Thursday, September 5, 2013

Thieving Abilities - Hear Noise

I'm going to take a crack at hear noise, among the most annoying, least defined and generally crappy game mechanics in AD&D ... and at the same time, absurdly difficult to improve.  I've never been able to do so, I don't know of any other mechanic that does more than simply inflate the base problems - first, that the character supposedly fails the roll and can't hear something that ought to be perfectly clear, and second, that what can and cannot be heard is far, far too subjective, and therefore almost wholly dependent upon the grudging or generous DM.

What's needed is a measurement ... and as it happens, in sound there is one.

Having spent some time trying to hammer down the mechanics of decibels, I've put together the following table:

Now, while I think this idea may in fact be easy to use, once one is adjusted, initially it is a bit of a mind fuck ... so take that ol' mind out and give it a good stretch.

The numbers were put together from an online calculator, so I admit that I did not do my own math; this one was a bit beyond me, I'm afraid ... but remember, we're not trying to pass a physics test here, we just want a workable system for D&D.  One where we can measure sound, not just point at it.

The feet/yard/miles columns are included so the DM can choose what particular measure is useful given the particular circumstances.  The examples of sound for decibels is more or less what you'll find online, though I've worked to remove all modern references, such as rock concerts, jet planes and jackhammers.  Those are very common examples you'll find on decibel sites.

How this works is this:  the distance indicated is that point where the sound being made has been reduced to barely audible (12 decibels).  It's that distance that the sound can just barely be picked up ... because, it must be understood, the amount of sound a thing makes is dependent on how far away you are from that sound.  That's obvious ... but where is the mechanic that demonstrates it?

If you half the maximum distance, you increase the level of sound by 6 decibels (not double!).  Half that distance again, and that again increases the sound by 6 decibels.  This is how the decibel mechanics actually work.  A sound that is 6 decibels higher can be heard at twice the distance.

So the hammer on an anvil can be just picked up at 3/10ths of a mile ... but it sounds very minute, like the clear sound of a pin dropped on a floor in an otherwise silent room.  At 273 yards, the hammer sounds as loud as the rustling of leaves.  At 137 yards, something like the collective sound of a nearby farm.  And so on.  At nine yards, as loud as a stream, at six feet, as loud as a dinner party.

Sound is such a deceptive thing, however, and the real problem with your perception of a hammer right now is that you tend to think of it as loud as it is when you're standing right next to it, actually using the hammer.  It's hard, really, to grasp how loud it sounds, exactly, from nine yards away.  Moreover, it is almost impossible for you to get a good grasp on how loud the rural countryside is, or a 'quiet garden.'  That's the mind fuck.  Every idea you have about the relationship between sound and actual recognized sound-making is purely subjective.  You simply have to discard that perception.

I would like a better set of sounds than 'quiet street' and 'noisy street' ... but that's going to take even more research than I've done now, and I'd like to start to get a handle on what kind of sounds I'm looking for.  The thing is, most examples that are given bear a certain modernistic similarities - I got better traction by looking up specific things.  I'm going to have to spend a bunch of time thinking up a list and then looking up those things and placing them on the decibel scale.

Meanwhile, the system in general should work. 

See, figure what the sound level is, then ask yourself as DM, is the party close enough to hear that?  Then consider this ... can the thief pick out the sound that's relevant apart from other sounds?  Not just that there's voices, but what the voices are actually saying.

That ought to be a gradient as well.  However, while I've been piecing out the above, I'm stumped for how to improve the thief's actual ability to decipher the noise.  I don't like the idea of a straight percentage roll ... which sucks.  Perhaps an ability to nail down sounds that fall into a certain range, and then that range expanding.  Between 57 and 62 at 1st level, then between 56 and 63 at second, etc.  Note that the range system above allows a very precise measurement of where the sound overlaps that range, and the difference in a few points could make a big difference.

But I'm just spitballin' here at the end.  I want to think about it more, try out the base template, and see if I don't get an epiphany.


Mic B said...

Wow, this is giving me food for thought and preventing my sleep!

I can't quickly find a way to eliminate the % roll but, setting a baseline of "certain to hear" level (at the listener's position) and adjusting for ambiant sound should work well.

Thing is, anybody could hear a 48dB sound... but not over a furious battle (i'm quite sure that a constant 96dB is having a lot of people involved!).

Also, no matter how loud a sound, is it's intermittend, it will prevent hearing the details of a conversation, but not the fact there is one. But a constant and relatively soft sound would never prevent anybody from hearing a scream.

Working with %, and going on with your base idea, i'd use the constant ambiant sound level as a penalty and the "target" sound level as a bonus to the roll to hear a single occurence of the target. To hear the details of a muted conversation, i'd impose a penalty of the worst of either ambiant or intermittent ambiant level.

Also, knowing that protective ear-wear are rated in dB reduction (anywhere from 10 to 35 according to my last shopping session), i'd use this as a guideline to reduce the effective dB level for obstacles (like a door), also reducing it by distance.

Then you'd need a "base distance" for the check, let's say 1 yard. Every doubling gets you a -6dB. But according to this, the 1st level thief would never hear anything, so I'd give him a flat bonus (undetermined)

now an exemple: the player's are trying to determine if somebody is hiding (respiration only, 6dB?)in an abandonned house (wood wall, -12) during the day right beside the town market (ambiant 66). The person is 4 yards away (-12). Then... a Detect Noise check at -84% should be made. This was presuming that a first level thief has a 100% chance of success, not a 15% base chance (so maybe a bonus of 80 should be given with this system?)

This would need a table with corresponding penalties for range. The dB levels would be used as bonus and penalties also. But still using the %...

But with just pinpointing sounds level... maybe instead of using a base range, use a relative one, like "minus XdB/level". Maybe simpler is better!

Scarbrow said...

I think you're on the right track here, Alexis. But I'm a little stumped because I don't know the precise mechanics for "Hear Noise" on AD&D (for what is worth, this post is currently #1 on Google for "hear noise ad&d"

Fortunately I have a copy of the AD&D Player's Handbook here. Combined with my precarious grasp of the system from Baldur's Gate, I understand the basic system was that the thief would roll a d%, and if the result was under his or her skill, the thief heard whatever noise it was. Which sucks, of course.

Now, what you've given here is a scale for noise, which is part of the solution. You are still lacking a compare mechanic for that scale, if I understand correctly. As you say: "can the thief pick out the sound that's relevant apart from other sounds? "

I think you have to take surroundings into account. Since you state "the distance indicated is that point where the sound being made has been reduced to barely audible (12 decibels)." I must suppose that sound can't (usually) be heard if there are louder noises around you. For if there is a soft whisper 3 feet from you, but there is rustling of leaves around you, it's just not going to register for common people. Which is where the thief comes in.

I suggest that any person can pick up a sound that is twice as strong as the surroundings. So the aforementioned whisper will be heard by any and all at 1 feet even through the rustling of leaves around them (two steps down the table is +12dB, so the 12 apparent decibels at 3 feet become 24dB, which is +6dB from the rustling leaves around you (18 dB), which is twice as much perceived sound. That would be my baseline.

Then I would grant thieves the ability of picking sounds that were of one lower rank (half perceived intensity) per level, so a lv 5 thief standing on a quiet garden would be able to detect the pin dropped at the road 13 feet away. Or, if you still need to reduce the feature to probabilities, I would apply then the base d% roll. So you first decide the amount of noise (pin drop or dragon road) , then adjust it by the distance between the sound source and the thief (plus dampening from walls, doors and such), etc, then compare it to the thief's surroundings, then adjust by thief's level. If and only if a thief of such level would be able to perceive that amount of noise, would he roll for "hear noise".

I suppose it will still require a few adjustments. But it's a start. How do you see it?

Koewn said...

Here's a guy walking and running in plate armor, which sounds like a good thing to put on a table for Hear Noise:

Light Googling says treadmills at pace run around 70 decibels, and it *seems* like he's overpowering that noise pretty handily. But that assumes the microphone is the one on the camera.

I'd expect any adventurer worth their salt to not be quite *that* clanging, though :)

Neklan Krasna said...

"But a constant and relatively soft sound would never prevent anybody from hearing a scream."

Oh, but that's the whole point MicB! What you think of as a soft constant sound, white noise, Does prevent you from hearing that scream, way more than you think it does!

The conversation came up on this blog not so long ago about how loud the "silent outdoors" can be.

So, Alexis, the table you posted indicates whether it is possible to hear something. The thief's hear noise ability is whether they can recognize the noise. I had the experience last night of not "hearing" plenty that my girlfriend was saying, despite the fact that it must have been clearly audible. The question in my mind is, should the hear noise encompass both the ambient noise and the thief's ability to realize that the noise is important? Should it also encompass the likelihood of the noise occurring at all? I absolutely don't think the latter, but I do think the roll should include the ambient noise, for two reasons. It's quite difficult to figure out what the ambient noise levels are, and also it's a legitimate question how lulled is the thief by those white noises.

So, that was really a long way of arguing against scarbrow's baseline. Instead what I'm saying is that, given the sound is audible, what is the likelihood that a thief will pay attention to the sound over the ambient noise. So, treat each doubling of apparent volume as a doubling of the chance, and apply modifiers as necessary for unusual ambient noise levels. (perhaps just subtract the one from the other) Then roll the percentile chance as given.

A level 1 thief stands on the bank of a stream and a sword strikes a shield 205 feet away (I'm calling that hammer on anvil.) The apparent loudness of the stream is 5x and the apparent loudness of the strike is 4x remember that the stream is a special case source of white noise, over and above the rustling of leaves, etc. and so counts as a modifier. Thus, I'd argue that the thief's chance to hear the sword strike is 15%/(2*(5-4)), or 6%
In an empty forest, the thief would have a 60% chance of identifying the swordblow, and on a still night, a 75% chance. I have no idea what the chances should be for the fighter standing next to them.

5x and 4x above represent the number of doublings of the distance until the sound is inaudible.

How's that strike you?

Mic B said...

Neklan Krasna: I think we're on the same page here. I was thinking that lower ambiant noise would not prevent an "adjaccent" stronger noise from being heard. Not a distant one. Also, we must remember that Detect Noise is an active ability, so the Thief would be very attentive to a particular sound (maybe what's behind that door, or what those 2 over there are talking about...)

Not sure I got your formula right, but it seems good. I'd use the strongest ambiant sound, with maybe a modifier for the lesser one if it's significant.

For the fighter's chance, I'm from AD&D 2nd and I think the only difference is that the thief can distribute 30 points among the 8 abilities pretty much as he wants. In that case, any class has the same % chance to Detect Noise as a 1st level thief's base chance (15% for humans)

Neklan Krasna said...

I guess I just proposed an alternative to your "decibel ranges" idea without actually justifying why I thought the one was better than the other.

There are really two concepts rolled into the skill of "hear noise" The first is the ability to recognize that a noise is important. This is the thing that makes you jump out of bed, or suddenly hush everyone around you so that you can listen more closely, and is similar to the skill of noticing that something is out of place visually. The second is the ability with concentration to understand what that noise is. The latter concept is further broken down into two more cases, the first in understanding spoken word, and the second in identifying the difference between, say a hammer on an anvil and a sword on armor.

I often listen to the news on the radio while I'm cooking dinner, and I find it very difficult to understand what is said while the faucet is running, or a pan is sizzling, if I'm not interested, but I've found that if I pay close attention, I can understand it without fail. I'd argue that thieves have trained themselves to be hyper-vigilant, so they're going to be able to do what I do with concentration, without, but it still seems that there are a lot of variables involved, not the least of which is their own mental processes, but which do also involve those things it just doesn't make sense to model in game, i.e. exactly how much background noise there is.

as an aside, I'd note that the ability to understand two people talking is strongly dependent on the type of background noise (easier to make out over rustling leaves than over other voices, even if the leaves and the voices are at the same volume) and also is heavily influenced by your ability to read gestures, expressions, and lips.

Anonymous said...

Let's do some empiricism... Referring to your helpful table, and to evidence from experience, signals three steps below the ambient noise barely can be discerned. I would give any character a 10:20 shot at those... -1 for each additional step below ambient, +1 per thief level above 1st.

Bill Dugger said...

You've made a couple important mistakes in your post, Alexis.

One sound is twice as powerful as another with a mere 3 dB difference, not 6 dB. A quick explanation: it's common knowledge that a 10 dB difference is a 10x multiplier, so since 2^3 is around 10, you want to have 3 doublings in 10 dB. 3 is the closest whole number to that. (The exact answer is sqrt(10), I believe).

Using the inverse square law from physics, if a sound is twice as far away, it's actually one-quarter as powerful, not one-half.

I'm trying to come up with a simple system that uses this. I'll probably post it tomorrow.

Alexis Smolensk said...

Thanks Bill,

I did think I was out of my depth.