On the other hand, if the head is tilted on either side there will be ITD from the speaker on the floor ! That can perceive it's location too, even it's emitting only low freqs.
Tilting can be very subtle, and happen subconsciously. But we know that ITD resolution is very good, so even small tilting maybe too much.
How to avoid this is a good question ?
Tilting can be very subtle, and happen subconsciously. But we know that ITD resolution is very good, so even small tilting maybe too much.
How to avoid this is a good question ?
well, under reverberant conditions spatial hearing gives special weighting to different localization cues
the ongoing (phase) ITD that is the most potent localization cue under anechoic conditions seems to be the least important in such situation because it can be ambiguous and misleading due to multiple reflections:
most probably the ongoing (phase) ITD cues are disregarded by our hearing under reverberant conditions in the presence of other localisation cues
on the other hand the localisation cue that is present and reliable under reverberant conditions is the onset (first wavefront arrival) ITD
and it seems (see my argument earlier in this thread) that onset ITD works in the range of 0.4<4.0 kHz - the frequency range critical for wavefront shape determination
so on to the question:
how about keeping the crossover point for a floor coupled woofer below 400 Hz - just like Roy Allison did in His designs?
I did some additional tests with the 10" dipole woofer close to floor. Same configuration as before.
I put 8th order LR low pass filter (48dB/oct) and varied the corner freq.
Test signal was pink noise.
Here are the results:
1.5kHz low pass: Less treble can be heard from the woofer than with 24dB filter. It's hard to say the height of the image. It seems to be rather sensitive to head side tilting. Tilting head 30 degrees to either side and image is on the floor.
1kHz low pass: The image seems to be at the ear height although it is not very sharp. Tilting of head sideways shifts image close to floor.
700Hz low pass: The image is clearly at the ear height and 1 or 2 meters behind the speaker. Tilting head side ways causes image shifting downwards.
400Hz low pass: Same as 700Hz, image at ear height and head sideways tilting pulls down the height a bit.
What can be said as a conclusion? Clearly below 1kHz the image stays at the ear height if the head is straight upright. Head position seems to important however also below 1kHz, looks like ITD comes into play there.
- Elias
I put 8th order LR low pass filter (48dB/oct) and varied the corner freq.
Test signal was pink noise.
Here are the results:
1.5kHz low pass: Less treble can be heard from the woofer than with 24dB filter. It's hard to say the height of the image. It seems to be rather sensitive to head side tilting. Tilting head 30 degrees to either side and image is on the floor.
1kHz low pass: The image seems to be at the ear height although it is not very sharp. Tilting of head sideways shifts image close to floor.
700Hz low pass: The image is clearly at the ear height and 1 or 2 meters behind the speaker. Tilting head side ways causes image shifting downwards.
400Hz low pass: Same as 700Hz, image at ear height and head sideways tilting pulls down the height a bit.
What can be said as a conclusion? Clearly below 1kHz the image stays at the ear height if the head is straight upright. Head position seems to important however also below 1kHz, looks like ITD comes into play there.
- Elias
Last edited:
how about keeping the crossover point for a floor coupled woofer below 400 Hz - just like Roy Allison did in His designs?
According my test 400Hz may be still be perceived to the floor depending on the head orientation.
If we settle the floor woofer at 400Hz max, the more severe problem is what to do above 400Hz ?
Any element emitting strong direct sound to the listening position and placed lower than ear height will very likely perceived as being low too. Not very good. The only remedy may be the typical music spectrum which is fortunately very wide containing treble which can be used to nail the image at ear height if the tweeter is placed at ear height.
- Elias
Had to "wade-through" a lot of posts to get to this one..
The poorer localization is a result of some loss in direct sound pressure, and particularly the break in the pressure gradient.
You can here this effect at 40 seconds in and better at 1.16:
JBL 4428 compact 3 way studio monitor - YouTube
There may very well be a certain bandwidth where doing this is fundamentally more correct.
At lower freq.s a break in the pressure gradient also effects localization (..though you can still have horrible/high intensity competing modes that wreck even an unbroken gradient in a typical room setting.)
I think it's likely he is "locking in" on the polar behavior of the 10" driver. The intensity at 10 degrees relative to his listening position puts an emphasis downward.
Elias should try the woofer center aimed directly at his ears and then adjust the vertical "tilt" in each direction to try and hear an effect.
I think it's likely he is "locking in" on the polar behavior of the 10" driver. The intensity at 10 degrees relative to his listening position puts an emphasis downward.
Elias should try the woofer center aimed directly at his ears and then adjust the vertical "tilt" in each direction to try and hear an effect.
The question is do we detect vertical pressure gradient below 1kHz ? I'm not sure about the answer.
The speaker was placed directly ahead so it's in the median plane.
I woulnd't count on that as a principle as the sufficient presence of the treble frequencies is dependent on the program material, see graphs I posted above
in fact musical "treble" start well below 1500 Hz
thank You Elias - one of those test that needed to be done
although pink noise is a specific test signal, perhaps You could try also pure tones and filtered clicks
also tilting the head by 30 degrees sideways is a bit too much - it is not what a normal human does when tracking a sound source unconsciously
in real life it's more like 5-15 degrees
couple of interesting things:
-ultimately it is not HRTF from high frequency leakage that causes pulling the image down
-the importance critical range 0.4<4.0 kHz was confirmed in the sense that the effect of tilting was frequency dependent and with 400 Hz low pass the effect was minor: "pulling down the height a bit"
- and the above means that it was the onset ITD that was responsible
furthermore probably it is not the onset ITD from "from the speaker on the floor" (ITD of the direct sound) - after all when one tilts his head sideways both ears remain in the same plane perpendicular to the direction of the speaker in front - consequently there is no change in the distance of the ear from the speaker as a source of the direct sound
OTOH tilting sideways results in an obviously detectable change of the ITD of the sound reflected off the floor
If we settle the floor woofer at 400Hz max, the more severe problem is what to do above 400Hz ?
yes, therefore Roy Allison placed midrange and tweeter units in His designs front firing at the ear level, consequently losing advantages of the upfiring confirugation
but it is not the only solution, for instance no problem in placing the midtweeter element up-firing at the ear level - or perhaps even lower - depending in the directivity of the element, because:
- HRTF is not a problem as long as there is an optimal direct-to-ceiling reflected sound ratio in the HRTF range
- we can use the directivity of the midtweeter element facing up on a baffle to keep the >400 Hz range away from the floor - and I believe that Your test confirmed the critical importance of the first floor reflection
in the light of the foregoing it's no suprise that in an extreme case even a big and highly directive coaxial placed on its magnet directly on the floor can produce a realistic soundstage without any pulling of the image to the floor:
see Radugazon's: Experience 2
if we place a speaker close to the floor and facing up then even the vertical geometry of the system works for us - the direction of the first floor reflection hits well below the ear level
and it is no suprise that You never experienced Yourself any "pulling of the image to the floor" in You own flooder tests
Last edited:
I did try that. Higher angles aims the high freq sounds more to the listener ears. This was bad because it helped the localisation of the woofer, especially without the tweeter.
Higher angles 45 deg and beyond there start to loose bass due to dipole back wave cancellation.
I also tried vertical 0 deg tilt. This was actually hardest to localise. Now the high freq beam goes to listener feets.
In a free-field condition - assuming the source is directed at you and has its pressure gradient unbroken, yes.
Try it out in a field sometime. Spin around with your eyes closed and then trigger the sound - and look to see if you are pointed in the right direction. Then try it with a heavy isolation foam ear-plug in one ear.
Of course this is all in relation to a synthetic source (..mono). Stereo naturally creates a competing source - though hopefully the channel balance should prevail to give you a new "source" (..or "sources"), but I suspect that it's quality is conditioned on the speaker pair's spread relative to each speaker, and the listener. It's not something I've really tried however (..where I was listening only to bass reproduction).
Last edited:
It is due to HRTF in a sense that ITD is formed by it, although it is not a high freq phenomena.
I don't think there is a floor "reflected sound" below 400Hz in this configuration.
The localisation is more likely due to the fact that head is not a sphere and neck is not connected in the middle bottom of it, nor the two ears are placed at the side spots of it but little behind. This results that with head sideways tilt, there will be change in ITD from the speaker itself.
- Elias
Yeah, what could be more enjoyable on a warm summer day, spinning around on a flower field, with your eyes closed... birds are singing... sun is warming your skin...
... And listening to pink noise 
I don't think there is either. 400 Hz is almost 3 feet in length, it's almost certainly a bounded condition.
-now diffraction and increase in intensity at the floor..
No, most likely he placed the midrange and tweeter to (almost) ear level because if they are at the floor level too the image stays allways on the floor
Note that his speakers are not flooders.
Yeah, what could be more enjoyable on a warm summer day, spinning around on a flower field, with your eyes closed... birds are singing... sun is warming your skin...
Actually I was thinking of a real sound (even if reproduced).