As we are DIY, we have more choices. I suspect many of us woudl not be afraid to cut a hole in the wall and can build boxes in unconventional shapes.
So, my thoughts are such. For ultimate performance, is in-wall actually a better design? Moves the edges out way further, No 2 foot surfaces to mess up imaging. Of course, many rooms would not be suitable.
I was at my sisters where their HT has the three fronts in-wall with pretty decent plates. Imaging was very good and the top end needed less eq for raggedness. In many rooms, this would allow the speakers not to dominate the architecture. In a HT, not that big a deal but for us who listen in a living room and also live in them. Not good for the ego "mine is bigger or more expensive" market of course. Does not work for Maggies. I have not heard things like the B&G ( gone sob sob) or the Martin Logans in wall.
Any experiences? What may I be overlooking?
Another question I was wondering about is how far away can we get away with a tweeter from the mid? I have a difficult HT arrangement requiring L&R to be at the ceiling. Center brings down the voices mostly, but still image can jump around. I could move the tweeters down a foot leaving the mids up where there is physically room.
So, my thoughts are such. For ultimate performance, is in-wall actually a better design? Moves the edges out way further, No 2 foot surfaces to mess up imaging. Of course, many rooms would not be suitable.
I was at my sisters where their HT has the three fronts in-wall with pretty decent plates. Imaging was very good and the top end needed less eq for raggedness. In many rooms, this would allow the speakers not to dominate the architecture. In a HT, not that big a deal but for us who listen in a living room and also live in them. Not good for the ego "mine is bigger or more expensive" market of course. Does not work for Maggies. I have not heard things like the B&G ( gone sob sob) or the Martin Logans in wall.
Any experiences? What may I be overlooking?
Another question I was wondering about is how far away can we get away with a tweeter from the mid? I have a difficult HT arrangement requiring L&R to be at the ceiling. Center brings down the voices mostly, but still image can jump around. I could move the tweeters down a foot leaving the mids up where there is physically room.
Tvrgeek, your sister's HT has good imaging because the entire wall acts like a huge baffle that restricts coverage to 180*, also often improving the reproduction of the lower frequencies by about 6dB (more amp headroom, less power required). In cinemas, THX recommends a baffle wall behind the screen along with a toe-in so as to maintain good stereo imaging far back into the crowd. So, in my opinion, in-wall is a good design if you do not have a projection screen. However, if you have one, just build a THX style baffle wall behind it and forget it.
In-wall speakers cannot use the toe-in that many people (including THX) say is beneficial, as most walls are flat. Nevertheless, if you can somehow get a curved wall surface then it should be possible to get the toe-in as well.
tvrgeek said:
In-wall speakers cannot use the toe-in that many people (including THX) say is beneficial, as most walls are flat. Nevertheless, if you can somehow get a curved wall surface then it should be possible to get the toe-in as well.
Hmmm... in wall doesn't allow for mine is bigger than yours?!
No, it all depends how you integrate it in the room:
https://www.google.com/imgres?imgur...iO2KeiX8Ba8M&w=310&h=163&itg=1&source=sh/x/im
Or
https://images.app.goo.gl/zPUojV4CEHL4UpNc7
And yes those are MTM with 15" in the first picture, 2x12" in the second... by designers which knows what they are doing Tvrgeek. 😉
I've got experience with inwall in the same kind of room than the one linked ( iow pro control room).
What do you want to know exactly?
Newvirus,
Why toe in should be mandatory? It depend of the room, the layout of it and the kind of loudspeakers. The second picture ( with the 2x12") is a multi purpose control room where some cinema mixing is (was) performed regularly.
Here is an example of a domestic realisation performed by a member of this board ( meh with constant directivity).
http://www.cowanaudio.com/
No, it all depends how you integrate it in the room:
https://www.google.com/imgres?imgur...iO2KeiX8Ba8M&w=310&h=163&itg=1&source=sh/x/im
Or
https://images.app.goo.gl/zPUojV4CEHL4UpNc7
And yes those are MTM with 15" in the first picture, 2x12" in the second... by designers which knows what they are doing Tvrgeek. 😉
I've got experience with inwall in the same kind of room than the one linked ( iow pro control room).
What do you want to know exactly?
Newvirus,
Why toe in should be mandatory? It depend of the room, the layout of it and the kind of loudspeakers. The second picture ( with the 2x12") is a multi purpose control room where some cinema mixing is (was) performed regularly.
Here is an example of a domestic realisation performed by a member of this board ( meh with constant directivity).
http://www.cowanaudio.com/
Last edited:
Of course, huge baffle is exactly what I am talking about. No step, no diffraction, and if far enough in, no reflection issues.
Corners put a reflective surface too close unless you can put sufficient treatment on it. 2 inches of OC 704 does wonders. For a HT, corners are too far from the screen. Optimal is in the 2 foot range to the sides and center to top in height opposite of the center. ( low center, push them up a little and vice-versa) For music listening, corners are WAY too far apart. For PA, great.
Brick? Well, maybe you could frame it out so you have some insulation? Then the screen could be truly flush.
THX curved eh? Well not impressed with THX anyway. Toe or not depends entirely on the drivers. Filling a 200 seat room is not the same as a living room.
No Krivium, toe-in is not at all mandatory, I was just mentioning a possible difficulty in achieving it, in case it is desired.
Ok.
I worked in different CR like the second one with in wall LCR and they were more than ok. And if you take a look at all the pictures there is some kind of 'wings' to include a bit of toe in. M. COWAN pictures even show how it is usually implemented.
Of course dedicated cinema mixing auditorium replicate theater ( by using same gear and more or less same layout) but this is not mandatory to have great results.
I worked in different CR like the second one with in wall LCR and they were more than ok. And if you take a look at all the pictures there is some kind of 'wings' to include a bit of toe in. M. COWAN pictures even show how it is usually implemented.
Of course dedicated cinema mixing auditorium replicate theater ( by using same gear and more or less same layout) but this is not mandatory to have great results.
Tvrgeek,
Corners could work great if you use some kind of directivity control (wg or horn). By choosing lower than 90° horizontal you'll minimize a lot the reflection you could face.
Try to find pictures of AllenB system here ( there is even an explanation from his choice as well as description of rendering). Pispeaker's/Geddes kind of approach ( there is differences between them but they are in the same family to me).
Otherwise you seems to have covered most aspect of the build already!
Corners could work great if you use some kind of directivity control (wg or horn). By choosing lower than 90° horizontal you'll minimize a lot the reflection you could face.
Try to find pictures of AllenB system here ( there is even an explanation from his choice as well as description of rendering). Pispeaker's/Geddes kind of approach ( there is differences between them but they are in the same family to me).
Otherwise you seems to have covered most aspect of the build already!
IMHO, the only thing you have to focus on is if the controlled off axis directivity will work as good as a good free stand loudspeaker you like for this purpose.
At iso power response fundamental curve since it is made the same with fine EQ, not sure the two systems, free stand or in wall have the same soundstage as the controlled directivity off axis should differ. You may like one more than the other...some like large baffle for the soundstage while others like more narrow baffle at iso power response. Advantage of a 180 degree vertical spread is also a huge sweet spot... youcare less about the heigth of your sofa and chairs...
We do not talk about bsc and group delay due to difraction bffle step corection as it can be solved with FIR corection nowadays in the good free stand loudspeakers.
The clear advantage on in wall drivers is to make them infinite load bebind and not acoustic suspension spring...no back wave on the rear of the cone equals in theory for better information, no ?
At iso power response fundamental curve since it is made the same with fine EQ, not sure the two systems, free stand or in wall have the same soundstage as the controlled directivity off axis should differ. You may like one more than the other...some like large baffle for the soundstage while others like more narrow baffle at iso power response. Advantage of a 180 degree vertical spread is also a huge sweet spot... youcare less about the heigth of your sofa and chairs...
We do not talk about bsc and group delay due to difraction bffle step corection as it can be solved with FIR corection nowadays in the good free stand loudspeakers.
The clear advantage on in wall drivers is to make them infinite load bebind and not acoustic suspension spring...no back wave on the rear of the cone equals in theory for better information, no ?
Last edited:
Here is AllenB's example i thought about.
https://www.diyaudio.com/community/threads/build-thread-corner-guided-3-way.243680/#post-3656833
And Wayne Pahram paper ( PiSpeakers):
https://www.pispeakers.com/Pi_Speakers_Info.pdf
Why bother with FIR Diyiggy? Bsc is minimal phase so an eq will solve the issue ( low shelf eq) and usually there is virtualy no diffraction with infinite baffle ( nothing that can't be treated with a bit of absorbtion as Tvrgeek stated).
https://www.diyaudio.com/community/threads/build-thread-corner-guided-3-way.243680/#post-3656833
And Wayne Pahram paper ( PiSpeakers):
https://www.pispeakers.com/Pi_Speakers_Info.pdf
Why bother with FIR Diyiggy? Bsc is minimal phase so an eq will solve the issue ( low shelf eq) and usually there is virtualy no diffraction with infinite baffle ( nothing that can't be treated with a bit of absorbtion as Tvrgeek stated).
Last edited:
why bother with FIR?
IMO it makes sense when you need to use a lot of EQ, so digital permits to DSP and manage phase easilier with FIR where it makes sense, so in the bass to low mid area. I was talking for the narrow baffle about the BSC as on a wall the diffraction should be way less i.e. the wave length of the front wall I surmise ?
However a linkwitz transform that still plays in the half of the 1k Hz so around 400 hz to 600 hz for the middle F3 baffle step point, I imagine a linkwitz transform in IIR or electronic part or passive will affect a lot the group delay where the ears are sensible to that phenomena (not so after 1 k hz)...
That's what I wanted to input... as an open question as theorical for me, not experienced in the true life halas.
However, however and this is a question ; if you say a the baffle step effect is minimal phase and say one cut off in the F3 so middle of the baffle step loss, is it still minimal phase if you use a LP and HP filter with a Q > 0.5. And at best if one used a 6BUT the high pass of the baffle step should have a 90° phase delay ?
IMO it makes sense when you need to use a lot of EQ, so digital permits to DSP and manage phase easilier with FIR where it makes sense, so in the bass to low mid area. I was talking for the narrow baffle about the BSC as on a wall the diffraction should be way less i.e. the wave length of the front wall I surmise ?
However a linkwitz transform that still plays in the half of the 1k Hz so around 400 hz to 600 hz for the middle F3 baffle step point, I imagine a linkwitz transform in IIR or electronic part or passive will affect a lot the group delay where the ears are sensible to that phenomena (not so after 1 k hz)...
That's what I wanted to input... as an open question as theorical for me, not experienced in the true life halas.
However, however and this is a question ; if you say a the baffle step effect is minimal phase and say one cut off in the F3 so middle of the baffle step loss, is it still minimal phase if you use a LP and HP filter with a Q > 0.5. And at best if one used a 6BUT the high pass of the baffle step should have a 90° phase delay ?
Last edited:
Diyiggy,
You mix different things together:
Bsc is minimal phase or it could not been compensated for by use of eq ( whatever the way you choose to apply it: passive, active, dsp,...as long as it is IIR).
As such once it is applied there is no reason group delay will be worst than initially, in fact it should be the inverse as the frequency response will be 'flater' overall.
To correct it i've never used anything more than a shelf filter (eq). From there you have 2 possible choice either a low shelf and use a boost, either a high shelf in cut. Both could give same final eq profile but one method will bring things than the other doesn't: increase of headroom and overall a more transparent result* at the cost of a lost of overall sensitivity. You guessed this is when you use high shelf in attenuation rather than a low shelf in boost.
You talk about Linkwitz Transform ( which is accomplished through IIR you are right) but i must say i don't see why in that case. I use this extensively but not to compensate for bsc, only to extend freq bandwidth ( or restrict it) or modify the 'native' Q of a closed box ( eg:from 0,7 to 0,577). But maybe you thoughts or have a different method i don't know about?
Anyway here again it doesn't modify the fact that here again this modifications are minimum phase too: once applied the system is closer to a theorical ideal (whatever it is) and the phase behavior will be close to the target you choose.
In both case there is no need for FIR to compensate for phase as you already corrected it by changing frequency behavior with IIR ( phase and amplitude are linked, correct one you change the other too).
Then you talk about xover behavior. In this case then yes you can run into trouble with phase with IIR depending of the type of global solution you have choosen ( or not it'll depend of the choice you made: bessel, LR 12db, LR24db, ...).
In this case then yes you could 'compensate' an issue using FIR. But... in my experience using a FIR profile to compensate a 'global' filter behavior ( using a stereo FIR profile before/upstream a passive filter to correct both Lp and Hp behavior for example) will be less transparent than using complentary FIR xovers ( so no more passive xover and multiamp required).
I'm not the only one to have come to this conclusion and some other members came to same observation by themself.
So i disagree when you say you need FIR when you use a lot of eq: if you use an hardware dsp ( loudspeaker management system) it sound cleaner to mix both IIR eq and complementary FIR filter for xover. The same is true from software dsp ime as long as you try to implement same approach
Once you introduce room correction or other more complex tricks*, then things could be different though and in this case it may be 'easier' to integrate xover/eq/'room correction' into one profile for each loudspeaker's way rather than to have different blocks of treatments chained the 'usual' way. Or you may prefer to have less steps for 'convenience' ( keep a profile 'proprietary', ease of use,...).
*eg: Fulcrum audio licensed FIR treatments to compensate artefacts for coaxials with 'Altec style' horns induce.
You mix different things together:
Bsc is minimal phase or it could not been compensated for by use of eq ( whatever the way you choose to apply it: passive, active, dsp,...as long as it is IIR).
As such once it is applied there is no reason group delay will be worst than initially, in fact it should be the inverse as the frequency response will be 'flater' overall.
To correct it i've never used anything more than a shelf filter (eq). From there you have 2 possible choice either a low shelf and use a boost, either a high shelf in cut. Both could give same final eq profile but one method will bring things than the other doesn't: increase of headroom and overall a more transparent result* at the cost of a lost of overall sensitivity. You guessed this is when you use high shelf in attenuation rather than a low shelf in boost.
You talk about Linkwitz Transform ( which is accomplished through IIR you are right) but i must say i don't see why in that case. I use this extensively but not to compensate for bsc, only to extend freq bandwidth ( or restrict it) or modify the 'native' Q of a closed box ( eg:from 0,7 to 0,577). But maybe you thoughts or have a different method i don't know about?
Anyway here again it doesn't modify the fact that here again this modifications are minimum phase too: once applied the system is closer to a theorical ideal (whatever it is) and the phase behavior will be close to the target you choose.
In both case there is no need for FIR to compensate for phase as you already corrected it by changing frequency behavior with IIR ( phase and amplitude are linked, correct one you change the other too).
Then you talk about xover behavior. In this case then yes you can run into trouble with phase with IIR depending of the type of global solution you have choosen ( or not it'll depend of the choice you made: bessel, LR 12db, LR24db, ...).
In this case then yes you could 'compensate' an issue using FIR. But... in my experience using a FIR profile to compensate a 'global' filter behavior ( using a stereo FIR profile before/upstream a passive filter to correct both Lp and Hp behavior for example) will be less transparent than using complentary FIR xovers ( so no more passive xover and multiamp required).
I'm not the only one to have come to this conclusion and some other members came to same observation by themself.
So i disagree when you say you need FIR when you use a lot of eq: if you use an hardware dsp ( loudspeaker management system) it sound cleaner to mix both IIR eq and complementary FIR filter for xover. The same is true from software dsp ime as long as you try to implement same approach
Once you introduce room correction or other more complex tricks*, then things could be different though and in this case it may be 'easier' to integrate xover/eq/'room correction' into one profile for each loudspeaker's way rather than to have different blocks of treatments chained the 'usual' way. Or you may prefer to have less steps for 'convenience' ( keep a profile 'proprietary', ease of use,...).
*eg: Fulcrum audio licensed FIR treatments to compensate artefacts for coaxials with 'Altec style' horns induce.
Thanks for the elaborate answer.
Makes me think I didn t understand what minimal phase expression meant.
What I believed is if one correct the power response then the group delay was affected and only a delay could be dealed in FIR domain for a correct transcient timing. I missed the point...my bad.
Makes me think I didn t understand what minimal phase expression meant.
What I believed is if one correct the power response then the group delay was affected and only a delay could be dealed in FIR domain for a correct transcient timing. I missed the point...my bad.