Hi
I think if you are in the development stage or fine tuning something then measurements are critical. They can tell you which way is toward the more faithful device.
I say faithful because for me that is the goal, that is how closely it reproduces any signal I feed it. Depending on the individual recording, they often but not always sound better too the more faithful the device is.
Since I have no control over what kind of music will be played, it has to be faithful which is neutral.
The problem is, we don’t hear like a microphone measures something I recall rambling on about here some time back. That doesn’t make a microphone useless, not at all, in fact compared to your ears, the microphone is less forgiving.
What we hear is an image composed by our auditory process which was two separate signals one from each ear. In addition depending on the incoming angle and height, the shape of your ear puts a weird pattern of comb filtering in the pressure response.
Rather than hear these things like they appear in a measurement, we only hear them as the height or angle of the source.
In a room, one has a true jumble of room effects and while a listening position measurement of a typical speaker might be a horrifying looking + - 10 or 20dB, we really hear through much of that or at least focus on the sound we want. Yet, the “extra” stuff away as much as possible (like room treatments and / or directivity) and more of the recorded image gets through.
We are totally unaware of the auditory process (we know nothing else) but it is both brain, eyes and ears. Hard to believe it but your brain will even choose what it knows to be based on sight over what ones ears tell. You only hear reality when your eyes are closed and that knowledge removed. Try the McGurke effect;
http://www.youtube.com/watch?v=G-lN8vWm3m0
I have urged people that if they have a measurement microphone and a sound card in the computer, they can make wonderful in home recordings which when you play them back, may make the hair of the back of your neck stand up because they ARE so familiar. Nothing beats things you know as reference material .
The whole thing falls apart when you try to get stereo (most recordings are not a “capture” of an event but composed at the mixing board) so use one mic and output to both channels.
Personally I believe we are usually doing the wrong thing when we capture “stereo”.
While measurements are my faithful partner in loudspeaker development, we have been doing something at work which I think has proven to be a useful reality check. So far as I am aware, no loudspeaker company has done this or let it help guide their product and yet it is not a measurement at all.
I figure you guys might be interested for evaluating all kinds of stuff and best of all, we only adapted an old standard technique.
At the point when I was able to make the first SH-50 Synergy horn produce a square wave over a broad bandwidth, covering all three frequency ranges and so appearing to be one driver, it also occurred to me that we would be approaching clear weather visibility of at least the neighborhood of the "theoretically ideal source". Anything that is ideal, can be inserted in a reproduction chain and not alter the sound.
When Mike and I started DSL, we started making generation loss tests like they used to do with tape and electronic circuitry EXCEPT we inserted the loudspeaker and a precision measurement mic and used a modern 24/96 recorder instead of a chain of identical units.
The speakers are on a forklift up in the air outside (no room effects), the mic at two meters and a modest level (90dB).
We used a 24/96 digital recorder and when the music track was played, the mic signal recorded and next time around the recorded mic signal played. On each generation, the speakers (or whatever) flaws or deviation from ideal becomes progressively exaggerated into a sonic caricature of all the warts.
AS a reality check they sometimes have recorded a direct music track too to keep track to the AD/DA process with each generation. One can go many generations though on that side.
The original plan was we would have an audition button where one could hear not only ours but our competitor’s speakers recorded identically. It sounded easy but the web guys dragged their feet and a few suggested we shouldn’t have our competitors, but then what’s the point without the reference.
A new company is working on a new web site but I don’t know if they will be there or not yet.
What is interesting is that some of the speakers we tried sounded pretty flawed listening to the first recording alone with no generation loss. These tended to be speakers that significant response aberrations and or if you played pink noise and moved the microphone left to right or up and down, that there was a noticeable to dramatic effect on the pink noise sound (remember this is up in the air and not from room reflections. It is radiating a complex sound field that is different from place to place.. These are things which are small changes over small distances like the speaker made something like a speckle pattern (an interference pattern).
If you guys are trying to zero in on something, I would urge you try the generation loss recordings, they provide no numbers but make flaws in the chain stand out like nobody’s business.
The weird thing, at least for some loudspeakers, after you hear a caricature and then go back and hear the speaker straight through, you can often hear that flaw a lot more clearly. Well I have no reason other than my feeling on it but I think it has been a useful thing for us anyway.
A weird thing happened, on the 15th the reps for the company had an open house where we demo’d a new loudspeaker for hi fi on a large scale for stadiums etc.
One of the people that attended, some one I didn’t know took a HD video and good sound and posted it on his facebook page and then on ours.
I have decent headphones on my computer and I though this captured the event rather well, for me one of those hair standing on the back of my neck things like home recordings.
Also though, it captures what it sounds like to have a single acoustic source with a lot of pattern control as well as one which will tolerate several generations of loss. If you have heard casual recordings where a loudspeaker is captured accidentally, you know what they can sound like or at least if they are not faithful.
In the beginning, the camera guy walks to the subwoofer which is out of the pattern of the flown full range speaker and while it probably felt cool, it overloaded his mic.
What is not obvious is how loud this was, at about 1:30 a fellow walks up to the guy next to the camera and speaks, adjust the volume to make the voice here in reasonable scale.
Most of the time, the camera guy is outside / under the pattern, the majority of the attendees were actually in a field up to 900 feet away, at around 2:10 he pans out to the field, the ridge a bunch were standing on is 450 feet away from the speaker.
http://www.youtube.com/watch?v=5MOG_sPejGA
If you have a broad band CD source, the spectral balance doesn’t change appreciably with distance or angle with pink noise OR music. If one has a single CD lobe over a broad band, then one can do another “old days” sound trick, you fly the speaker and aim the speaker at the farthest seat and use the underside of the lobe to contour the SPL. AS you get closer, you are increasingly out of the pattern.
In one football stadium where they used one of our JH-90’s for the entire stadium system, the closest and farthest seats were only 4dB apart in SPL.
The other cool part for stadiums is where a wind really messes up the sound with a line array (because it also radiates an interference pattern), these act like a single source radiating a simple homogeneous field and so the wind has very little effect.
As people (many in large scale sound) were only able to walk to 900 feet, a caravan of cars took off to listen from a greater distance. Mike took this video from the sunroof on his truck from about 1500, about twice as far as any foot ball stadium would require. If you recall seeing a large building near the speaker, you can see the building in the video 2 but not the speaker, you can hear the wind blowing on the camera’s mic, but not swishing the sound.
Here I think the possibilities for large scale audio “practical joke” or possibly a giant outdoor movie come to mind. In the 80’s I hid some large speaker in my backyard near a swamp and played frogs and crickets as loud as I could. Well I thought I was funny then, but you could hide a bunch of these and control them remotely and have much more effect.
http://www.facebook.com/photo.php?v=2587095730763
Anyway, consider the generation loss tests we do as something that can be applied to any aspect of the audio gear you folks are fiddling with. Our loudspeaker test would be perfect for home speaker too.
It doesn’t quantify anything, doesn’t tell you what’s wrong or how to fix it but you can sure hear it and most loudspeakers are unlistenable at only 1, 2 or 3 generations (which is the reason i have focused on this at work).
Best,
Tom Danley
Danley Sound Labs
http://www.danleysoundlabs.com/
http://www.facebook.com/DanleySoundLabs?ref=ts
