To me, the crossover at 60 Hz makes the plots unusable. First, I would never use a crossover to subs and second certainly NOT at 60 Hz. What you need are plots showing each source independently. Then you can decide how to divy up the sources based on the responses.
I don't know how many times I have expressed that my technique does not use the concept of "crossovers", and yet people use them all the time and wonder what went wrong!
Yes, I'm aware you don't recommend any crossover and I would be happy not to use it. I have done it before and it was great. But this new place of mine gives me a headache.
Speaking of that, does no crossover rule apply to all sorts of speakers? Mine go down do 40 Hz and until now I have never thought this might be a problem when adding subs.
Plots with crossover at 60 Hz are there for comparison as I tried to disable the speaker ports and wanted to see some variations. There are other plots without crossover and there are plots for each speaker as well.
Here, I took new ones yesterday:
WeTransfer
I see no reason to use a "crossover" on any multisub situation. I might in some cases use a 6 dB/oct HP filter on the mains if I felt that they could see significant excursions. But in my system there is no HP as the drivers are all large 15" pro units that would never even come close to excursion limitations in practice. Small drivers have no business doing LFs.
Of course there is always some LP filters on the subs, but seldom, if ever, will they be at the same frequencies as the mains or other subs. The idea is to blend the sources together with as much overlap as possible. Crossovers completely disable this possibility.
If your interest is the effect of ports, then I don't see the need for any crossover or the mains. Just show the subs with and without ports. The mains don't enter into the picture in this case.
Of course there is always some LP filters on the subs, but seldom, if ever, will they be at the same frequencies as the mains or other subs. The idea is to blend the sources together with as much overlap as possible. Crossovers completely disable this possibility.
If your interest is the effect of ports, then I don't see the need for any crossover or the mains. Just show the subs with and without ports. The mains don't enter into the picture in this case.
Earl,
Some additional information. My subs are closed design. Mains (with 12" LF units) have ports and I was testing the difference. Looking at the comparison plot opened ports add 6-7 dB at 25-60 Hz area. Not sure if that would make the integration more difficult.
On the other hand, I was testing delay on L/R speaker. Turned out I had to look further (up to 500 Hz) to see delay is not as good as I though in the first place (blue plot: ports open, green plot: ports closed).
What do you think (the plots are in wetransfer link above).
Some additional information. My subs are closed design. Mains (with 12" LF units) have ports and I was testing the difference. Looking at the comparison plot opened ports add 6-7 dB at 25-60 Hz area. Not sure if that would make the integration more difficult.
On the other hand, I was testing delay on L/R speaker. Turned out I had to look further (up to 500 Hz) to see delay is not as good as I though in the first place (blue plot: ports open, green plot: ports closed).
What do you think (the plots are in wetransfer link above).
Attachments
I'd want to start with them closed. (To me that should always be the case for mains used with multiple subs.) Then add some subs and see what happens.
Like I said best would be multiple traces with mono mains, and individual subs shown independently. Then you start tweaking. First try and find the best place for the LP filters on each sub (not the same frequency) and see how they blend with the mains. Then start going lower and lower from say 200 Hz and fix the peaks and dips with EQ. This can be global or individual, but global is easier. I set limited EQ on each sub and then fix the result with some global EQ.
Like I said best would be multiple traces with mono mains, and individual subs shown independently. Then you start tweaking. First try and find the best place for the LP filters on each sub (not the same frequency) and see how they blend with the mains. Then start going lower and lower from say 200 Hz and fix the peaks and dips with EQ. This can be global or individual, but global is easier. I set limited EQ on each sub and then fix the result with some global EQ.
I would make the mains mono and treat as a single source for the purposes of setting the EQ. The playback will of course be stereo, but since you are really only setting the EQ below say 120 Hz. the differences will be minimal. Thus you have 1 mains source and X subs, so you should have X+1 lines to show.
Honestly, this is a minor point - about what signal to use where in the mains - you can do it however you want, but going with different signals from each main separately is going to double the work, without I would suggest, nothing comparable in performance differential.
I am not familiar with "integration process". Is that a step? I don't see it as a "step" in what I do.
I am not familiar with "integration process". Is that a step? I don't see it as a "step" in what I do.
I understand most of the Geddes style multiple sub setup but cannot understand the lack of lowpass filters on the subs ?
Here's a measurement of my old tempest sealed sub at listening position (no EQ /XO). At 800Hz it's louder than at 20/30Hz.
A lot of people run subs a bit louder than the mains (I use the JBL synthesis curves for my system) So their sub could be outputting 800Hz material louder than their mains.
Unless Geddes assumes we're all using AV processors with built in crossovers on the sub channel I'm not sure I get it ?
Rob.
Here's a measurement of my old tempest sealed sub at listening position (no EQ /XO). At 800Hz it's louder than at 20/30Hz.
A lot of people run subs a bit louder than the mains (I use the JBL synthesis curves for my system) So their sub could be outputting 800Hz material louder than their mains.
Unless Geddes assumes we're all using AV processors with built in crossovers on the sub channel I'm not sure I get it ?
Rob.
Attachments
What you are missing is that I have never suggested not using "lowpass filters on the subs". I always have. The question is where! They should not all be at the same frequency. Some sub/positions work better for the upper LFs and thus their LP freq is higher. The idea is to blend everything together smoothly, not abruptly.
"No crossover" is not the same thing as "No LP".
"No crossover" is not the same thing as "No LP".
Last edited:
Hi, Earl,
Before that optimal placement of mains can is IMHO crucial so one can eliminate as many peaks/dip as he can.
Does that mean I should measure L+R together and then start adding the subs?
Sorry for the confussion, my english is not as good as I'd want it to be.
Integration process is a complete process you described in your thesis. Add the nearest sub and measure. Add second one, third one ... and then EQ.
Before that optimal placement of mains can is IMHO crucial so one can eliminate as many peaks/dip as he can.
You said you would make mains mono and treat them as a single source.
Does that mean I should measure L+R together and then start adding the subs?
Sorry for the confussion, my english is not as good as I'd want it to be.
Hi,
I understand what it is about, but i don't know how our brain deal with what we try to do, and would be curious to ear your point about this : If a acoustical instrument interact with room in real life, don't making response uniform over the room, is making stuff finally less realistic ? Not a ideal example, but in room double bass sound nice compared to outdoor...
I understand what it is about, but i don't know how our brain deal with what we try to do, and would be curious to ear your point about this : If a acoustical instrument interact with room in real life, don't making response uniform over the room, is making stuff finally less realistic ? Not a ideal example, but in room double bass sound nice compared to outdoor...
I said I understand
.Let me rephrase : do smooth response is natural and to what point is it desirable ? don't some peaks are more natural than no peaks ? It's not a "real" question...As perfect directivity goal, wich is not realistic since acoustical instruments doesn't have perfect directivity...Isn't there a futur for audio reproduction, with differential directivity/sound pressure over the room, for different tracks of a song ?
Ok...i go too far That's just a open question.
Rob,
When Geddes says don't use any 'crossovers', what he is saying and rightly so, is don't put a HP filter on the mains along with a LP filter on the sub(s) at the same frequency. That's a crossover. It simple doesn't work acoustically if achieving a smooth bass response is required in room. One should stagger the LP filter frequencies for the subs and that should only be dictated by measurements, period. Without measurements, this entire exercise is impossible. Now some speakers can get by without a HP filter on the mains and some cannot. In Geddes' setup (and mine, along with numerous others who have HUGE woofers on their mains), running the mains full range is the best option in all honesty. Indeed some fellas have mains with very small woofers and excursion limits are easily hit (along with alarming increases in THD and IMD, i.e. > 10% which becomes audible) and putting a HP filter on the mains makes sense, or honestly, just designing new speakers.
I don't use an AV processor. I use either a MiniDSP or Behringer DCX2496. It provides the greatest flexibility and that way I can let the AV processor do its own thing so to speak. Too many AV processors in my opinion do not have the ability to finely adjust all the parameters needed to get a smooth bass response.
You are confusing, conflating and obfuscating SOUND PRODUCTION with SOUND REPRODUCTION. There is a whole encyclopedia on psychoacoustic research on how bass propagates in small rooms, and small room acoustics. I highly suggest you purchase a copy of Tooles' Sound Reproduction book which goes through this in detail. Geddes' had a book as well, now available as a download I think that explains this. Some of these fellas have spent their entire careers proving this, not using hand waving gestures or throwing out opinionated columns.
Here is some of Toole's work:
http://www.wghwoodworking.com/audio/loudspeakers_and_rooms_for_sound_reproduction.pdf
http://www.wghwoodworking.com/audio/LoudspeakersandRooms-WorkingTogether.pdf
http://www.wghwoodworking.com/audio/multsubs.pdf
http://www.wghwoodworking.com/audio/MaximizingLoudspeakerPerformanceInRooms.pdf
Best,
Anand.
I said I understand
Let me rephrase : do smooth response is natural and to what point is it desirable ? don't some peaks are more natural than no peaks ? It's not a "real" question...As perfect directivity goal, wich is not realistic since acoustical instruments doesn't have perfect directivity...Isn't there a futur for audio reproduction, with differential directivity/sound pressure over the room, for different tracks of a song ?
Yes, that is quite brilliant to observe and relates to what I've been calling "Toole's critique" of EQ, the principle of adaptation level in perception (the visual or aural environment you are immersed in becomes your benchmark... until you go into another room), and obliquely the "El Greco fallacy" as well.
A FR can't be heard - you can only hear tone colour. But tone colour can be "off"; for example the Concertgebouw and the Philadelphia Academy of Music are very strong on bass. Too strong? Is it right for the resident orchestras to have fewer double-basses? Nearly every home I see pictured on this forum must certainly be quite "bright" in the treble and with unfortunate reverberation.
So, I would suppose, your room can be "off" too.
B.
Last edited:
Maybe if i wasn't a bit provocative, you didn't defend with such good reading collection,
. 8 years after joining diyaudio, my english level isn't anymore a barrier for those reading. I said I understand what it is about, so, it's implicit, I already know multi-sub is a evident solution in-room
Last edited: