No, it can't be interpreted in a contradictory way. It might appear so for non-native speakers of English but there's really no doubt about what he said.
Elias obviously feels attacked by the various comments that his perception of phantom sources might differ from the majority of all listeners. Now we have to wade through ironic and sarcastic posts without ever progressing to a factual discourse.
There's also no doubt that there can be a significant difference between sounds from a single speaker (sound mixed solely to the left or right speaker) and phantom sounds. It depends on the room and the speaker's directivity.
Not completely. I'd say the tight directivity control of the speakers , which minimizes the very early room refelctions is a key factor as well.
"More or less similar speaker(s)" may not be quite enough. A Ferrari and a Ford Fiesta are "more or less similar" cars, but they differ considerably in performance.
A very narrow sweatspot to me means a none directivity controlled speaker in a fairly dead room.
This cannot be the goal of accuracy because the perception of the producer/engineer is not defined !
Should I become a producer/engineer and start mixing tracks by listening the two tweeters and not hearing high freq phantom images??
Then, customers who bought my records would also be listening to the two tweeters, and not hearing high freq phantom images ! Nothing more, nothing less. 
If I, as a producer/engineer mixing records, am hearing two tweeters, anyone who is not hearing the two tweeters exactly in the similar fashion but might perceive high freq phantom images is not capable of accurate reproduction because he is not capable of "hearing like the producer/engineer heard it at the production stage" !!!
- Elias
You know what speakers I listen to in my room and still I do experience a very narrow sweet spot. That's why I'm very interested in seeing measurements of your room. They might contain other possible explanations what causes a larger sweet spot.
Your analogie of a Ferrari and a Ford Fiesta put a smile on my face
. My Ferrari is set up in a relatively live room, but apart from the floor refection (at about -17 dB) there are no significant discrete reflections. RT (yes, I am well aware it is a difficult concept in small rooms) is relatively flat with frequency, at about 0.4 ms. The room is almost 100 m^3.
All in all not too different from how you set up your Fiesta, right
?
Excuse me but it is soon to be 2012 ! Making steady state cancellation measurements sounds so 70's
Measure room impulse response at the listening position and do wavelet analysis !
Anyway, in room direct sound can only be separated from the reflections within the Heisenbergs time-freq limit.
- Elias
All in all not too different from how you set up your Fiesta, right
> insert "laughing hysterically" smiley face here <
P.S. Earl prefers Honda.
Last edited:
I suggest center channel-derivation for Elias
I suggest that he comes up with a speaker that solves his/the problem.
Quite right. "The proof of the pudding is in the eating".
I've listened to some of those "magic" concepts that promise to overcome the shortcomings of 2 speaker stereo (namely lack of spaciousness) but they all created more problems than they had solved. Maybe Elias will finally crack it?
maybe this topic would work
How many angels can dance on the head of a pin? - Wikipedia, the free encyclopedia
John L.
How many angels can dance on the head of a pin? - Wikipedia, the free encyclopedia
John L.
I'm with Markus too. Ergo, I'm not even convinced high-directivity speakers sound less spacious per se. In my experience a high-directivity speaker in a room with medium to slightly high reverberation, but with very early reflections damped, a good recording with reverberation and/or reflections can sound more spacious than with a wider-dispersion speaker in a room where VER do occur. However, this is quite recording-dependent. Dry recordings will sound drier than with wide-dispersion + VER.
The biggest disadvantage of my setup in my opinion is the smaller than normal sweet-spot. A second disadvantage is (as already mentioned by A_tewinkel) that the sweet-spot dip around 2 khz may be more noticeable (see Toole (9.1.3). For now, I can live with those two shortcomings - I've picked my poison.
The biggest disadvantage of my setup in my opinion is the smaller than normal sweet-spot. A second disadvantage is (as already mentioned by A_tewinkel) that the sweet-spot dip around 2 khz may be more noticeable (see Toole (9.1.3). For now, I can live with those two shortcomings - I've picked my poison.
In your playing around have you tried evaluating normal stereo recordings reproduced via Dolby Pro Logic I/II ? (Or other non-reverberation introducing multichannel modes, not the gimmicky stuff like "hall" mode etc)
Does your receiver have a mode where derivation of rear channels can be switched off, so that only a centre is being derived ? (And by that I mean not just turning off the rear channels, but that the matrix is changed so their sound components are not "removed" from the front channel mix)
If so I wonder what your impressions are ? In theory a steered matrix system shouldn't work too well on music even in something as simple as deriving a centre channel, but I think some of the newer systems are quite clever in this regard and may do a better job than we give credit for.
Exploring wider speaker separation with a derived centre channel is certainly something I intend to explore when the living space, money, and spare time align at the same time, and I think it all comes down to finding the best algorithm. (And so far in this thread we've seen at least 3 different ones referenced in various papers)
For experimentation what would be great is if there exists some PC based software that could take a two channel recording (line in or local source such as iTunes) and process it to a 3 channel L+C+R output, incorporating several different switchable algorithms so they could be directly compared.
Does anyone know whether any of the algorithms highlighted in research papers are available in PC software based DSP form ?
There may also be some hardware implementations - my PC has a Creative Audigy 2ZS which has an algorithm called CMSS 3D which can upmix 2 channel audio to multiple channels in interesting ways, but I don't know which algorithm it uses, nor do I have a way to directly compare it to other algorithms.
Likewise many receivers have different algorithms but other than Dolby Prologic and a couple of others, we really don't know what algorithm they might be implementing.
The ideal would be to find a fully software based DSP which can implement any of the different matrix schemes (using a multichannel sound card) so that some direct comparisons can be made.
Last edited:
Lack of spaciousness is only one of the most serious problems of conventional stereo triangle speakers. One other serious problem being the localisation of the speakers at high freqs.
I've have introduced several concepts in other threads to try to overcome these problems.
The basis to move beyond these problems appears to be looking conventional constant values with open mind. Is there anything constant in human sound localisation in amplitude, frequency or time domain ? No. Then why not take advantage of it and introduce speaker characteristics that helps the perception instead sticking himself in the old fashioned constant specifications (constant directivity, constant freq response etc.).
I think once after the speakers are made unlocalisable in any condition and some spacious in generated, the stereo records can be quite enjoyable.
But, I have not been able to achieve this with conventional stereo triangle. Thus the extraordinaties.
- Elias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.