@Defender and others, I'll clear up the db thing, but don't worry, I will not talk about the math formulas behind that, I'll give you an intuitive understanding as well as maybe clearing up some misconceptions.
First, what does dB measure? Nothing. That's right, dB is not a unit, it measures nothing. If you tell me that something is 4 dB, I'll ask you compared to what. This is because dB is a ratio. It is used in a lot of fields when you need to have close numbers for things that can be magnitudes apart, like earthquakes, power, electricity, sound, etc.
Ok, now that we know that dB are a ratio and are useless on their own, how could we make it so it's convenient to use? By setting a reference point of course. This way, we can have consistent ratios, because it will always be compared to the same reference. So there are multiple dB scales out there. What makes them different is the reference point which can be of a different value and unit. But there is one thing they all have in common, the scaling.
A dB scale will always obey this rule: every +10 means times 10. Why +10 = times 10? Because they are decibels. If we were talking about bels, then it would be +1 = times 10. Actually it was like that before, but the numbers were too small, so we said hey, let's use decibels instead. It's just how we decide to use centimeters to measure smaller stuff.
Ok, let's have some fun. Let's create a dB scale together. It's really easy, you will see, because all we need is a reference, remember, that's the only thing that changes between dB scales. So our scale will measure excitement in a game. We need to name that scale, so let's say it's the dB Origine scale, measured in dB O. Now let's set our reference of 0 dB O to the level of excitement while configuring the speech options in a game.
That's it, we've created a dB scale, the dB Origine scale. Now, let's say someone comes and say "you should try that game, the first level is an overall 10 dB O garanteed". This means that playing the level 1 of that game will give me 10 times more excitement than configuring speech options.
Now I might reply "10 dB O isn't much, you sure I should try this game?". If he says "trust me, the first level is only a tutorial, level 2 and onwards is a 20 dB O minimum!".
Now what does it mean? Does it mean that level 2 is twice as exciting as level 1? No! It's 10 times! This means that playing level 2 will give me at least 100 times more excitement than configuring speech options. Just to make sure it's clear, here are some key points in the dB O scale:
Minus 30 dB O: staring at a computer screen, doing nothing. This is 1000 times less exciting than configuring speech options.
Minus 20 dB O: dieing for the 50th time at the exact same place in a level. This is 100 times less exciting than configuring speech options.
0 dB O: Configuring speech options.
60 dB O: Beating a game for the first time, unlocking a new character, ending up first on the online leaderboard and receiving an e-mail congratulations from the developer. This is a million times more exciting than configuring speech options.
Now let's get back to sound. There are two main dB scales for sound. First one is dB SPL and the other one is dB FS which stands for full scale.
dB SPL is perhaps the most known one. The reference point is 0 dB SPL = the quietest sound we can possibly hear. It is measured in sound pressure which have watts equivalents, but let's not care about that. Here are some key points on the dB SPL scale:
0 dB SPL: the quietest we can hear
30 dB SPL: your bedroom at night
40 dB SPL: inside a library
60 dB SPL: Conversation 1m distance
120 dB SPL: treshold of discomfort
130 dB SPL: treshold of pain
140 dB SPL: a jet engine, 15m away
The dB SPL scale measures acoustic power. This is different than loudness because of how our brain treats the sound. While + 10 = 10 times acoustic power, it is generally said that +10 = twice the perceived loudness. This means that a library will seem twice as loud as our bedroom at night.
Say there is one person shouting 10m away from you. Let's say it's 70 dB. How many people does it take for you to hear them shouting twice as loud? While you might think it would take 2 people, it actually takes 10. This is because 10 people will produce 10 times more acoustic power which will be perceived as twice as loud: It would take 100 people to hear it four times as loud. It's quite fortunate it works like that if you think about it or else we would never go in a concert or go see a sports game.
Ok, we're almost done here. The other scale for sound is dB FS or dB Full-Scale. The reference point is 0 dB = the maximum a sound can be without distortion.
This is the scale you see in Reaper, in any DAW, in any recording instrument. Over 0 dBFS, there will be clipping, the sound will saturate. Now reaper can normalize the sound when rendering it to avoid clipping, it can make sure that 0 dBFS is the maximum amplitude in the sound it renders. However, if you hit 0 dBFS while recording an instrument, you're fucked. It clipped and there's nothing you can do about it. This is why you must set the gain correctly.
The idea is that you want to put it as high as you can without clipping. It is generally good to leave you some room, you can aim for -10 dBFS while recording, you can always boost it later. But you might be thinking, why don't I record at -30 dBFS or -60 dBFS, I can just boost it later? Short answer: the recorded sound will have a much lower quality. This is because of two things:
1) Amplitude Resolution also called bit depth.
2) Noise: this can be electrical noise, ambient noise, etc.
The bit depth determines the resolution of the amplitude of your signal. If you have 24 bit depth, it means every point will be encoded over 24 bits. This sets a limit to the quietest sound that can be encoded, which is only the lowest bit set to 1. You should play much above that limit.
Then there will be noise, there is always noise. There is electrical noise because of our cellphones, lights, laptops, lightning, cellphone towers, etc. This noise gets caught into the wire you record with because it is like an antenna. No matter what you do, there will always be noise, you can reduce it, you can filter it sometimes, sometimes you make it so low that it doesn't bother you, but there is always noise.
So if when recording the noise is at -40 dBFS and you also record at -40 dBFS, then you record yourr instrument as loud as the noise, which will give a terrible audio recording. If you record at -30 dBFS, it's much better, because the noise is now 10 times less present in your recording compared to your instrument. If you record at -10 dBFS, then it's 1000 times lower. If you record at 10 dBFS, trick question, the audio clips at 0 dBFS, so you will only hear distortion.
We don't hear a difference between -40 dBFS and -30 dBFS, because twice as loud of something that isn't loud much is still isn't much. You probably wouldn't have guessed that a library is twice as loud as a bedroom, because both are quiet, but you can notice the difference between someone talking and screaming, because talking is a memedium loudness and screaming is 2-4 times louder, so it gets loud. If you eat 1% of a cake, it's not much. If you eat 2%, although it's twice as much, still not much. Now if you eat 25% of the cake, that's a reasonable amount, but if you eat 50%, although twice as much, it's quite a lot more cake you've eaten.
Finally, I used dB O, dB SPL, dBFS in this post so it's clear, but depending on what you are doing, people find it too long to use the full name, so they only say dB. If you mix, record audio, work in a DAW, assume dB = dBFS, if you are talking about sound power of common things like concerts, machinery or in every day life, assume it's dB SPL. And if you are talking about excitement in games, assume it's dB O
Reading is one form of escape. Running for your life is another. ― Lemony Snicket