-designed 3-way speakers in conventional rooms will measure like this at the listening position (except bass boost). I say 3-way because that's about the right number of drivers to cover the frequency range with pistonic operation (decreasing size of drivers will lead to widening directivity).
Last edited:
ra7, I slowly worked my way to the above curve. I've tried a whole lot of different curves and sounds. The downward sloped Sean Olive curve was one of the first that stuck for longer than average. I usually run my FIR correction, chose the preferred curve at that time and over time do very minor cuts in JRivers graphic equaliser if needed. After an extended period I try to integrate the graphic EQ cuts into the next curve to evolve. That's how I ended up with what I called a gentle JBL curve over time.
I know the theory you speak about, but without access to an anechoic room or other possibilities to measure huge towers without including reflections (like outside) I worked backwards. And that basically led up to the same point in the end. I need the close by walls (to the speakers) to get the wanted response in the lows.
The one thing I did work on is getting the room response as equal as I could to the response from the speaker. At least as far as I could measure that.
If I measure in stereo, let REW gate from 0 to 40 ms, or 40 ms to 500 I keep the same shape in my graphs. Processing alone wouldn't have worked to get that. The damping panels helped me get that part right. All processing I do is relatively short windows, meaning I shape the sound of the speaker more than the sound of the room.
Green curve is REW average of left and right gated 0 to 40 ms
Blue curve is REW average of left and right gated with a 4 cycle window
Red curve is REW average of left and right gated 40 to 500 ms
When the ideal method isn't possible, use the next best thing I say
. Anechoic is out, close up measuring of arrays is pointless so the listening position was the best place to attack the problem. Especially after I averaged multiple measurements covering my entire couch/seating area and got basically the same curve as at the listening position as the average result. After that I had no more doubt this would/could work.
P.S. A very important difference between this case and multi-way speakers is that I didn't need to design crossovers to make this work. It is a row of small full range speakers all acting as one big "line". Very different from figuring out proper crossovers at the listening position. No strange off axis behaviour other than the (already known) off axis response of these small drivers.
Can't mention enough what a big part the room is playing in all this. On one side I need it (for my low end) and on the other side you want to get rid of it as a stamp on all your music. I only softened the effect of the room, by damping panels and the rest with DRC-FIR. I'm under strong advisement from my other half not to add any more damping. Got to work with what you've got... I'm slowly letting her get used to the idea that I'm adding more speakers and another amp. That didn't go all smooth: "Even more speakers? Isn't it good enough the way it is?" "But what if I can get it even better honey?"
I know the theory you speak about, but without access to an anechoic room or other possibilities to measure huge towers without including reflections (like outside) I worked backwards. And that basically led up to the same point in the end. I need the close by walls (to the speakers) to get the wanted response in the lows.
The one thing I did work on is getting the room response as equal as I could to the response from the speaker. At least as far as I could measure that.
If I measure in stereo, let REW gate from 0 to 40 ms, or 40 ms to 500 I keep the same shape in my graphs. Processing alone wouldn't have worked to get that. The damping panels helped me get that part right. All processing I do is relatively short windows, meaning I shape the sound of the speaker more than the sound of the room.
Green curve is REW average of left and right gated 0 to 40 ms
Blue curve is REW average of left and right gated with a 4 cycle window
Red curve is REW average of left and right gated 40 to 500 ms
When the ideal method isn't possible, use the next best thing I say
. Anechoic is out, close up measuring of arrays is pointless so the listening position was the best place to attack the problem. Especially after I averaged multiple measurements covering my entire couch/seating area and got basically the same curve as at the listening position as the average result. After that I had no more doubt this would/could work.P.S. A very important difference between this case and multi-way speakers is that I didn't need to design crossovers to make this work. It is a row of small full range speakers all acting as one big "line". Very different from figuring out proper crossovers at the listening position. No strange off axis behaviour other than the (already known) off axis response of these small drivers.
Can't mention enough what a big part the room is playing in all this. On one side I need it (for my low end) and on the other side you want to get rid of it as a stamp on all your music. I only softened the effect of the room, by damping panels and the rest with DRC-FIR. I'm under strong advisement from my other half not to add any more damping. Got to work with what you've got... I'm slowly letting her get used to the idea that I'm adding more speakers and another amp. That didn't go all smooth: "Even more speakers? Isn't it good enough the way it is?" "But what if I can get it even better honey?"
Last edited:
Getting closer...
Now am I confident enough this is going to work and paint the enclosures before I test it?
Nice looking Daggers you got there. Actually, the 4-sided Dagger was something I suggested to another member a while back who was loathe to build the spiral Nautaloss (basically a coiled up 4-sided Dagger).
It really is a nice coloration free enclsoure, just fill with stuffing of variable density - tightly stuffed at the apex and looser as you get closer to the driver and leave a few inches of free space around driver. Since you have some parallel walls, it may help to line the flat walls adjacent to driver basket with foam/felt to prevent back reflections (sideways).
Thanks for the tip on Infected Mushroom - cool music!
https://www.youtube.com/watch?v=EjKGFLbSFBw
Looking more closely that's quite the bump you have at low frequency... mine is about 3 dB but extends more downwards to ~25 Hz with 17 Hz -3 dB (or in other words at the level of 200 and up).
More bass than that seems to overpower the midrange, less than 3 dB and I start missing the shaking floor
at 85-88 dB listening levels. I've had the extension in the low end stop at 35 Hz for a while. It just doesn't give you the same dynamic thrill when listening.Just two more honey..... Beolab 90.
I don't think I ever told her what my ideas for the stools were that stand beside each tower...
I haven't updated this picture yet, but the stool you see on the left is now accompanied by the exact same one on the right side. The amp etc. has moved to a position below the TV.
I always figured if she accepts their place I can easily replace them with a pair of down firing subs with the same black cover. Up till now this has not been needed. It was a back-up plan
. I might eventually play with that though. No hurry as the arrays have plenty output down low. The only reason would be to increase the headroom at the first octave.Thanks for the inspiration
.The enclosures will have a layer of real wool felt on all sides, just like in my Line Array's and stuffing will be either fibre glass fill or Twaron's Angel Hair or perhaps a combination of those two (I've seen that to work excellent in the midrange). I agree about some free space behind the driver. I also put a layer of felt between the driver and the stuffing:
I don't think I've shared this picture before, a view just before I closed up the Arrays.
Last edited:
No, never saw this picture before. Thanks for sharing - you did a pretty careful job soldering and shrink tube wrapping each terminal quick connect. That's a 100 jobs ! I would have just crimped and called it a day! in the Dagger I wouldn't put felt in front of stuffing. I would let the gradual stuffing gradient do its impedance matching thing. The neat thing is it's easily adjusted to taste and can be checked with an impedance sweep. The flatter the peak and less ripples the better.
in the Dagger I wouldn't put felt in front of stuffing. I would let the gradual stuffing gradient do its impedance matching thing. The neat thing is it's easily adjusted to taste and can be checked with an impedance sweep. The flatter the peak and less ripples the better.
I'll be looking at the impedance curves with and without that blanket of wool felt. It really made a difference with the arrays and their filling. It might not be so in this case though. If I end up using the Twaron Angel Hair it might not be needed at all. The Twaron did measure really well in midrange. But I have to say the simple fibre glass insulation does work real good on the entire full range, combined with wool felt it is a hard to beat combination. I'll probably publish the impedance curves soon. I'll let my paint job on the enclosures cure an extra day before I start testing this stuff(ing).
Up to a couple of cm behind the driver, yes... All based on the impedance tests:
Posts with Impedance testing start here...
Posts with Impedance testing start here...
Last edited:
Wesayso,
Your curve is in pair with most of the curves I have seen on french forums when the first digital pre with EQ were launched : especially after 1 K hz ! they spend a lot of times to agree on Something very close to yours after 1 K hZ (often with the most dynamic design if I remember : horns + CD !)
Do you believe the 5 dB bump in the low end is due to the room, I find this 20 hZ quasi 200 hZ at 5 dB very cool !
Did you try a large notch (idem ; progressiv lowering to -5 dB) between 1 or 2 K hZ to 8 kHz (wilde curve with -5 dB at around 4 to 5 K hZ approximatly) then a progressive climb at 0 dB to 20 kHz ?
Hey your tower has a Professional finish, superb !
Your curve is in pair with most of the curves I have seen on french forums when the first digital pre with EQ were launched : especially after 1 K hz ! they spend a lot of times to agree on Something very close to yours after 1 K hZ (often with the most dynamic design if I remember : horns + CD !)
Do you believe the 5 dB bump in the low end is due to the room, I find this 20 hZ quasi 200 hZ at 5 dB very cool
!Did you try a large notch (idem ; progressiv lowering to -5 dB) between 1 or 2 K hZ to 8 kHz (wilde curve with -5 dB at around 4 to 5 K hZ approximatly) then a progressive climb at 0 dB to 20 kHz ?
Hey your tower has a Professional finish, superb !
Basically I've tried everything. All curves I could think of and anything in between. The curve I ended up with (with still some minor tweaks here and there) is what I posted. But I'm also still working on optimising my processing settings (DRC-FIR).
The low end definitely is due to room gain with the corner loading, especially on one side. I still mix and match to overcome room modes below ~70 Hz. The right side does 50-70 Hz better, the left side excels at the first octave. So to never need too much boost I swap a bit of energy down low between the channels.
A raise at higher frequencies above 8 KHz can sound good on some material but degrade other songs. The curve I worked on plays most songs very well and is most pleasant on a very wide choice of music. The arrays don't discriminate any genre. Basically I like everything I throw at them although not every song excels at imaging or displays time coherency that well. Unfortunately some material is from bands I like a lot. It's not that they sound bad, but the good recordings sound way more engaging.
Thanks for the compliment, I tried to make it look and feel like it could have come from a store. Not the over the top (to me) audiophile kind, more the "we are meaning business" type of look. Does that make sense?
The low end definitely is due to room gain with the corner loading, especially on one side. I still mix and match to overcome room modes below ~70 Hz. The right side does 50-70 Hz better, the left side excels at the first octave. So to never need too much boost I swap a bit of energy down low between the channels.
A raise at higher frequencies above 8 KHz can sound good on some material but degrade other songs. The curve I worked on plays most songs very well and is most pleasant on a very wide choice of music. The arrays don't discriminate any genre. Basically I like everything I throw at them although not every song excels at imaging or displays time coherency that well. Unfortunately some material is from bands I like a lot. It's not that they sound bad, but the good recordings sound way more engaging.
Thanks for the compliment, I tried to make it look and feel like it could have come from a store. Not the over the top (to me) audiophile kind, more the "we are meaning business" type of look. Does that make sense?
Thanks for the tip about the upper 8 K. I putted such a notch but narrower centered at 4.5 k Hz on my aluminium tweeters (width : 2.5 k to 6.5 k iirc ! worked great for me.
I have a basic EQ on the streamer of my NAS which feed the AYA 2 2014 DAC : have to try this curve you linked above ! (tweeter beginns at 2.6 K on my speakers !). It push also to choose drivers in a normal 3 way design with less efficienty that the low mid !
Your are all crazy enclosures makers in this Full Range section ! I surmise yours costed more than the 50 drivers !
The array lines are ok to my eyes, I don't like the curved one of Dr Keele (too big for the curve, it's Strange for a living room. It's not when the curved speaker is not big like the little ESL 57 in an other style !)
PS : yes the good looking makes sense in a living room. Your design is too complicate for my skills, but I have to read this thread as well as the transcient of such a system should be very good. Have you heard the two ways of Dr Keele with the little tweeters ?
I have a basic EQ on the streamer of my NAS which feed the AYA 2 2014 DAC : have to try this curve you linked above ! (tweeter beginns at 2.6 K on my speakers !). It push also to choose drivers in a normal 3 way design with less efficienty that the low mid !
Your are all crazy enclosures makers in this Full Range section
! I surmise yours costed more than the 50 drivers ! The array lines are ok to my eyes, I don't like the curved one of Dr Keele (too big for the curve, it's Strange for a living room. It's not when the curved speaker is not big like the little ESL 57 in an other style !)
PS : yes the good looking makes sense in a living room. Your design is too complicate for my skills, but I have to read this thread as well as the transcient of such a system should be very good. Have you heard the two ways of Dr Keele with the little tweeters ?
Last edited:
Thanks for the tip about the upper 8 K. I putted such a notch but narrower centered at 4.5 k Hz on my aluminium tweeters (width : 2.5 k to 6.5 k iirc ! worked great for me.
I have a basic EQ on the streamer of my NAS which feed the AYA 2 2014 DAC : have to try this curve you linked above ! (tweeter beginns at 2.6 K on my speakers !). It push also to choose drivers in a normal 3 way design with less efficienty that the low mid !
Your are all crazy enclosures makers in this Full Range section
The array lines are ok to my eyes, I don't like the curved one of Dr Keele (too big for the curve, it's Strange for a living room. It's not when the curved speaker is not big like the little ESL 57 in an other style !)
PS : yes the good looking makes sense in a living room. Your design is too complicate for my skills, but I have to read this thread as well as the transcient of such a system should be very good. Have you heard the two ways of Dr Keele with the little tweeters ?
If you find wesayso´s enclosures too complicated, you could try this design:
http://www.diyaudio.com/forums/multi-way/193015-stupid-cheap-line-array-21.html
Attachments
NO, I haven't heard the Keele CBT. I've read a lot about it though. I would have thought it would make a bigger stir than it did. For a while it had a good buzz just before it came out. But somehow I think it didn't quite have the appeal it needed. I can't judge the sound but I agree on the somewhat awkward shape for a living room. Arrays that fit on a piece of A4 paper however seem ideal to me!
Before this build I've studied all I could get my hands on to get a bit of a grip on the art of musical reproduction. In that time I tried to absorb most of the different theories and views on this matter. After making my choice for this type of speaker I continued to question if it was applicable to my proposed build. I still do sometimes. I read new theory and try to see where my speakers fit in that picture. But my list of positives for an array outweighs the negatives and that's why I believed in this principle. Now I'm just glad I did it.
A lot of what I studied and read about I now try in practice. To educate myself, mostly and share my findings with those interested. At the point where I had my enclosures still in bare wood with polyurethane coat and it cracked I had to think real hard weather or not to throw it online. Partly it felt like failure, although it was a risk I was willing to take at the time to preserve the wood finish. I always knew it could happen. But do you display a failure like that? I'm glad I did. I'm here to learn among other things, got a lot of positive reactions and I'm glad I shared the pain. After that I figured: throw it out there, you won't always be right, but what other path is there to learn new things than to be wrong every now an then...
After all is said and done, yes... these were not cheap to build, even excluding the personal time. I knew that going in. But it was definitely worth it to me. I can build you a pair if you want. But it will never be cheap. At one time I even dreamed of doing this as a job/business. Not to get rich, but to do what I love doing. If it pays the bills it would have been good enough for me. But the asking price would be something like € 25000,00 for a pair. To me that's a crazy high amount of money to spend on loudspeakers. Plus the 'customer' would have to wait a year for me to finish them. When is a product really worth it's price?
To me, the time and money invested still feel like time and money well spend. I loved the journey, the design part, problem fixing and just adore the listening part. It was planned from the start to combine the computer with the audio to see where I could take this. After all, that was/is my real job/business, ICT.
So far I haven't been disappointed. The back up solution, a Behringer DEQ 24/96 is still in it's box after the purchase in 2011. Today I fiddled around with JRiver setting up the ambience channels. That reminded me why that is the ultimate solution. The ease and possibilities a piece of software like that brings is so liberating! I can basically send everything I want to the ambient channels. My own mind/fantasy being the only restriction. I guess you guys can tell I really love this fiddling... I'll shut up now.
But before I go: the number of views on this thread is unbelievable to me, almost 250K. Thanks everyone for the support, critics and fun here!
Before this build I've studied all I could get my hands on to get a bit of a grip on the art of musical reproduction. In that time I tried to absorb most of the different theories and views on this matter. After making my choice for this type of speaker I continued to question if it was applicable to my proposed build. I still do sometimes. I read new theory and try to see where my speakers fit in that picture. But my list of positives for an array outweighs the negatives and that's why I believed in this principle. Now I'm just glad I did it.
A lot of what I studied and read about I now try in practice. To educate myself, mostly and share my findings with those interested. At the point where I had my enclosures still in bare wood with polyurethane coat and it cracked I had to think real hard weather or not to throw it online. Partly it felt like failure, although it was a risk I was willing to take at the time to preserve the wood finish. I always knew it could happen. But do you display a failure like that? I'm glad I did. I'm here to learn among other things, got a lot of positive reactions and I'm glad I shared the pain
. After that I figured: throw it out there, you won't always be right, but what other path is there to learn new things than to be wrong every now an then...After all is said and done, yes... these were not cheap to build, even excluding the personal time. I knew that going in. But it was definitely worth it to me. I can build you a pair if you want
. But it will never be cheap. At one time I even dreamed of doing this as a job/business. Not to get rich, but to do what I love doing. If it pays the bills it would have been good enough for me. But the asking price would be something like € 25000,00 for a pair. To me that's a crazy high amount of money to spend on loudspeakers. Plus the 'customer' would have to wait a year for me to finish them. When is a product really worth it's price?To me, the time and money invested still feel like time and money well spend. I loved the journey, the design part, problem fixing and just adore the listening part. It was planned from the start to combine the computer with the audio to see where I could take this. After all, that was/is my real job/business, ICT.
So far I haven't been disappointed. The back up solution, a Behringer DEQ 24/96 is still in it's box after the purchase in 2011. Today I fiddled around with JRiver setting up the ambience channels. That reminded me why that is the ultimate solution. The ease and possibilities a piece of software like that brings is so liberating! I can basically send everything I want to the ambient channels. My own mind/fantasy being the only restriction. I guess you guys can tell I really love this fiddling... I'll shut up now.
But before I go: the number of views on this thread is unbelievable to me, almost 250K. Thanks everyone for the support, critics and fun here!
I had hesitation two or three years ago to bought some RD75 both because waf, (big looking array) and the problems a guy had here with his pair bought at Part Express !
(I lurked at Stig Erik design with it)... I have some regret today, not sourcable anymore !
Foam also opens some space to make easier the making process of the front bafles without CNC and woods workshop !
(I lurked at Stig Erik design with it)... I have some regret today, not sourcable anymore !
Foam also opens some space to make easier the making process of the front bafles without CNC and woods workshop !
I lusted after those as well... but got worried when I read a couple of failure reports.
No CNC was used by me to create these. Just a router, jig saw and some water yet cut templates. And time, a lot of time. But Koldby's proposed enclosures would be do-able, even without CNC. It's not going to be as quick or easy though.
Last edited:
Slippery enclosures, hard to photograph with my phone...
Painted with roller to get the orange peal in satin black. They're still a bit too glossy for me. But for now they'll do.
Edit: it's silk gloss, I guess I want Satin. I toned down the arrays gloss but that was a strong 2K boat paint. These were painted
with the cheapest paint I could find.
Painted with roller to get the orange peal in satin black. They're still a bit too glossy for me. But for now they'll do.
Edit: it's silk gloss, I guess I want Satin. I toned down the arrays gloss but that was a strong 2K boat paint. These were painted
with the cheapest paint I could find.
Last edited:
Indeed I covered it, but no problem to try and explain it again. These two speakers, powered with the Scan Speak 10F will be used behind the listening position, spread wide and pointed up.
The signal they will be getting will be (L-R) and (R-L) bandwidth limited to 200 Hz up to 3500 Hz, attenuated compared to the arrays, dialling it in at the listening position. Further they will be delayed to about 20 ms behind the main impulse peak.
What that will do (hopefully) is create the perception of a bigger listening space. So far I've done a lot of things to lessen the effect of early reflections. They are still there but down by 20 dB or more after the first peak. This virtually created back 'reflection' will make the listener 'believe' he's listening in a big room and make him/her more at ease, because we as humans are always scouting the room for clues. We do it with our eyes, but also with our ears. In the studio world I've noticed they call it a termination. It makes us focus on the part before it. And as a consequence stop scouting the room for (room) clues.
There, that's the theory behind it in a nutshell. The above is my starting point going in. Most of it I believe strongly due to the experiments I've done so far.
If I am correct here you won't notice the back channels by themselves. But they do create a sense of envelopment. They are to early to be noticed as echo's, too late to be perceived as part of the direct sound.
Also known as a Haas kicker. I've read about it in 2008/2009 somewhere and knew I had to try it in my house. But my living room isn't big enough so I go for a virtual solution.
The signal they will be getting will be (L-R) and (R-L) bandwidth limited to 200 Hz up to 3500 Hz, attenuated compared to the arrays, dialling it in at the listening position. Further they will be delayed to about 20 ms behind the main impulse peak.
What that will do (hopefully) is create the perception of a bigger listening space. So far I've done a lot of things to lessen the effect of early reflections. They are still there but down by 20 dB or more after the first peak. This virtually created back 'reflection' will make the listener 'believe' he's listening in a big room and make him/her more at ease, because we as humans are always scouting the room for clues. We do it with our eyes, but also with our ears. In the studio world I've noticed they call it a termination. It makes us focus on the part before it. And as a consequence stop scouting the room for (room) clues.
There, that's the theory behind it in a nutshell. The above is my starting point going in. Most of it I believe strongly due to the experiments I've done so far.
If I am correct here you won't notice the back channels by themselves. But they do create a sense of envelopment. They are to early to be noticed as echo's, too late to be perceived as part of the direct sound.
Also known as a Haas kicker. I've read about it in 2008/2009 somewhere and knew I had to try it in my house. But my living room isn't big enough so I go for a virtual solution.