A thought I would like to discuss:
In general:
Low frequency is omni- directional.
Flush mounted drivers are 2 Pi radiators until they are acoustically large.
Waveguides provide a pre-defined angle of radiation. Constant Directivity (CD).
Most seem to agree that CD is desirable. Direkt sound and reflections have identical frequency response. However, low frequency will always be omni-directional. This creates a conflict.
There is usually an abrupt change in radiation angle around the frequency where the driver/waveguide/baffle turns acustically small.
How would this scenario compare to an approach where the radiation pattern gradually increases towards omni-directional like this for example:
https://www.diyaudio.com/community/...he-easy-way-ath4.338806/page-562#post-7225299
Assume point source behavior of all frequencies. It is also free to choose radiation angle of the very high frequencies.
Please share your thoughts.
In general:
Low frequency is omni- directional.
Flush mounted drivers are 2 Pi radiators until they are acoustically large.
Waveguides provide a pre-defined angle of radiation. Constant Directivity (CD).
Most seem to agree that CD is desirable. Direkt sound and reflections have identical frequency response. However, low frequency will always be omni-directional. This creates a conflict.
There is usually an abrupt change in radiation angle around the frequency where the driver/waveguide/baffle turns acustically small.
How would this scenario compare to an approach where the radiation pattern gradually increases towards omni-directional like this for example:
https://www.diyaudio.com/community/...he-easy-way-ath4.338806/page-562#post-7225299
Assume point source behavior of all frequencies. It is also free to choose radiation angle of the very high frequencies.
Please share your thoughts.
Last edited:
Why is this? It takes a certain type of waveguide to do this.
A continually expanding horn needs to be larger to hold the lower frequencies as narrow.
To me and most folks in the PA world, DI means Direct Injection or Direct Input, as in 'DI Box'. What is your usage, just to clarify.
AllenB, I thought of a full range point source to not complicate the discussion with different sources of sound and their interference.
Mainly to understand how CD-radiation above a specific frequency and omni below sounds compared to equal radiation angle (as CD) at very high frequencies and then gradually increased radiation angle towards 4 Pi as frequency drops.
MrKlinky,
Thanks for pointing that out. Should have written CD as constant directivity. Had directivity index in my mind at the moment. Post updated.
Mainly to understand how CD-radiation above a specific frequency and omni below sounds compared to equal radiation angle (as CD) at very high frequencies and then gradually increased radiation angle towards 4 Pi as frequency drops.
MrKlinky,
Thanks for pointing that out. Should have written CD as constant directivity. Had directivity index in my mind at the moment. Post updated.
Good question Petter Persson, you are at core of things! It is question how wavelength relates to physical world, rooms, loudspeakers and observers. Your question ties loudpeaker to acoustics and psychoacoustics. Also it ties loudspeakers to set of compromises due to wild difference in wavelength between low and high frequencies. Sound power directivity index (DI) of a loudspeaker is kind of summary of it, omni low frequencies have low DI and high frequencies usually have higher DI and its all due to wavelength, size and shape of transducers and over all construct of the loudspeaker affect what the relation is.
Premise is that we'd like direct sound with flat frequency response in order to hear records without much coloration, right? Sound power shows how much sound is radiated to other (all) directions. If loudspeaker is in a room and we assume boundaries of the room reflect sound ideally then sound power is good estimation of average nugget of sound that hits an ear. Direct sound is just one tiny slice of sound radiated from loudspeaker. How all of it is perveiced depends on how hearing system works.
Lets put technical stuff aside for a second and think philosophically: what is the goal of loudspeakers? of course very good sound, right? Very good sound is defined by you, pressure variation in room reaching ears (body) is perveiced by your brain such way that is pleasing to you. Without knowing much details about anything one could start reasoning how to put sound to room so that it is perceived with maximum pleasure? If its is domestic setting acoustics of a room are probably not optimal, positioning of speaker(s) and observer(s) might need to be practical instead of ideal and so on. Before even getting into DI we see context varies so probably various DI would work on various situations. To be safe one could draw conclusion that when DI is "flat", at least smooth, which indicates average sound of the speaker emitted to any direction could be described "flat" then the sound is as good as it can be to any direction. This means any observer anywhere in the room would get roughly similar "flat" frequency response. Sounds good right? Yeah, but reaching such situation is not necessarily easy or cheap or practical or mandatory. If there is only one observer and positioning of things can be changed and room acoustics can be improved then importance of sound power reduces and importance of "direct sound" increases.
So, basically yeah one could claim constant directivity is always desirable but one could get enjoyable sound just by manipulating positioning and / or acoustics even if DI wasn't constant ( or otherwise ideal ). I think it is more practical to come up with rising DI system than constant DI. Easiest wide bandwidth constant DI is probably OB speaker with enough ways.
Premise is that we'd like direct sound with flat frequency response in order to hear records without much coloration, right? Sound power shows how much sound is radiated to other (all) directions. If loudspeaker is in a room and we assume boundaries of the room reflect sound ideally then sound power is good estimation of average nugget of sound that hits an ear. Direct sound is just one tiny slice of sound radiated from loudspeaker. How all of it is perveiced depends on how hearing system works.
Lets put technical stuff aside for a second and think philosophically: what is the goal of loudspeakers? of course very good sound, right? Very good sound is defined by you, pressure variation in room reaching ears (body) is perveiced by your brain such way that is pleasing to you. Without knowing much details about anything one could start reasoning how to put sound to room so that it is perceived with maximum pleasure? If its is domestic setting acoustics of a room are probably not optimal, positioning of speaker(s) and observer(s) might need to be practical instead of ideal and so on. Before even getting into DI we see context varies so probably various DI would work on various situations. To be safe one could draw conclusion that when DI is "flat", at least smooth, which indicates average sound of the speaker emitted to any direction could be described "flat" then the sound is as good as it can be to any direction. This means any observer anywhere in the room would get roughly similar "flat" frequency response. Sounds good right? Yeah, but reaching such situation is not necessarily easy or cheap or practical or mandatory. If there is only one observer and positioning of things can be changed and room acoustics can be improved then importance of sound power reduces and importance of "direct sound" increases.
So, basically yeah one could claim constant directivity is always desirable but one could get enjoyable sound just by manipulating positioning and / or acoustics even if DI wasn't constant ( or otherwise ideal ). I think it is more practical to come up with rising DI system than constant DI. Easiest wide bandwidth constant DI is probably OB speaker with enough ways.
Last edited:
I wrote "In general there is only one "flaw" with this speaker and it is its tendency to create a "loge" or opera box effect. Its like a delicate doll house with all the details and relations perfectly intact but it is still a doll house. It's small. This effect seem to come from the narrow directivity of the WG. It may be constant but it is rather narrow. So these speakers can not re-create the full size and depth experience of a 10th row center seat in a concert hall. Shade.
Recordings with dry and tight ambience integrate better as this is how my room sound. Recordings from big venues with a lot of acoustics is revealed immediately in two ways - the ambience is reproduce really fine with lot of details and dynamics - but it is like listen from an opera box - its limited in size and the border becomes really obvious as the acoustics in my room is so totally different."
here: https://www.diyaudio.com/community/threads/hornflower-2-way-point-source.386363/page-4#post-7133870
All in all it boils down to the "Stereo system" faults/lack of standard - the sound-mic-2-speaker-sound chains theoretical ability to really reproduce a live event. It cant work and it dosen't really. But we strive to do the best attempts from what we got.
Just now I'm in the state that wide dispersion speakers with the help of a decent room gets the best disappearing act and recreation of believable room but that the more directive speakers recreate a more correct wave and sound objects relations.
Its like: either a live size house with bent walls and putruding windows - but still a quite proper house or, a perfectly laid out doll-house where you cant get rid of the "frame".
Pick our poison.
My thinking around the question asked is that any abrupt change in a speakers characteristics (FR, distortion, phase, directivity) is probably not a good thing...
//
Recordings with dry and tight ambience integrate better as this is how my room sound. Recordings from big venues with a lot of acoustics is revealed immediately in two ways - the ambience is reproduce really fine with lot of details and dynamics - but it is like listen from an opera box - its limited in size and the border becomes really obvious as the acoustics in my room is so totally different."
here: https://www.diyaudio.com/community/threads/hornflower-2-way-point-source.386363/page-4#post-7133870
All in all it boils down to the "Stereo system" faults/lack of standard - the sound-mic-2-speaker-sound chains theoretical ability to really reproduce a live event. It cant work and it dosen't really. But we strive to do the best attempts from what we got.
Just now I'm in the state that wide dispersion speakers with the help of a decent room gets the best disappearing act and recreation of believable room but that the more directive speakers recreate a more correct wave and sound objects relations.
Its like: either a live size house with bent walls and putruding windows - but still a quite proper house or, a perfectly laid out doll-house where you cant get rid of the "frame".
Pick our poison.
My thinking around the question asked is that any abrupt change in a speakers characteristics (FR, distortion, phase, directivity) is probably not a good thing...
//
Hi TNT, yeah that is the generalization, we can say smooth DI is always better than the same general trend with aberrations.
Btw. go listen closer and the image widens, use toe-in. Start with 45 deg toe-in and go listen on axis, bam, no operabox. Problem with toe-in and narrow coverage speakers is that first lateral reflections are reduced, this increases clarity but also limits sound stage between the speakers. For this reason you need to be listening close enough so that the speakers are far enough apart so that perceived sound stage width seems natural, not too narrow, not too wide. If its too wide a hole appears on the center without separate center speaker. Move back and forth on the center normal between speakers to find spot you prefer and then figure out how to produce that on your practical listening position.
Also, don't limit yourself to single pair of speakers, even for stereo material. If you wanna remove the operabox / doll house just add another set of speakers (>=90 deg from listening position) to provide envelopment, but this time its fully controlled by you (DSP). You could add center speaker to stabilize phantom image if you will.
Anyway, it is some work to get "perfect sound", enough clarity and enough envelopment / spaciousness, just matter of time to align implementation to subjective preference.
If its not possible to adjust listening triangle to your speakers and you'd still grave for wider bigger sound then I think you need wider coverage speakers. Easiest is to make traditional direct radiating system.
Btw. go listen closer and the image widens, use toe-in. Start with 45 deg toe-in and go listen on axis, bam, no operabox. Problem with toe-in and narrow coverage speakers is that first lateral reflections are reduced, this increases clarity but also limits sound stage between the speakers. For this reason you need to be listening close enough so that the speakers are far enough apart so that perceived sound stage width seems natural, not too narrow, not too wide. If its too wide a hole appears on the center without separate center speaker. Move back and forth on the center normal between speakers to find spot you prefer and then figure out how to produce that on your practical listening position.
Also, don't limit yourself to single pair of speakers, even for stereo material. If you wanna remove the operabox / doll house just add another set of speakers (>=90 deg from listening position) to provide envelopment, but this time its fully controlled by you (DSP). You could add center speaker to stabilize phantom image if you will.
Anyway, it is some work to get "perfect sound", enough clarity and enough envelopment / spaciousness, just matter of time to align implementation to subjective preference.
If its not possible to adjust listening triangle to your speakers and you'd still grave for wider bigger sound then I think you need wider coverage speakers. Easiest is to make traditional direct radiating system.
Last edited:
In general I agree with you TNT, normal consumer living room size speakers (CD & non CD alike) are characterized by this delicate doll house like aspect, it’s what most audiophiles take as great sound. The difference in scale and dynamics is really akin to a portable radio vs a small hifi system but most folks have only been exposed to loud hifi vs truly dynamic sound representation.
But having heard really big CD/horn systems i know it’s possible to increase the size of that opera box to something that approximates the scale and impact of live music.
But having heard really big CD/horn systems i know it’s possible to increase the size of that opera box to something that approximates the scale and impact of live music.
I don't believe a wider pattern is the answer to get rid of the doll house effect. As it simply would enlarge anything and everything you play on it, also when it isn't in need of such an enlargement. Your miles may vary, but that's not what I would want.
I've been trough that with my own speakers, that, when left "untamed" they projected a huge soundscape. Fun and engaging at first, but it really is throwing the same sauce on everything. Take away the first reflections and essentially the sound is "tamed" and you'd get something resembling that doll house experience. You get way better placement in imaging, but the size and sheer feel and energy of 'space' is missing.
When you take away those first reflections, and do have later arriving reflections from lateral angles, plus you make sure to help out battle the cross talk you'll probably hear due to that reduced level of reflections, the imaging size tends to vary with the source material. Large when it needs to be large, small if it is in the recording. Look for the studies from David Griesinger for more info on how that works. In fact, I bet his Logic 7, now abandoned, could translate 2 channel music to a surround system using those principles.
At least that would be my 2 cents.
I would say that a smooth shaped Directivity Index is key whatever your preference is, be it a wide or more narrow pattern.
I've been trough that with my own speakers, that, when left "untamed" they projected a huge soundscape. Fun and engaging at first, but it really is throwing the same sauce on everything. Take away the first reflections and essentially the sound is "tamed" and you'd get something resembling that doll house experience. You get way better placement in imaging, but the size and sheer feel and energy of 'space' is missing.
When you take away those first reflections, and do have later arriving reflections from lateral angles, plus you make sure to help out battle the cross talk you'll probably hear due to that reduced level of reflections, the imaging size tends to vary with the source material. Large when it needs to be large, small if it is in the recording. Look for the studies from David Griesinger for more info on how that works. In fact, I bet his Logic 7, now abandoned, could translate 2 channel music to a surround system using those principles.
At least that would be my 2 cents.
I would say that a smooth shaped Directivity Index is key whatever your preference is, be it a wide or more narrow pattern.
The limitations of stereo are not a deal breaker. I would not compromise it with more spaciousness.
Without reflections, the room is no longer a limitation and the soundstage reaches as far as it wants beyond the room.
Without reflections, the room is no longer a limitation and the soundstage reaches as far as it wants beyond the room.
@wesayso, I agree that spraying the room with sound do of course set a stamp of sameness of everything - not desirable. And in here lies the problem.
I still think that a standard sound source (a clap?) should be recorded in a good sounding position in the venue and follow the recording as meta-data for the reproduction system to be utilised for (auto, per recording) linearisation and sound scape forming to be used in a "one-time" calibrated (vs the standard clap) reproduction system for true end2end corrected reproduction. Most probably, this system need to entail at least a center channel as @tmuikku indicates - perhaps even more.
//
Sorry for this blurrb - its OT actually 🙂
I still think that a standard sound source (a clap?) should be recorded in a good sounding position in the venue and follow the recording as meta-data for the reproduction system to be utilised for (auto, per recording) linearisation and sound scape forming to be used in a "one-time" calibrated (vs the standard clap) reproduction system for true end2end corrected reproduction. Most probably, this system need to entail at least a center channel as @tmuikku indicates - perhaps even more.
//
Sorry for this blurrb - its OT actually 🙂
Last edited:
Therein lies part of the problem? If it really is a problem. All my favorite music is locked up in the Stereo format. Making me want to get the most out of it. I can be perfectly happy with what Stereo has to offer, I merely created a hobby out of it to maximize that enjoyment 😉.
A surround system would have a lot more potential to portray an event in a more truthful or valid way. If the standards for recording of such an event are sufficiently regulated.
A surround system would have a lot more potential to portray an event in a more truthful or valid way. If the standards for recording of such an event are sufficiently regulated.
You are starting with an assumption/premise that is false. Not every speaker is a monopole, you know?
Really? Listening tests don't support it (e.g. studio speakers vs good home speakers in a room in a home rather than studio) nor does a consideration of how sound is radiated by musical instruments. What appears to be preferable is a smooth but rising directivity. How that directivity varies optimally is unfortunately a function of the room, the recording, the preference of the user and the expectation for the type of music (e.g. orchestral music vs studio rock). At the moment there seems to be a tendency to go for a straight line rising directivity (e.g. LS60, Revel,...) which I guess looks good on graphs in promotional material but I'm not sure it will preferred to a non-straight line if the wiggles are better matched to the room absorption.
Good topic though and one of the most important when it comes to achieving genuine high sound quality in the home.
Oh really?! Oh well, times change, so guessing 'Q' has replaced it.
Indeed! IME, some of the best 'stereo' reproduction has been in mono. 😉
Petter Persson - Listen to a planar speaker (i.e. Magnepan). Planars don't have the problem of omnidirectional lows and forward-firing highs. They give the impression of listening to real sounds rather than speakers.
A long time ago, some conventional speakers had both front and rear firing tweeters. I suspect that not enough listeners cared about being omnidirectional to keep the rear-firing tweeter from being cost-optimized out.
Ed
A long time ago, some conventional speakers had both front and rear firing tweeters. I suspect that not enough listeners cared about being omnidirectional to keep the rear-firing tweeter from being cost-optimized out.
Ed
Thanks for a good debate everyone.
That's why "In general" and is written in the first post, so we could have a value adding discussion. Also, "Assume point source behavior of all frequencies" was ment to keep the discussion to radiation/directivity patterns rather than interference between different drivers.
This is a good example of what I aimed to discuss. The JBL is a good example (though not a monopole) where constant directivity abruptly change to 2 pi, and then 4 pi. The Revel is no monopole either and has a different driver setup, baffle size (and so on) than the JBL. The results are intresting, but I would not go further to claim the constant directivity horn for the battle loss. I see it as the speaker as a whole was outperformed by the Revel.
I guess the only way to isolate radiation pattern effects is to build sets of large enough MEH similar to what mark100 and A for Ara have done...