Hi Tom, that's a long time that I use your records that are of great help for comparing different speakers or filters settings. What I like in the Harley is the little plane that flies over at 35 seconds. With headphones it's in the head, with speakers it's in the sky, approx 40° elevation, that's much better.
Are you planning other records ?
..withe more conventional speakers (with increasing directivity at higher freq.s and a fairly large diffraction signature due to the "box" shape):
If 0 degrees is parallel to side-walls, then
Toe-out of 25 degrees, loudspeakers about 4 feet apart, and me about 5 feet from loudspeakers.
Basically the "inside" diffraction edge of the baffle not quite "pointed" at me.
Loudspeakers further away from any wall than I am to loudspeakers.
..and yes, head movement does cause an imbalance.
(..though I don't find this to be particularly off-putting.)..withe more conventional speakers (with increasing directivity at higher freq.s and a fairly large diffraction signature due to the "box" shape):
If 0 degrees is parallel to side-walls, then
Toe-out of 25 degrees, loudspeakers about 4 feet apart, and me about 5 feet from loudspeakers.
Basically the "inside" diffraction edge of the baffle not quite "pointed" at me.
Loudspeakers further away from any wall than I am to loudspeakers.
..and yes, head movement does cause an imbalance.
Speaker-phones? Sorry, couldn't resist
Speaker-phones? Sorry, couldn't resist
It would seem like that, but it isn't at all like that.
Sadly though I'm all to aware of this: headphones is what I have to work with currently.
(..that Smyth Realiser is looking *real* fine at the moment.
)It would seem like that, but it isn't at all like that.
Sadly though I'm all to aware of this: headphones is what I have to work with currently.
(..that Smyth Realiser is looking *real* fine at the moment.
You could buy my Headzone Pro
You're welcome.
You would need to equalize your speakers at 90° and 120° to compensate for the fall-off above 8kHz.
Hi Tom,
Very interesting. I've only listened to them on in-ear earphones (it's too late at night to try them on speakers) but they sound very much like binaural recordings, with one difference - there is no forward localization of frontal sounds, yet it is not completely "in the head" like a pan-potted stereo recording - sideways localization beyond the head is very strong. Most sounds also localize as being considerably higher than the listening point.
The Harley recording is the best sounding clip, and also best at illustrating the strengths and weaknesses of whatever your approach is.
When the Harley starts it's location is in the head, but as soon as it drives off to one side by about 30 degrees (I'm only guessing about the angle obviously) it seems to localize properly as both a long way to the left and forward. However like all the other sounds it sounds higher than the listener.
As soon as it does a drive by from left to right and passes through a 60 degree frontal angle range the image location completely flattens and sounds like its just above my head, then stretches out in depth again as it passes about 30 degrees to the right.
This is quite interesting and makes a lot of sense if you've used some kind of directional mic's which don't have a HRTF. For locations where the sound is a long way to the left or right, and when the high frequencies are attenuated (like when the bike drives presumably out of direct view) the IID and IPD are probably providing the bulk of the localization and the image position is convincing, with distance cues also provided by the reverberation.
As soon as the bike drives by, comes into direct line of sight (greater direct signal, less reverberation, greater high frequencies, smaller IID/IPD) the brain is expecting the high frequency HRTF variations to take over, but they don't exist in the recording, so the brain thinks "wait, what the ?" and the image quite abruptly collapses to a flat plane near but above the head.
So sideways lateralization is very good, as good as a true binaural recording, but there is simply no depth for any sounds directly ahead. (Or presumably behind)
All sounds also seem to be coming from an upwards angle of around 20 degrees, although I noticed a similar upwards angle effect listening to an actual binaural recording recently, so it could be that my earphones have some emphasis at ~8Khz.
As far as recording quality goes, all are excellent, but a bit on the "dull" side. Bass is indeed very heavy on in-ear earphones that are flat down to about 25Hz, possibly unrealistically heavy. (Mind you, recordings I've taken of real world sounds with a measurement mic do tend to sound bass heavy when played back) Treble sounds fairly rolled off.
Do you have any plans to try a binaural head recording as well ? I would love to hear and compare the same auditory scene recorded both with your existing set-up, and a binaural dummy head
(I can't seem to find enough good binaural recordings...)Edit: For those wanting to compare Tom's recording to a good example of a binaural recording, try this:
http://www.youtube.com/watch?v=IUDTlvagjJA
(Like Tom's recordings I localize most sounds in this clip at a significant upwards angle, but localization is otherwise convincing)
Last edited:
Simon, earlier you talked about the room being a bigger factor than the speaker, Toole and maybe Olive did some work that showed the speaker is the bigger factor. In in Sound Reproduction somewhere.
Dave, looking and all those curves, it looks like "smoothness" or lack of squigliness(sp? he he) is a big factor barring the the solo dip.
Dan
Dave, looking and all those curves, it looks like "smoothness" or lack of squigliness(sp? he he) is a big factor barring the the solo dip.
Dan
I think you're misquoting me or extrapolating something specific I said into a generalization, as I don't remember making any generalizations that the room was more important than the speaker.
Do you have any particular posts you can point to so that I can see the context of what I might have said ? In some aspects of speaker performance the room does indeed dominate, like the response below 250Hz, or the level of reverberation, but I would never say that the room is dominant overall, and in many important performance metrics like non-linear distortion and headroom / dynamic range compression the room plays no role at all.
Absolutely. Sean Olive did a study where he made a number of measurements on a large group of speakers and analyzed the measurements to see which ones would correlate with the same speaker's ranking in blind listening tests. In the end there were 4 primary factors: on axis smoothness, on axis flatness, off axis smoothness, and bass extension (-10dB point). Those factors alone were sufficient to rank order the speakers.
David S
Hello,
Here first WAV file with Bark wavelet:
Click my signature to download the free software to do these plots !
- Elias
Here first WAV file with Bark wavelet:
Click my signature to download the free software to do these plots !
- Elias
Attachments
Back to the roots? I suspect not a lot of people are keen on going back to a DOS-like interface (I'm not). Why not try to get it in one of the more userfriendly apps like REW?
How were on axis smoothness and flatness defined ?
Was smoothness a measure of the slope of errors in the frequency response ? In other words, narrow steep variations in response such as high Q resonances were considered "not smooth" while gradual low slope errors were considered smooth even if the total amplitude error was greater ?
Also did on axis flatness represent adherence to a level flat response, or would a response that (for example) was a straight line (on a standard log/log graph) but smoothly fell by 2dB from bass to treble qualify as a flat response ?
(REW has a very good interface for comparing several measures, because it keeps everything instantly available. But there is no CSD plot. Arta has not this problem, but the only way of keeping a measure is to make a Pir, that's fastidious. Arta is also more intuitive for gated measurements.
A convenient solution is to save the WAV file with REW, then analyzing it with the wavelets that are a "no pity" test. It's DOS-like interface is very fast. It's also useful to run Arta and REW together)
A convenient solution is to save the WAV file with REW, then analyzing it with the wavelets that are a "no pity" test. It's DOS-like interface is very fast. It's also useful to run Arta and REW together)
I'm mostly on a Mac. We're used to software that makes it easier for the user, less to software that makes it easier for the programmer
I don't have Toole's book in front of me but it was well covered there. It is also in Olive's blog. From my recollection he had at least 3 measures of response, one was a 1/20th Octave smoothness of a std deviation nature. One was more 1/2 Octave? and then flatness was absolute deviation from a flat and level curve, i.e. a linear but downsloping curve would be downgraded.
Only on axis needed to be linear. Off axis curves or power response needed to be smooth but any shape would do. Toole's theory is that power response smoothness is the way to look at resonances.
What is this thread about? Speaker directivity, which is also equivalent to the relationship between on axis and off axis response.
David S.
This is the only software in the world that will do for example psychoacoustic Bark wavelet on your impulse response for free
Elias Pekonen Home Page - Wavelet Software
Feel free to improve it, it's GPL.
- Elias
Elias Pekonen Home Page - Wavelet Software
Feel free to improve it, it's GPL.
- Elias
Ah THAT quote. Yes, you have misinterpreted it.
If you read it carefully I wasn't saying that in general the listening room is more important than the speakers within it.
Rather, that in the specific case of that Audissey test suite, the single biggest improvement over a typical home listening set-up that wowed the reporter is quite likely to be the excellent room acoustics and optimal speaker placement, rather than their specific surround technology, or even the speakers in use, as their test room is so much larger than even the biggest home listening room that it puts it into a different category.
You can also be sure that the acoustics of the room have been carefully designed and treated for both absorption and diffusion to an extent usually not possible for those that must appease their other half
Eg it's a large custom designed acoustic space, not a small, slightly or not at all adapted living room.As a home audio enthusiast it's possible to have speakers as good or even better than what they used if you have the money, (or lots of DIY skill) but getting the same room acoustics would be nigh on impossible in a multi-use family room, especially a European sized one.
Speakers may be more important than the room, but for those already having good speakers reaching a certain level of performance, most of the remaining potential sound improvement would come from treating the room, and/or having a bigger room.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.