Volume Volume Volume - relative and absolute

[BloomingLate]

Hi there,

so I'm having a bit of a problem with mixing which is worsened by the fact that volume levels are not consistent across the devices that I use. Basically this is usually what happens:

I start working on a mix on my laptop. Initially I just need to get the right notes in my project file, but soon enough I need to playback the sound to see if what I'm doing is working. At that point the crappy onboard speakers are no good because they lack bass. To get to a semi-workable level I need to set the volume to 40. To get a more complete picture I then go to my headphones, but level 40 would destroy me so I have to tune it back down to like 8 for it to be comfortable. Alternatively, I could use my Trust speakers, but they may or may not be somewhat broken (it is hard to tell because I'm so used to the sound). I have no idea what a normal volume level for that thing would be. I sometimes have to turn the dial way over 50%, while I used to get by with just 25% (though maybe I was compensating with the computer volume control).
Yet another alternative is hooking up an external device that lets me connect to my Sharp stereo system. But even at volume level 12 (which is where purchased music starts to sound properly loud for me but too loud for my wife), the volume is just too low. Using my mp3-player headphones I can barely hear my mixes at 50%.

With all these relative values and different results between devices I'm having trouble knowing if what I'm doing is actually "right". Add to that the fact that my ears may plug up from internal pressure and now I can't tell anything is balanced or loud enough.

Is there a formula or a way to kind of figure out what the "absolute" volume output is relative to these different devices? So that I may have confidence that what I produce is still loud or not loud, irrespective of how a given speaker or headset presents it? Does this tie in to being able to read the Db meters on mixer channels?

At this point I suspect my monitor headphones produce the most realistic and best signal, but working with them on quickly wears me down and ruins my hearing.

[Meteo Xavier]

Is there a consistent and objective authority on how loud an audio track should be? A little bit YES and NO. There's a whole world of debate on this subject (Google "Loudness Wars", just as an example) and is a very prominent audio subject for the absolute audio production masters to bring to the forefront of attention. You can get to a consistent standard of loudness based on numbers if you like, but there isn't really a universal standard that applies to audio production all throughout the world. You can build that consistency yourself, you can find an experienced audio professional to ask on that (maybe slip'm a few bucks to loosen their lips), or just pay a mastering engineer to handle it and use THAT as the example to go for.

That's not a very concrete answer. Think of it like schools of martial arts. You can make up your own fighting style out of scratch, you can study under a strict discipline, you can do something in between and still potentially learn how to fight effectively, but there's no universal perfect answer for the question of how to learn martial arts.

What I do is set my audio computer's standard volume at 61. THAT is the audio level that my personal taste and experience says is the perfect spot there. If the audio track I'm working on sounds too loud at 61, then I turn the entire track way down. If it's not loud enough at 61, I turn the whole track way up. But then, if I'm ready to publish the song or album, I do hire out for a mastering engineer to take care of it because I don't do mastering and just trust that they know what they're doing.

[JohnStacy]

I use a pink noise generator.  I set it to -12db, then mix relative to that.  One track at a time I level them so that they can barely be heard over the pink noise.  This makes the levels consistent, and most of the time it ends up putting the mix at between -9 and -6 depending on the project.  9/10 times that is actually a decent volume that isn't obnoxiously loud, but it also leaves a lot of headroom if for whatever reason the mix needs to be made louder.

[timaeus222]

Well, for one, at least start your mixing on headphones, and not your laptop speakers. They are certainly not designed for mixing, and you know they have no bass, and so they will not sound right anyway...

-----

There is no exactly correct loudness, but you can get close to what is too loud for you. I probably listen to music about 3 dB louder than other people, but it allows me to hear more detail, it still doesn't hurt my ears, and I still write music with consistent loudnesses. YMMV.

I have a loudness reference that I like to use. This song is about as loud as I would listen to before I don't want to turn it up any higher. Try it when your ears are cleared, on your system when using the headphones specifically, and turn the internal system volume up until you want to turn it down, and find that balance (if you are using "Audio Enhancements" or whatever it is on your OS, turn that off, it messes with you). Establish a consistent perceived loudness for your different listening scenarios for the same song, that ALSO uses the same internal system volume. That is, try to make the same song feel about as loud across the board... but keep the internal system volume the same to make it easier on yourself.

To further adjust volumes on speakers, turn the knob on the hardware, instead of changing the internal system volume. That way you are at least keeping one setting constant across your comparisons.

[BloomingLate]

14 hours ago, Meteo Xavier said:

Is there a consistent and objective authority on how loud an audio track should be? A little bit YES and NO. There's a whole world of debate on this subject (Google "Loudness Wars", just as an example) and is a very prominent audio subject for the absolute audio production masters to bring to the forefront of attention. You can get to a consistent standard of loudness based on numbers if you like, but there isn't really a universal standard that applies to audio production all throughout the world. You can build that consistency yourself, you can find an experienced audio professional to ask on that (maybe slip'm a few bucks to loosen their lips), or just pay a mastering engineer to handle it and use THAT as the example to go for.

I knew I was getting myself into controversy once again. Okay, that is good to know. I'm nowhere near the point where I am producing music for consumption, so I'll have to keep this in mind for if or when I get there.

At any rate, I suppose there is no way to ensure that a song will sound good for every person on any device, so that is not a goal to pursue to begin with.
Now that I think about it, it does make perfect sense to start with my own relative perception.

 

11 hours ago, JohnStacy said:

I use a pink noise generator.  I set it to -12db, then mix relative to that.  One track at a time I level them so that they can barely be heard over the pink noise.  This makes the levels consistent, and most of the time it ends up putting the mix at between -9 and -6 depending on the project.  9/10 times that is actually a decent volume that isn't obnoxiously loud, but it also leaves a lot of headroom if for whatever reason the mix needs to be made louder.

That's an interesting approach. I quickly Googled it and found some online tone generator that helped explain the differences between white, brown and pink noise (a bonus is that I now have a name for that awful sound I hear when my soundcard "crashes"... its called white noise, and I hate it :P). -9 and -6 would be approaching the "yellow" part on the Db meter, correct? I noticed that my recently bought trance tracks are all maxing out into the red area and I generally find that uncomfortable. Between -6 en -4 is best for me.

 

9 hours ago, timaeus222 said:

I have a loudness reference that I like to use. This song is about as loud as I would listen to before I don't want to turn it up any higher. Try it when your ears are cleared, on your system when using the headphones specifically, and turn the internal system volume up until you want to turn it down, and find that balance (if you are using "Audio Enhancements" or whatever it is on your OS, turn that off, it messes with you). Establish a consistent perceived loudness for your different listening scenarios for the same song, that ALSO uses the same internal system volume. That is, try to make the same song feel about as loud across the board... but keep the internal system volume the same to make it easier on yourself.

To further adjust volumes on speakers, turn the knob on the hardware, instead of changing the internal system volume. That way you are at least keeping one setting constant across your comparisons.

I've been thinking about using an existing song as a reference. Your approach makes enough sense. However, would I be at risk of "blowing up" speakers by turning up the knob of the hardware without adjusting the internal volume? My Trust speakers may need to go all the way up to 75% to sound decent with my comfortable laptop level and at that point they start cracking somewhat. I can hear them produce white noise when no sound is going through (is that normal?).
I actually suspect I "blew them up" (not in a definite sense) before, and maybe I've been doing that consistently with other devices because of the hearing thing. Oops!

Thanks everyone, for your answers! I think I know what to do now.

[timaeus222]

10 hours ago, BloomingLate said:

I've been thinking about using an existing song as a reference. Your approach makes enough sense. However, would I be at risk of "blowing up" speakers by turning up the knob of the hardware without adjusting the internal volume? My Trust speakers may need to go all the way up to 75% to sound decent with my comfortable laptop level and at that point they start cracking somewhat. I can hear them produce white noise when no sound is going through (is that normal?).

I actually suspect I "blew them up" (not in a definite sense) before, and maybe I've been doing that consistently with other devices because of the hearing thing. Oops!

The point of turning the knob on the hardware instead of adjusting the internal volume is that you aren't changing one of the only things that could be consistent across your listening methods. If your speaker 'noise baseline' is audible, it may be that the impedance on them is high (it takes more power to achieve the same volume as something with less impedance) and you should consider getting new speakers with lower impedance (or an amp to boost their base volume, so that you don't have to turn the knob up so high).

[Nabeel Ansari]

@timaeus222 I think you're going in a perceived loudness direction which is a little more advanced than the kind of issue BloomingLate has. The issue here is simply that OP doesn't understand the dB scale, which is the "absolute loudness" measurement he's looking for.

BloomingLate, you can raise the master track of your song up to 0 dB FS, which is the digital limit for clipping. You should always mix to 0 dB because that's the standard for mastering. 0 dB is marked at the top of the loudness meter in your DAW software.

The dB number has absolutely no bearing on the perceived sound energy without a consideration of dynamic range (you can still have soft music where its loudest peak is 0 dB). If you don't like a high amount of sound energy, mix to 0 dB but avoid any master compression or limiting so that nothing goes over. In other words, avoiding 0 dB doesn't mean you're avoiding making the music sound too loud, you're just annoyingly making people raise their volume knobs relative to all the other music they listen to.

To explain your own example, trance music isn't loud because it's at 0 dB (the "red" part), it's loud because it's very compressed with little dynamic range, so the sound energy over time is packed and you feel it harder in your ears.

For a practical solution to your problem, you can also render your mix so it never hits 0 dB (to truly avoid the need any master compression and limiting) and then just Normalize it. This will make your music at least hit the same peak that other music does, and shouldn't require the listeners to vastly pump up the volume to hear. However, I would wager that without any compression whatsoever, people will still be raising their volumes. Most music is compressed in some form nowadays, and I can't remember the last album I saw with full dynamic range (besides classical music, which is impossible to listen to in environments like the car because of said dynamic range).

As for the volume levels of your devices (headphones, laptops, stereo), none of that stuff matters at all. If someone's listening device is quiet and they need to dial it to 70% to hear anything, that's their problem. If your music is mixed to the same standards as everyone else, then it will sound the same on their system as any other music they listen to, and that's what you shoot for. This is the 0 dB thing I was talking about before.

How loud it sounds is a matter of handling dynamic range using stuff like compression, and that's what Timaeus is talking about with referencing a track to match the perceived loudness. That stuff is its own rabbithole and takes a lot of learning and experience to understand how to do properly.

tl;dr If you mix it so that you go up to but never cross 0 dB, you will never blow out speakers/headphones and your signal won't distort. This is one of those things that should just be automatic for every piece of music you create.

[timaeus222]

33 minutes ago, PRYZM said:

@timaeus222 I think you're going in a perceived loudness direction which is a little more advanced than the kind of issue BloomingLate has. The issue here is simply that OP doesn't understand the dB scale, which is the "absolute loudness" measurement he's looking for.

Yeah, that's true, and it's good that you cover that below; however, it also looked to me like he was having issues with figuring out how he can match up all of his listening methods to hear a consistent loudness across them, because he was unsure of what to adjust (without re-rendering, presumably) and ended up adjusting the internal volume control several times in the process... So, after establishing a setting from previous listening methods (say, headphones), it's now changed (say, after trying speakers), which would then lead him to go back and adjust again... which could be a frustrating cycle.

Hopefully either of these approaches (mine + PRYZM's) help though! I think both were good to have.

[Meteo Xavier]

11 hours ago, BloomingLate said:

At any rate, I suppose there is no way to ensure that a song will sound good for every person on any device, so that is not a goal to pursue to begin with.
Now that I think about it, it does make perfect sense to start with my own relative perception.

No, there is, that's what the mastering phase of audio does. Most consumer devices don't give all their frequencies out when they play audio because that requires heavier, more expensive stuff inside the device to do, so they typically only focus on getting the mid/high frequencies out because that's what people are really trying to hear in the first place. Most people are not in-depth music/audiophiles and don't need to hear all the frequencies - it's a cost-cutting measure and that's why your laptop and consumer speakers typically don't come equipped with it, and why you've been hearing as you have.

You CAN potentially teach yourself how to do this and it does start with relative perception, it'll just be a journey like everything else worth doing. I never learned mastering because after learning all I know on composition, production and mixing, it just wasn't worth it to me. Get you some AKG K-headphones and see if you can take some mastering or audio production courses somewhere and you'll be amazed how well you can accomplish this obstacle.

[Nabeel Ansari]

2 hours ago, timaeus222 said:

Yeah, that's true, and it's good that you cover that below; however, it also looked to me like he was having issues with figuring out how he can match up all of his listening methods to hear a consistent loudness across them, because he was unsure of what to adjust (without re-rendering, presumably) and ended up adjusting the internal volume control several times in the process... So, after establishing a setting from previous listening methods (say, headphones), it's now changed (say, after trying speakers), which would then lead him to go back and adjust again... which could be a frustrating cycle.

Ah I see. In that case, I get what you're saying. Definitely want to never touch the internal volume. Mine's always at 100%, but that's because I have separate volume knobs for my headphones and speakers.

I have 0 consistency in what my (interface physical output) volume is set. The output knob to my monitors will move randomly at least 10 times a day just listening to random stuff (like other music, YT vids, etc.) But when I'm making music, and doing a final loudness check on my stuff wrt perceived loudness, I definitely pull up something professionally done and released on iTunes to reset it back to where I'm comfortable listening to mastered music. At this point, I can just visually see where that knob position should be that I hear "powerful and comfortable" for professionally done music.

It's around 30% on the knob dial for speakers, and 100% for the headphones (with the Sonarworks calibration giving -7.9 dB, and having the 250 ohm DT 880's, my interface really needs to try hard) And then I just take that sort of familiar volume comfort zone and then mix my desired perceived loudness there; this way, I know exactly how my music is going to contrast with other albums people might be streaming alongside me, because I picked that volume level while listening to other stuff. And I know that other stuff is at 0 dB, so I put mine to 0 dB too. If they're both 0 dB, and they both sound just as loud, then they are just as loud, absolutely, on all devices.

Timaeus hit the nail on the head. Before you mix anything you should set your listening volume while listening to a reference track, preferably something professionally done and commercially released.

[BloomingLate]

16 hours ago, PRYZM said:

@timaeus222 I think you're going in a perceived loudness direction which is a little more advanced than the kind of issue BloomingLate has. The issue here is simply that OP doesn't understand the dB scale, which is the "absolute loudness" measurement he's looking for.

 

16 hours ago, timaeus222 said:

Yeah, that's true, and it's good that you cover that below; however, it also looked to me like he was having issues with figuring out how he can match up all of his listening methods to hear a consistent loudness across them, because he was unsure of what to adjust (without re-rendering, presumably) and ended up adjusting the internal volume control several times in the process... So, after establishing a setting from previous listening methods (say, headphones), it's now changed (say, after trying speakers), which would then lead him to go back and adjust again... which could be a frustrating cycle.

Hopefully either of these approaches (mine + PRYZM's) help though! I think both were good to have.

PRYZM is right in that I don't understand the dB scale, and timaeus222 was right to catch my frustration with the discrepancies between devices that causes me to adjust mixes to a given device, only to learn that I completely ruined it for another device. My general inexperience and confusion makes me mash two problems together, so you both did me a service by seeing them as two seperate things.
However, now I feel like I'm in way over my head with all this talk about mastering and whatnot.

I guess this is where I am at now with my understanding:

1. I think I understand that dB level refers to the signal strength and that something called "clipping" occurs when the signal goes over that 0. Not sure what clipping "looks" like yet, but I suppose I can Google that myself. At any rate, I see that the volume level of my device does not impact how high the dB meter will reach, so those are two seperate things.
2. I may be confused about the difference or relation between Hertz and dB. I guess Hertz refers to the range of sound a human hear can pick up at all?
3. Audacity has at least two options: enhance sound and normalize sound. Do I understand correctly that enhancing the sound would raise everything to the chosen value (including the already strong parts) wheras normalize only affects the parts that are subpar at that point?
4. So I need the total output of tracks/channels (the Master) to hit 0dB but that doesn't imply that all individual channels need to hit 0? Its about the cumulative effect?

As you may be able to tell I am at seriously n00b level at this point in my development. Like I said in my post about discouragement: it is difficult to know where to begin learning sometimes because stuff can be overwhelming. I found that a lot of online tutorials and articles assume a familiarity with key terms and concepts that I simply don't have as someone starting from scratch. Some concepts like Compression are not quite landing in my brain yet. Am I correctly understanding Compression now as something that reduces the gap between dynamic ranges? Something that may raise the hearability of a sound but not necesarily the power of the sound. That is to say, a whisper will still be a whisper, not a shout, only it will be a loud/hearable whisper?

So many questions, so many things to be confused about :D

 

[BloomingLate]

15 hours ago, Meteo Xavier said:

No, there is, that's what the mastering phase of audio does. Most consumer devices don't give all their frequencies out when they play audio because that requires heavier, more expensive stuff inside the device to do, so they typically only focus on getting the mid/high frequencies out because that's what people are really trying to hear in the first place. Most people are not in-depth music/audiophiles and don't need to hear all the frequencies - it's a cost-cutting measure and that's why your laptop and consumer speakers typically don't come equipped with it, and why you've been hearing as you have.

You CAN potentially teach yourself how to do this and it does start with relative perception, it'll just be a journey like everything else worth doing. I never learned mastering because after learning all I know on composition, production and mixing, it just wasn't worth it to me. Get you some AKG K-headphones and see if you can take some mastering or audio production courses somewhere and you'll be amazed how well you can accomplish this obstacle.

Okay, I see. Thanks for correcting me on that. Now I understand why nearly every computer and speaker system I ever had sounded so crappy. I always found myself EQ-ing the lower ends of the spectrum when listening to music 0_0; I don't know what brand of headphones I have, but when I did the research it seemed like it was very suitable for our purposes (no bass boosting or whatever).

Random side note: I hadn't been in an electronics store for like a decade so I was in for quite a shock: drones, HD cameras and TV screens...And I learned that they're moving in a completely wireless direction for headphones, which I do not like! I rather tangle with the chords than having to deal with batteries all the time :D

[timaeus222]

9 hours ago, BloomingLate said:

I guess this is where I am at now with my understanding:

1. I think I understand that dB level refers to the signal strength and that something called "clipping" occurs when the signal goes over that 0. Not sure what clipping "looks" like yet, but I suppose I can Google that myself. At any rate, I see that the volume level of my device does not impact how high the dB meter will reach, so those are two seperate things.
2. I may be confused about the difference or relation between Hertz and dB. I guess Hertz refers to the range of sound a human hear can pick up at all?
3. Audacity has at least two options: enhance sound and normalize sound. Do I understand correctly that enhancing the sound would raise everything to the chosen value (including the already strong parts) wheras normalize only affects the parts that are subpar at that point?
4. So I need the total output of tracks/channels (the Master) to hit 0dB but that doesn't imply that all individual channels need to hit 0? Its about the cumulative effect?

As you may be able to tell I am at seriously n00b level at this point in my development. Like I said in my post about discouragement: it is difficult to know where to begin learning sometimes because stuff can be overwhelming. I found that a lot of online tutorials and articles assume a familiarity with key terms and concepts that I simply don't have as someone starting from scratch. Some concepts like Compression are not quite landing in my brain yet. Am I correctly understanding Compression now as something that reduces the gap between dynamic ranges? Something that may raise the hearability of a sound but not necesarily the power of the sound. That is to say, a whisper will still be a whisper, not a shout, only it will be a loud/hearable whisper?

So many questions, so many things to be confused about

No worries. It's good that you are actually asking, and you seem to be getting it so far. I don't use Audacity that much, but I can give you my thoughts on the rest.

1. Visually, you would have to inspect the mixer dB meter to see that something went over 0 dB. Sometimes (if the DAW is programmed that way) it'll show red, which should indicate clipping. If you haven't heard it before, it sounds like a crackling, but it might be subtle if you don't exaggerate the volume level. I know I didn't perceive clipping in my first year learning music production without looking at a mixer.
    
2. You can think of hertz (frequency) as the horizontal span of what we can hear, and dB (loudness) as the vertical span of what we can hear.
   
   Hertz is a frequency unit (as opposed to, say, length units of meters, or mass units of kilograms). It just means that for every second that passes, one wave cycle passes by, so 5 Hz means that 5 cycles pass per second. It is also written Hz, or 1/sec.
   
   Some producers might say that higher frequencies "brighten up" a sound, etc. Lower frequencies sound "boomy", "rumbly", sometimes "muddy", etc. Higher frequencies span the treble range (generally around 4000 - 20000 Hz, give or take) via high notes, and low frequencies span the bass range (generally around 20 - 1000 Hz, give or take) via low notes. Lastly, midrange (the rest) is where most sounds lie, like the human voice, lead instruments, guitar, etc. It's where you hear presence, and all listening devices hit some of this range.
   
   On the other hand, dB is basically just loudness. Louder, or softer, at a fixed frequency.
    
3.  
4. Yes, the net result is what you should pay attention to. As a general property of audio waves, their amplitudes (loudnesses) will add, and also cancel, at various points, and so, if every channel hits 0 dB, the net result will exceed 0 dB. Therefore, NOT all channels "need" to reach 0 dB by themselves.

[Nabeel Ansari]

9 hours ago, BloomingLate said:

1. I think I understand that dB level refers to the signal strength and that something called "clipping" occurs when the signal goes over that 0. Not sure what clipping "looks" like yet, but I suppose I can Google that myself. At any rate, I see that the volume level of my device does not impact how high the dB meter will reach, so those are two seperate things.
2. I may be confused about the difference or relation between Hertz and dB. I guess Hertz refers to the range of sound a human hear can pick up at all?
3. Audacity has at least two options: enhance sound and normalize sound. Do I understand correctly that enhancing the sound would raise everything to the chosen value (including the already strong parts) wheras normalize only affects the parts that are subpar at that point?
4. So I need the total output of tracks/channels (the Master) to hit 0dB but that doesn't imply that all individual channels need to hit 0? Its about the cumulative effect?

Here's some answers in more plain english:

1. Clipping sounds like crackling distortion. Just keep raising the volume and eventually you'll hear it. That's the rule in computers; when it goes past 0 dB, it will clip. Because the computer can't process stuff that is louder.

2. dB is just a measure of amplitude (loudness). Hz is completely different. When you have a pure tone, it's a wave moving at some number of times per second. 1 Hz is one pulse per second, 2 Hz is 2 per second, etc.

The range of human hearing is 20 Hz to 20,000 Hz (that's generous, it's different per person. My ears stop at 17,000.) http://www.szynalski.com/tone-generator/

Go to this site and you will quickly understand what Hz means in the context of what you hear. Low Hz is bass, high Hz is treble. Raise and lower the Hz on that website and you'll see as you go from the lowest number to the highest number, you're going from lowest pitch to highest pitch.

3. Normalize just means it finds the loudest point in your song and raises the entire song equally all at once so that THAT particular loud point is 0 dB. Usually normalize in programs lets you tell it what to normalize to (-3 dB, -1 dB, 0dB, etc.).

4. The overall loudness of your song is whatever is passing through the master channel. That's what you're hearing, and the master channel loudness meter tells you what it is.

[BloomingLate]

Alrighty, thanks for the additional comments, both of you. That helped in clearing some of the confusion.

Random fact: I found out through that tone generator that and the use of some sub-bass that these low tones are generally what cause my ears to plug up u_u; I could hear up to 15.000 but sometimes I needed to turn my head to actually pick up the sound. Sometimes I worry that I destroyed my hearing by not heeding my grandmother's warning about listening to loud music on headphones