What is the speed of "sound"?

Hi all,

I've been trying to measure the time it takes for an llSay or llShout to reach a remote object, and wonder if anyone has already done some experiments on this? It seems that an object with an llListen will "hear" closer objects that speak to it before it hears more distant objects, and of course, lag plays a considerable factor as well.

Currently I am placing several test cubes within 30 metres of a listener object, the listener object sends a shout to the test cubes who in turn shout a response back to the listener object, and I am getting wildly varying results.

A single test cube placed within 10 metres of an empty sim is providing a result of around 1/3 second - repeated tests on 11 test cubes spread over 30 metres with 11 agents in the sim is providing a total elapsed period of anywhere between 6.5 and 12 seconds to receive all responses.

The more avs in a sim obviously the greater the lag (or lower sim fps) - so I'm wondering if there's a formula to predetermine how long a response should take to receive with lag and distance taken into account.

Cheers,
Dominus

Dominus, I've heard of several tests on this exact issue. Here is one such thread.

He's not asking about the speed of the listen event.

He's asking if there's a delay given the amount of distance separating two objects in a sim between sending and receiving data from a say/to a listen event.

I'm torn between saying yes and no. On the one hand, it would be a lot simpler I would think to ignore distance and fire off the listen event in each script in range. On the other, the delay in calculating the distances and such may be a factor.

Given the intricacies involved, I don't think it's possible to figure out without asking a Linden and hoping they know. Too many variables (llSay/Listen delay, Sim FPS, etc).

Ah. My assessment is it's based on a couple factors:

1) Sim lag. Definitely.
2) What's actually being "checked" by the server or listen.

I cannot easily answer this without offering some alternatives and conjecture. In my experience, for example, an llShout from distance has been roughly the same speed as an llSay nearer to the source. This is to say that it's what I've seen - as opposed to how it actually works.

As for an alternative, one can use a timer or other event and llKey2Name to receive messages at a near-infinite distance within the same sim in a timely manner. Similarly, if the data need not be from a source in Second Life, one can write some very simple XMLRPC to send data faster, to anywhere in Second Life, with a rough delay of only a couple seconds.

I would say, then, that it all depends on what you're expecting to do. If you'd like a simple example, take a shot at my pinball machine in Tantalum. That's a perfect example of distance versus timing, as the "shout" is placed in a mobile ball and the listeners are in one area.

As just about everything in SL is plesiochronous (love that word*) I would expect that internal mutliplex delay, scheduling of client updates, variation in internet latency, and client load would swamp any real variation in propogation of "chat" messages and that trying to measure them would be folly. Futhermore, there is a greater latency in messages coming from an adjoining sim, such that an agent 20m from you in the same will hear you sooner than one across a sim border.

Looked at another way, if llSay (~20m) and llShout (~100m) have identical latencies - as mentioned above - the speed of sound could be considered infinite as there is zero propogation delay over 80m. For that matter, updates of the visual space (what you see) is bounded by the same event list. Thus, the speed of light is roughly the same as the speed of sound. Max Plank would probably go insane in SL; and I shudder at the relativistic effects of such an overlap. But now I'm being silly.

(*plesiochronous adj. -- nearly synchronous, but not.)

You mean... almost, but not quite, entirely out of synch?