Thanks for your post. For different days, and resetting microphone the results are very consistent.
Though I have had chances to do a measurement right now I just can't stop listening 😛... After all, that's why I made them!
In time, I'm sure...
Not really sure what changed but it is engaging involving and dynamic.
I've had that separately but not all at once. So I'm enjoying it for now.
Sure I notice some things are still a bit off, I played with toe-in toe-out and did not re-measure in between. At least I get a taste of what it CAN do, and like it very much!
If you ask me what did the trick I honestly don't know yet. Apparently small changes can have a huge impact. But as of now I'm still unable to extract a recipe from that that will work each time.
In time, I'm sure...
Not really sure what changed but it is engaging involving and dynamic.
I've had that separately but not all at once. So I'm enjoying it for now.
Sure I notice some things are still a bit off, I played with toe-in toe-out and did not re-measure in between. At least I get a taste of what it CAN do, and like it very much!
If you ask me what did the trick I honestly don't know yet. Apparently small changes can have a huge impact. But as of now I'm still unable to extract a recipe from that that will work each time.
Last edited:
Have you run any "Windows update" lately 😛.
Enjoy listening that high performance level 🙂.
The only experiments I would be inclined to do with such a speaker are with toe in/out, pulling them about three feet away from walls (just for listening), and adding a sub <70Hz.
I've always found that reducing cone movement brings greater realism.
Here is link to a little something from the treasure trove; not likely your general listening. A live recording with no compression, no limiting, just a little fade out at the end of audience response.
https://www.dropbox.com/s/xsg001o1d9e50gq/audio_04.wav?dl=0
Dial up the volume and enjoy!
I've always found that reducing cone movement brings greater realism.
Here is link to a little something from the treasure trove; not likely your general listening. A live recording with no compression, no limiting, just a little fade out at the end of audience response.
https://www.dropbox.com/s/xsg001o1d9e50gq/audio_04.wav?dl=0
Dial up the volume and enjoy!
Last edited:
Thanks for the clip, it is indeed different from my main stream material. As soon as I have the house to myself I'll sit down and enjoy.
Though I agree with your suggestions above my challenge lies in optimising this setup I have, as it is the compromise I could agree on with my girl. It would still make sense to me to try some bottom end reinforcement at some point. In time I will probably try that as I do notice I enjoy the sound reinforcement at the bottom end very much down to 20 Hz.
Though I agree with your suggestions above my challenge lies in optimising this setup I have, as it is the compromise I could agree on with my girl. It would still make sense to me to try some bottom end reinforcement at some point. In time I will probably try that as I do notice I enjoy the sound reinforcement at the bottom end very much down to 20 Hz.
Correct me if you think I'm wrong about this, but with real music (not test signals) doesn't every sound, every expression of any instrument get delivered in an amplitude envelop that could easily have significant energy in the sub 40HZ frequencies?
If true, then having bass that is acoustically flat down to 20HZ should improve the sense of realness, even on program material that has no bass instrument.
The one catch I've noticed on my flat to 20HZ system, is that many recordings that have bass down that low seem to come with EQ that tries to make up for the typical speaker that only barely makes it down to 40HZ. So there can be too much. It gets flabby. This is where my 4 section Baxandall tone control circuit has worked well for me.
If true, then having bass that is acoustically flat down to 20HZ should improve the sense of realness, even on program material that has no bass instrument.
The one catch I've noticed on my flat to 20HZ system, is that many recordings that have bass down that low seem to come with EQ that tries to make up for the typical speaker that only barely makes it down to 40HZ. So there can be too much. It gets flabby. This is where my 4 section Baxandall tone control circuit has worked well for me.
The most fun I've found are the 'mistakes' down low that shouldn't be there 🙂. But I know what you're saying here. When its right, it sounds more right with the sub frequency extention. I actually tried cutting the low end for a while. I started missing the feeling it gave me. I guess the same could very well be real on the other part of the spectrum. Though I know and realise I don't hear as much there as I used to, I still believe it matters. I use my 9 year old son to tell me I overcompensated 🙂. Though the full rangers will never do it with the same ease as a tweeter, something about this coherent sound keeps reeling me in. Again, in time I'll add an ambient tweeter experiment.
Generally speaking women have better hearing then men... I used to have my wife come listen to one or two tracks she liked. Then in very few words she would shatter my pride. But sadly she was always right and that is why I work in IT and not speaker design... 🙂
Percussion is where the low end extension really makes a big difference to me. You almost forget how good a drum can sound until you hear it down to 30HZ acoustically somewhat flat (assuming a very good recording of coarse).
An Audiologist (medically trained) once told me that it's normal for the inner ear cavity to build up pressure, which dampens the ear drum membrane, causing a change in the frequency response. Much like when you are driving through the mountains and the sense of bass seems suppressed until your ears "pop" and all of a sudden things sound so much clearer. This is one of the many reasons why ears aren't to be trusted over measurements. The trick is to figure out which measurements and how to do them right.
An Audiologist (medically trained) once told me that it's normal for the inner ear cavity to build up pressure, which dampens the ear drum membrane, causing a change in the frequency response. Much like when you are driving through the mountains and the sense of bass seems suppressed until your ears "pop" and all of a sudden things sound so much clearer. This is one of the many reasons why ears aren't to be trusted over measurements. The trick is to figure out which measurements and how to do them right.
Hi Bob
Why not take this test, the poll is still open. Attach a txt file of your findings as not to spoil others taking the vote. I'd love to see what you think of these clips.
http://www.diyaudio.com/forums/full-range/279352-subjective-blind-comparison-2in-4in-drivers-round-4-a.html
You've visited it, but did you try it?
I agree on your perception of percussion. And I believe Hans Groothuis whom I mentioned earlier was using that exact masking principle you mentioned to make his speakers sound larger than they actually were.
Why not take this test, the poll is still open. Attach a txt file of your findings as not to spoil others taking the vote. I'd love to see what you think of these clips.
http://www.diyaudio.com/forums/full-range/279352-subjective-blind-comparison-2in-4in-drivers-round-4-a.html
You've visited it, but did you try it?
I agree on your perception of percussion. And I believe Hans Groothuis whom I mentioned earlier was using that exact masking principle you mentioned to make his speakers sound larger than they actually were.
Last edited:
Time again for a more regular program addition after messing with the conjugation. As said I've been extremely pleased with removing the RC part of the conjugation network. I still have the big RCL in place though.
No separate investigation done there yet. So far I'm not measuring big differences there, except for the already shown change in impedance plots.
I didn't do a direct measurement session to compare. I had other things on my mind at this time.
By making small changes in my overall Pré EQ (needed to boost the high and low end to fall within DRC's comfort zone) I am able to steer the responses and shape of the signal and phase summation. Prior measurements I did were just Left, Right and a control session of the Left plus Right. Usually the sum displayed a drop in high frequencies. Caused by a phase difference between the left and right signal. I noticed small but sure changes by re-visiting my pré EQ and in that way the shape of the STEP of the raw EQ-ed response. I could also soften the ringing after the main pulse by choosing my EQ more carefully.
So back to measurement sessions to optimise that part of the deal!
Here's the left, right and the sum I have now (gated 4 cycles):
To add to this story, the green line with wobbles at 130 and 200 Hz is the right channel with an unavoidable room effect. At 4 cycles it is worse than the cycles before it. I show 4 cycles (REW chooses the smoothing here) to show the first wave front arriving at the listening position.
Good to see the little changes in pré EQ help to get proper summation at higher frequencies. The sum clearly follows the curve of the Left and right signal now.
You can also see the trade I make in lower frequencies where in the left channel I have a problem at 70 Hz (room mode) and compensate for that with the right channel. The sum is nearly flat.
Here's the 4 cycles gated left plus right sum and it's phase...
Pretty self explanatory I'd say.
If we look at it in a Wavelet:
A 40 ms Wavelet to show the first wave front at the listening position. It clearly is in a somewhat life room (partly treated living room) but all in all I'm not unsatisfied with it. Pretty straight line for the highest SPL level. The low frequencies aren't done yet so it's normal that isn't straight. I can't filter 4 cycles here.
If I do that with the group delay (4 cycles gate) and 1/3 smoothing it looks like this:
Obvious by looking at the phase plot but still...
I'd say goal pretty much reached in creating a time coherent arrival at the listening position! Best part of coarse is it sounds better and better with each little improvement I manage to make.
Sometimes it takes a while to find a bit extra but as long as I'm able to think up ways I'll go on.
No separate investigation done there yet. So far I'm not measuring big differences there, except for the already shown change in impedance plots.
I didn't do a direct measurement session to compare. I had other things on my mind at this time.
By making small changes in my overall Pré EQ (needed to boost the high and low end to fall within DRC's comfort zone) I am able to steer the responses and shape of the signal and phase summation. Prior measurements I did were just Left, Right and a control session of the Left plus Right. Usually the sum displayed a drop in high frequencies. Caused by a phase difference between the left and right signal. I noticed small but sure changes by re-visiting my pré EQ and in that way the shape of the STEP of the raw EQ-ed response. I could also soften the ringing after the main pulse by choosing my EQ more carefully.
So back to measurement sessions to optimise that part of the deal!
Here's the left, right and the sum I have now (gated 4 cycles):
To add to this story, the green line with wobbles at 130 and 200 Hz is the right channel with an unavoidable room effect. At 4 cycles it is worse than the cycles before it. I show 4 cycles (REW chooses the smoothing here) to show the first wave front arriving at the listening position.
Good to see the little changes in pré EQ help to get proper summation at higher frequencies. The sum clearly follows the curve of the Left and right signal now.
You can also see the trade I make in lower frequencies where in the left channel I have a problem at 70 Hz (room mode) and compensate for that with the right channel. The sum is nearly flat.
Here's the 4 cycles gated left plus right sum and it's phase...
Pretty self explanatory I'd say.
If we look at it in a Wavelet:
A 40 ms Wavelet to show the first wave front at the listening position. It clearly is in a somewhat life room (partly treated living room) but all in all I'm not unsatisfied with it. Pretty straight line for the highest SPL level. The low frequencies aren't done yet so it's normal that isn't straight. I can't filter 4 cycles here.
If I do that with the group delay (4 cycles gate) and 1/3 smoothing it looks like this:
Obvious by looking at the phase plot but still...
I'd say goal pretty much reached in creating a time coherent arrival at the listening position! Best part of coarse is it sounds better and better with each little improvement I manage to make.
Sometimes it takes a while to find a bit extra but as long as I'm able to think up ways I'll go on.
Last edited:
Awesome data at listening position even serious headphones if they in reality is smooth gets probably distanced in phase timing thanks sharing.
As Freddi once reminded me of this quote by Howard Beale who said, "I have seen the face of God..."
It doesn't get any better than this.
Congratulations Wesayso! What are you going to do with all your free time now that the Towers have reached their pinnacle?
Oh, and don't forget, there is zero group delay!!! Amazing! I can't even fathom what this sounds like. Pure heaven.
p.s., where do you get this "Wavelet" capability in REW? Is that the same as the spectrogram? Doesn't look the same.
Last edited:
Cool! I never knew REW did wavelets - this software keeps getting better and better - always there unbeknownst to me 🙂
More than mighty clean 😀 I/O soundcard loopback makes wonder, had it been real room think room wasn't there anymore cause completely filled with rockwool and or glasswool.
As Freddi once reminded me of this quote by Howard Beale who said, "I have seen the face of God..."
It doesn't get any better than this.
Congratulations Wesayso! What are you going to do with all your free time now that the Towers have reached their pinnacle?
Oh, and don't forget, there is zero group delay!!! Amazing! I can't even fathom what this sounds like. Pure heaven.
p.s., where do you get this "Wavelet" capability in REW? Is that the same as the spectrogram? Doesn't look the same.
Though I am quite pleased so far with what I've managed to accomplish in a relatively normal living room I'm not about to stop now! 😀
I'd still like to find a way to optimise the room as I expect my ceiling to be a big barrier in achieving more. I'll run some experiments to find out.
I also want to test ambient speakers playing delayed band limited queues from behind. Not sure how to yet but I think I have the perfect drivers for that test 😉.
I don't think I would have gotten this far with different type of speakers. The dynamics are awesome, imaging is very good and the coherence is THE main player here. Still have some voicing to do, that's the hardest part with all the different source material out there.
To ArtsyAllen I'd like to say: it can be done with full range arrays. One could win something with a tweeter line but I'm not prepared to loose the coherency. After a lot of work and tests a little mid-side processing gave me the phantom center sound I was looking for. I'm getting quite close to the sound I loved from the XT25 ring radiator tweeters and in time I'm sure I'll nail it. Hope you're still watching, I'll give you a push on your own thread.
My understanding is that vertical line arrays, floor to ceiling, effectively not only cancel the effects of floor and ceiling bounce, but actually use that in a way that turns them into an advantage. Roger Russel explains how this works on his website. Corners are what I would still want to dampen. If I were to add tweeters to what you have, I'd make sure the crossover was well above 6kHZ, so the coherency would not be damaged too much.