Am i the only one who can perceive or have the experiences listening to tube amps and the singer is positioned on the ceiling? This includes the use of top concentric Tannoy and very expensive tube amp. Just like headphones, the speakers are on the left and right but the sound can be high above the head. This is for anyone who might be confused why two speaker in horizontal plane can produce vertical imaging.
You might forget the possibility that your ears/brain is harder to please than theirs. My 'parameter' to remove this subjectivity is the time one spends (in a day) listening to music. But then my friend used earbud to listen to music all day at work from his laptop
In a vertical plane, the tweeter is positioned slightly forward than the woofer (the sound is originated close to the position of the voice coil). To make them time coherent, we can make a slanted baffle or tweeter at the bottom, depends on the position of the ears (e.g. with listener sitting on the sofa) and how much phase change has been made by crossover.
I have wished people like PMA is more interested to investigate this than the... you know what.
No, that came long after.
At that time, you were doing the Symmetry 4 quadrant power amp prototype, and it had 10 times more THD than the breadboard version. We both scratched our heads for the longest time, and, on the last hour of my last day, I said: "Sorry, my plane is gonna leave, we have to call it a day". I reached out to turn off the Sound Technology, and my hand accidentally moved the probe by an inch, and the THD dropped 10 times! It was repeatable, and we spent a while moving the probe back and forth and laughing. I called you later from the airport, and you said you found a polluted trace on the PCB!
Last edited:
@MaxHeadroom
Dan, maybe the last remark to your question if the impulses in the groups were same. Attached is a record of all 3 groups of 3 impulses in persistence mode, that means all shown at the same screen. Both electrical impulses and JVC headphone responses. We can see that both input impulses overlap and responses overlap and there are no differences between the impulses. This shows that all the perceived audible "differences" are a result of brain evaluation of the ear signal. As we know, this evaluation is time dependent, so the impulses in the train of impulses are perceived differently, even if they are same. What we hear does not necessarily means that it exists, it exists in our mind only.
Dan, maybe the last remark to your question if the impulses in the groups were same. Attached is a record of all 3 groups of 3 impulses in persistence mode, that means all shown at the same screen. Both electrical impulses and JVC headphone responses. We can see that both input impulses overlap and responses overlap and there are no differences between the impulses. This shows that all the perceived audible "differences" are a result of brain evaluation of the ear signal. As we know, this evaluation is time dependent, so the impulses in the train of impulses are perceived differently, even if they are same. What we hear does not necessarily means that it exists, it exists in our mind only.
Attachments
What I have perceived with 2-speaker stereo is the elevation of center mono phantom sources, a well known effekt and easy to explain with HRTF: we receive a quadruplet of impulse reponse with +30° and -30° angles of incidence, from each speaker to each ear. The brain is forced to interpret this as if it were a 0° impulse from a single source. The comb-filter type of effect from the summing has patterns that have some resemblance to the HRTF we would have from an elevated center mono real source, and that's why the brain is fooled to hear mono center images elevated. How strong this effect is will depend on many variables, personal HRTF's dominating, but also any reflections and intrinsic characteristics of the speaker. I've found phase response to slightly change the elevation effect, for example.
For the same reason mono phantom center sources never have the same timbre that hard L/R panned signals do, a problem the better mixing engineers are aware of and EQ their signals differently depending on the panning, and if the positioning is also made with time offsets yet another set of effects on timbre and perceived elevation kick in, depending on the apparent source position.
How strong the systematic mono phantom center elevation and timbre shift effect really is can be judged with a trinaural playback system which has a center speaker added and the L/R speaker are at 45°. The feeds to the speakers are derived by simple static matrix operations which result in a mono source being 6dB louder on the center than on the side speakers which makes the center speaker almost as dominant as if if it were the only speaker and the artifical elevation is completely gone.
Trinaural has the advantage that it swaps the roles of phantom and real sources and a way better rendering of the sound field around the head that is more immune to head position and rotation, plus it has better distinction between close-up focussed and diffuse ambient images. It "decouples" the speakers from the listening impression much better than two speakers do. This is because no speaker is ever off, all three are always outputting signal (exeption being exactly half-panned signals which emanate from only two speakers). Only disadvantage is that HRTF encoded stuff like Q-sound does not work ver well (by design), plus the need for a 3rd speaker (which must be identical to the L/R speakers, of course).
Last edited:
Thank you for an outstanding post.
I am building what I now realise is a trinaural horn speaker system. My thinking is to have two small horns either side of a central horn to handle upper frequencies in stereo, while the larger central horn is summed to mono for lower minds and general bass duties. As a restless listener, my aim is not accuracy but musicality.
You say that all three speakers in a trinaural system need to be of the same size. I understand this, and shall keep it in mind during testing when my build is complete.
And that's exactly what is all about hifi.
Who cares about what happens, in a transistor or a wire, out of the gear designer ?
All this strange process, to try to capture sound pressures variations in the air to electric signals, transform those signals in coded pulses and sample them, transform back those pulses in an analog electric signal strong enough to move a speaker somewhere else, to transform back those electric signals to sound pressures variations, is just about this. To create in our minds the illusion that some musicians are playing in our living rooms.
We could imagine lot of other methods to do this. And, with the technology we use, for the moment, the goal is not linearity or low distortion. The goal is to make this illusion as credible as possible.
Just imagine that we are able to capture the stimulus that our nerves transmit to our brain, record-it, and excite the nerves of the listener the same way ?
We would have a perfect system, with no local acoustic problems etc.
And lose, in the same time, all the creativity that allow our actual technology: Musical instruments that do not exists (Synthetisers, vocoders, electric guitars sounds, etc.), artificial reverberations that create rooms acoustics that do not exists neither, musicians that seems to play together, while they never met them in the same spacetime ...
Ps: Good and interesting input, KSTR. We could talk about the advantages and disadvantages of the 5+1 and the need for our brains to be accustomed to this kind of system, that do not work so good for music at this time for various reasons.
Last edited:
Everyone who is interested in communicating constructively about what is physically happening in a commonly understood language.
The actual dominant requirement is (close to) identical phase responses. You can use a different speaker for the center channel when it matches in phase.. or when you force it to match either with DSP phase linearisation applied to each speaker (preferred) or with additional allpass functions to add the phase response of the center to the others and vice versa, which of course increases the total phase response significantly -- but which then can be rectified by an overall DSP phase correction on the source material.
If the phase responses don't match, the center image stays there (unlike to normal stereo playback) but off-center sources become unstable, actually "flipping sides" with frequency when the offset reaches 180deg.
I mentioned my several weeks lasting experience in Supraphon recording studios for classical music. This happened in 1977. Until that time, I was a sort of a "HiFi audiophile". After the experience with the recording of the classical music, I started to be absolutely skeptical about home "hifi" compared to what I seen and heard in studios and stopped any attempts to change or improve my home audio system, considering it useless. This lasted for another 25 years until I started to be interested in hifi audio again. But quite soon I have realized that the situation has not changed much and I returned to a rational engineering approach and trying to understand more of the human psychoacoustics.
Gee whizz, this is gold dust to me. Eventually, when financial constraints allow, I will will invest in something like the miniDSP 2x4 HD and a calibrated mic, but for now I will be using a combination of both mono and stereo guitar pedals to equalise and delay delay the sound. Believe me, there is a lot of mileage to be had from the audio use of guitar pedals. They are not just for guitars.
I do understand the perspective of using measurement, but I also understand the value of mapping out the psychoacoustics of playback by ear alone. To combine both is something really special to look forward to.
Again, I thank you.
I don't understand this. Please, elaborate ?
On my side, i'm absolutely not an "audiophile". My home system was build in the spirit of big studios monitors, in the unique attempt to can have, at home, a system more neutral, well known by me, in order to have a different point of view of my studio work.
Multiplying the listening systems is the only way to ensure that we don't had made obvious mistakes on the tonal balance of a mix, due to some bad or special characteristic of the studio monitors. We depend on them.
What is a studio monitor ? The contrary of most audiophiles speakers.
High efficiency, often with horns because it is the only way to get high listening levels and dynamic. Horns are often a problem. During decades they were bad designed. Full of high Q resonnances, that produced this kind of nasal sound. In fact it is not correlated with the horn principle, but its bad design and bad implantation.
The second reason is it is the only way i know to have the same emissive surface at the Xover, IE the same directivity. The third reason is it allows to minimise the ways, (only two) that brings, with the crossover phases turns & a lot of problems of consistency.
That was the reason why I worked with a friend, a clever and talented speaker guy, that had the idea to design a horn, no more based on flat waves propagation, but spherical waves. It was before the Le c'learch work.
We made-it from huge and massive pieces of wood to avoid any internal resonnances.
His horns are fantastic. More natural and fluid than any membranes, with all the advantages of reduced power on the side (less room's resonnances) constant directivity in their angles of projections etc. It was a hard job to find a very good driver for them, but, after a lot of tries and measurements of various drivers, we found one working not that bad after passive correction of their response curve.
An other thing to mention is the compression in the driver, and the charge of the horn, make amortissement of the resonances of the membrane , excited by transients are lot better. THis reveal a lot of litle details and micro dynamic, ordinary lost in a kind of fog.
For the first time, I had a piano at home, recorded flat, that sounded like... a Piano. And what i believe was a reference I could believe in.
The drawback is, now, I cannot afford most of the hifi speakers. hence my disappointment with those KEF LS50 i recently bought.
The second drawback is ,like said J.C. about his planars, very few records I love are perfect, because those speakers have no mercy. And, on the long run, dynamic is bringing fatigue: too much details. It is like real life concerts, more than two hours, you die. So, often, I prefer, like John, to listen to a little poor system, like my PC one, for casual listenings. just enjoy the music easy way. Sometimes, better is the enemy of good ?
Last edited:
We like to say that we want our systems to sound like live music, but in fact many would be very disappointed if it did. I have heard good orchestras in very good halls, and it was virtually impossible to pick out the position of any particular instrument with one's eyes closed. But that is the illusion most sought after by audiophiles who go on and on about resolution, detail, and "sound stage". Sadly, reality just isn't audiophile grade!
Hi Pavel.
I have revisited the test and discovered I had reset a SRC enable function in VLC Player so was running at 44k.
On second listen with dac definitely running at 96k and knowing what to listen for ..........I found several spots around the central listening spot and around the room where I had good central apparent sound source for the first triple and Left skewed and Right skewed pulses for the second and third sets respectively.
I also found spots where I had good central source and could not tell set 2 from set 3 and other spots where set 2 and set 3 apparent positions were reversed.
These results were at standing height or half standing, and when sitting with ears at tweeter height I found it harder to discern direction skew but sets 2 and 3 definitely sound longer and smeared compared to set 1 where the pulses sounded 'crisper' and 'shorter' in all listening positions.
These are all very fine differences and moving the head sub-mm is plenty enough to mask or reverse findings.
If the pulses were repeated with less delay and longer pulse sequence I think the direction discrimination would improve.
Conclusion, These short and intermittent (stacatto) pulses and only three repetition sets are not quite enough information for the brain to reliably filter ambient/room reverb sounds and then reliably discern fine source direction changes.
If you want to put more sequences up I can take another listen.
Dan.
Edit Re the 22k filtered file, I found the LPF filtered file sounded 'brighter' and 'wrong' in the manner that HC filters can cause a slight high mids lift before roll out (and rumble filters can cause a slight bump in mid bass).
Last edited:
A good friend of mine uses very high end equipment to play vinyl records through headphones. Expensive record cleaning machine, expensive turntable, expensive cartridge, expensive preamp, expensive headphone amp, expensive headphones. He's very pleased with the experience, likes it much better than loudspeakers.
Good post, nezbleu. Live music and at home reproduced music are two different disciplines. However, either we want to sound it as close to original as technical means enable, or we ask for a " pleasing" sound. That's the key of all misunderstandings here. And I am quite sure it depends on music we listen to. Like Geddes x Linkwitz.
- Status
- Not open for further replies.