How do you make a track louder without maxing out your headroom?

[Simplex09]

Hey guys!

I have been trying to mess around with recording tracks and playing them back but I have ran into a bit of a issue that I can't seem to overcome. My tracks seem to be really quiet like they need more volume compared to other tracks I have heard on the internet, From my understanding you can add volume by a few ways (Volume slider, Gain, Normalize, Gain Through a compressor) but even when I do that it seems like its still seems quiet or I max out my headroom and its -0db(red). So I was reading on the internet and on this forum that you don't have to worry about the gain on your audio interface because you can just add more gain later on gives you more headroom later on to edit. So I aim for about -20db which I think the tracks I recorded they are about -18db on advantage. For example after I add gain through the ways I mentioned above for me to hear the exported audio I have windows sound on 60% and my phone maxed out I know phones don't have good speakers but I just wanted to hear it how others may hear it. The amp sounds really full in the room and its quiet loud.

I did use a 3 mics because I wanted to see the same thing recorded with different mics positions

Mic1 > outside the dust cap about 2 fingers away from the grill
Mic2 > Outside dust cap farther away about 4 fingers away
Mic 3 > About 3 feet away from the cab.

What I used to record this was;
-Fender Classic Player Jaguar with bridge dimarzio super distortion and PAF
-Fender Twin Reverb 1980 plugged into a 1960a Marshall Cab.
-DS1 Pedal
-Celestion 1987 G12T-75
-Mic 1 Shure SM57
-Mic 2 Sm57 Copy
-Mic 3 AT2020
-Cakewalk

So the three audio files that I uploaded its raw audio I didn't add any gain or effects at all.

Anyways thanks for your help I wanted to make the post as detailed as I could possible could if I missed anything just let me know!

Thanks!

 

[Deylight]

Hi there. I found getting loudness took a lot of studying and determination. From your description, it sounds like you are trying to get loudness from a live recording without using compression. I think if you want your final recording to be "competitively" loud, you will have to do something like what I describe below. But maybe someone else here knows another approach?

Here's my current method: I have headroom like you describe in a mix from a DAW file. It's quiet.

I bring that first mix into a second DAW file, and put it in 2 tracks. The first track I don't touch. The second track I put a couple compressors on. This technique is called parallel compression, there are a lot of youtube vids about it. The compressed track has a LOT of compression, and its volume is maybe 12db less than the untouched track. The result is increased perceived loudness. In the stereo out that sums both tracks, I'll put an adaptive limiter, with perhaps 4db of gain and a peak set to -.1. If you look at a meter plugin on the stereo out, it should tell you a LUFS number, which means perceived loudness. Professional mixes have a LUFS of between -10 and -6. It's not easy to get there without distorting your mix. After getting a mix file from that second DAW file that has just the 2 tracks plus stereo out, (I call that second DAW file "Master 1") I take that compressed mix into a 3rd DAW file, Master 2, where I basically do the same thing. I just learned about the free Kazrog KClip clipper plugin, which you can also watch vids about, using that in Master 1 and the free Loudmax plugin in Master 2, I get another few db of LUFS loudness, and then it sounds ... pretty loud.

The art is to make it sound spacious and pleasing and dynamic, while at the same time being earsplittingly loud. Right now I'm settling for loud, and trying with each new mix to get closer to pleasing and dynamic. I hope this is useful.

 

[Farview]

The -18dbfs should be your average level, not your peak level.

Stuff you hear on the internet does have all the headroom used up. The final product does most of the time. It just the recording process that the headroom is important.

 

[TalismanRich]

I'm not exactly sure where you're losing your way here. Maybe it's your understanding about volume, gain, and normalization. Gain in recording isn't the same as adding "gain" on a guitar with a distortion pedal. In recording, adding gain is the same as adding volume, at least until you exceed the inherent headroom of the system. At that point you don't add "volume", you add distortion. Normalization is simply looking at the recorded signal, finding the highest peak level and bringing that to 0dBFS (or whatever setpoint you want to use).

Your three guitar tracks are all around -30dB for peak level. You easily increase the input gain (input volume) for the microphones another 20 dB and still be safe from distorting. And if you have that distorted guitar, you're probably going to find that your average level is really high. Highly distorted guitar tracks don't have a lot of dynamics.

Here's your track as posted, and with the gain normalized to -1dB (in this case, adding 28dB). You can see how much headroom you have to play with.

Typically you don't normalize individual tracks unless they are so unbalanced from the other tracks that it makes mixing difficult. If your fader only gives you 20 more dB, then normalizing can help. If everything comes out looking like your posted tracks, then you need to increase the input level on your interface. THEN once you get a good sounding mix, you can increase the volume until your track maxes out a few dB shy of 0, and THEN look to see what your other statistics are (LUFS-M, LUFS-I, Peak).

Instead of normalizing, I will usually do a dry run for the mix, which tells me the peak and LUFS numbers. If the peak is -12, then I'll kick the master fader up 10 or 11 and dry run again. If things look right, then I do the actual render to a WAV file and MP3 file.

 

[Simplex09]

Hi there. I found getting loudness took a lot of studying and determination. From your description, it sounds like you are trying to get loudness from a live recording without using compression. I think if you want your final recording to be "competitively" loud, you will have to do something like what I describe below. But maybe someone else here knows another approach?

Here's my current method: I have headroom like you describe in a mix from a DAW file. It's quiet.

I bring that first mix into a second DAW file, and put it in 2 tracks. The first track I don't touch. The second track I put a couple compressors on. This technique is called parallel compression, there are a lot of youtube vids about it. The compressed track has a LOT of compression, and its volume is maybe 12db less than the untouched track. The result is increased perceived loudness. In the stereo out that sums both tracks, I'll put an adaptive limiter, with perhaps 4db of gain and a peak set to -.1. If you look at a meter plugin on the stereo out, it should tell you a LUFS number, which means perceived loudness. Professional mixes have a LUFS of between -10 and -6. It's not easy to get there without distorting your mix. After getting a mix file from that second DAW file that has just the 2 tracks plus stereo out, (I call that second DAW file "Master 1") I take that compressed mix into a 3rd DAW file, Master 2, where I basically do the same thing. I just learned about the free Kazrog KClip clipper plugin, which you can also watch vids about, using that in Master 1 and the free Loudmax plugin in Master 2, I get another few db of LUFS loudness, and then it sounds ... pretty loud.

The art is to make it sound spacious and pleasing and dynamic, while at the same time being earsplittingly loud. Right now I'm settling for loud, and trying with each new mix to get closer to pleasing and dynamic. I hope this is useful.

Cool! thanks for going into so much detail about it gives me a lot to go read about. I don't mind using compression whatever is the best way to add the volume you want without its being maxed out on head room and causing digital distortion. So I just mean going into the read section on the master fader. But I really appreciate your help I will for sure look into your parallel compression I haven't heard about it before so I'm interested to do some research into it.

 

[Simplex09]

The -18dbfs should be your average level, not your peak level.

Stuff you hear on the internet does have all the headroom used up. The final product does most of the time. It just the recording process that the headroom is important.

Ok makes sense! I mean sometimes trying to learn from the internet can be a bit confusing if that makes sense because there is a lot of useful info and sometimes it can be information leading you down the wrong direction. So I'll do what you said and I'll increase the gain a bit more then it was so -18db will be average. Thanks for your help.

 

[Simplex09]

I'm not exactly sure where you're losing your way here. Maybe it's your understanding about volume, gain, and normalization. Gain in recording isn't the same as adding "gain" on a guitar with a distortion pedal. In recording, adding gain is the same as adding volume, at least until you exceed the inherent headroom of the system. At that point you don't add "volume", you add distortion. Normalization is simply looking at the recorded signal, finding the highest peak level and bringing that to 0dBFS (or whatever setpoint you want to use).

Your three guitar tracks are all around -30dB for peak level. You easily increase the input gain (input volume) for the microphones another 20 dB and still be safe from distorting. And if you have that distorted guitar, you're probably going to find that your average level is really high. Highly distorted guitar tracks don't have a lot of dynamics.

Here's your track as posted, and with the gain normalized to -1dB (in this case, adding 28dB). You can see how much headroom you have to play with.

View attachment 149308

Typically you don't normalize individual tracks unless they are so unbalanced from the other tracks that it makes mixing difficult. If your fader only gives you 20 more dB, then normalizing can help. If everything comes out looking like your posted tracks, then you need to increase the input level on your interface. THEN once you get a good sounding mix, you can increase the volume until your track maxes out a few dB shy of 0, and THEN look to see what your other statistics are (LUFS-M, LUFS-I, Peak).

Instead of normalizing, I will usually do a dry run for the mix, which tells me the peak and LUFS numbers. If the peak is -12, then I'll kick the master fader up 10 or 11 and dry run again. If things look right, then I do the actual render to a WAV file and MP3 file.

Ok! yeah I thought that gain = volume in any daw and gain = distortion on a guitar pedal. So normalizing would be used if you have two different sound files that were recorded at different volumes and you want your average volume the same? Now you normalized the audio file to -1db which added 28db of volume my question is what about headroom that can cause distortion for example;

I did the same as you but on all three audio files that I uploaded and the files are not maxing out the headroom into the red section but the master fader is getting maxed out which when i export into a mp3 it causes distortion. But that happens if you even add a song to the mix also so what would you suggest to turn down the guitars sound files a bit so they average out at -1dB all together?

Or do you just keep your files at lower volumes so you don't max out of your headroom using one file alone?
For example you have three audio files and the volume/gain is lower than -1dB so all together their total volume is what you are looking for?

So you wouldn't normalize individual tracks but would you send multiple sound files to a stereo bus and normalize those so it controls all three sound files or something similar to that?

Also you talked about a dry run so that's a raw unedited? So you would use the fader first then if needed afterwards then normalize if you need too? So my master fader only goes up to 6.0 is that the same meaning as 10 if that makes sense.

Anyways thanks for all your help I appreciate it sorry I'm still a bit of a beginner so I'm trying to just make sure I understand everything. And the post is a bit messy because I added some more details a few times.

 

[TalismanRich]

You sent 3 identical tracks that were normalized to the master? Of course it will be in the red. Audio is additive. Combine two identical files, and it adds up to 6dB more signal. In that case, turn the master down until it doesn't hit the red at all. You can try a little experiment to prove this. Take one of your guitar tracks and normalize it. Then copy that track to a second track. Mute one and see what your master peak is. Now unmute the second track and see what your master peak reads. It should be trying to do +6dB which means flat topped distorted peaks. Do it with 3 identical tracks and you'll end up around +12 and EVERYTHING will be clipped.

This will apply to any mix of audio. If your bass guitar's wave form is positive and your drummer hits his kick, it can easily make a big peak. Drums are big transient peaks so if you have loud drum peaks, everything else will average lower. That's why many time, drums are hit with a limiter. You get most of the impact without the huge transient peak.

Remember that if two signals are positive, they add. If one is going down and the other is going up, the sum will will be lower. Take those two guitar tracks that you added together above and invert one of the signals. You should get silence. That's called a null test.

SO when you are mixing, turn down your master so you don't get read peaks. That way your mix stays the same, just the overall level stays in range.

I don't know about your DAW, but in Reaper, under the render section, you can run the mix through the same process as if you were making a final stereo wave file, but instead of making an audio file, it analyses what you statistic are. It's a quick and easy way to get your final output file close to the goal.

 

[Simplex09]

You sent 3 identical tracks that were normalized to the master? Of course it will be in the red. Audio is additive. Combine two identical files, and it adds up to 6dB more signal. In that case, turn the master down until it doesn't hit the red at all. You can try a little experiment to prove this. Take one of your guitar tracks and normalize it. Then copy that track to a second track. Mute one and see what your master peak is. Now unmute the second track and see what your master peak reads. It should be trying to do +6dB which means flat topped distorted peaks. Do it with 3 identical tracks and you'll end up around +12 and EVERYTHING will be clipped.

This will apply to any mix of audio. If your bass guitar's wave form is positive and your drummer hits his kick, it can easily make a big peak. Drums are big transient peaks so if you have loud drum peaks, everything else will average lower. That's why many time, drums are hit with a limiter. You get most of the impact without the huge transient peak.

Remember that if two signals are positive, they add. If one is going down and the other is going up, the sum will will be lower. Take those two guitar tracks that you added together above and invert one of the signals. You should get silence. That's called a null test.

SO when you are mixing, turn down your master so you don't get read peaks. That way your mix stays the same, just the overall level stays in range.

I don't know about your DAW, but in Reaper, under the render section, you can run the mix through the same process as if you were making a final stereo wave file, but instead of making an audio file, it analyses what you statistic are. It's a quick and easy way to get your final output file close to the goal.

View attachment 149313

Thanks for your help! I downloaded Reaper to give it a try and I seem to have better volume results after your tips. I have been comparing my audio files to other recordings of others playing guitar and seems to be much better! So if audio is additive meaning if you have the same sound even from different mics it adds to the volume on the peaks or low per file that makes sense, So each file adds to the peaks. So I guess you could use a compressor to get your peaks down so you can turn the total gain up from my understanding. Appreciate your help! Sorry for delay in replying but I wanted to give it a try everything that was talked about. Thanks Again.

 

[rob aylestone]

Personally - I never used compression at all for 25% of my recording career, and when I bought one, I really struggled to make it work for me and not be a bigger problem than no compression. Once I could hear it and hear what reaction came from my fingers on the knobs - I suddenly got it. Parallel compression, in my humble view, is not an everyday technique, but one better suited to parts within a mix. I know some people use it on mixes, but maybe you're struggling a bit to get basics sorted, so advanced stuff might be best held off for a bit. For what it's worth, going into the red (in Cubase) is something warning me it's going wrong. Channels, groups, masters, inserted processing - red is bad. Basic things like sending loads of channel's reverb to the same inserted effect can easily go over the top.

 

[Simplex09]

Personally - I never used compression at all for 25% of my recording career, and when I bought one, I really struggled to make it work for me and not be a bigger problem than no compression. Once I could hear it and hear what reaction came from my fingers on the knobs - I suddenly got it. Parallel compression, in my humble view, is not an everyday technique, but one better suited to parts within a mix. I know some people use it on mixes, but maybe you're struggling a bit to get basics sorted, so advanced stuff might be best held off for a bit. For what it's worth, going into the red (in Cubase) is something warning me it's going wrong. Channels, groups, masters, inserted processing - red is bad. Basic things like sending loads of channel's reverb to the same inserted effect can easily go over the top.

That's quite interesting! I have been finding its mostly about reading about how it works online and giving it a go for yourself. Turning it off/on seeing it if improves or doesn't.

 

[rob aylestone]

The snag is (as happened to me), I had not yet discovered what compression sounds like - for years I was listening for the wrong thing. Once it hit me I could not understand why I'd not noticed. Try it first on a naked drum, bass or vocal track so you can hear it not buried in a busy mix.

 

[TalismanRich]

The trick with a compressor it determining the parameters that produce the best sound for you. These are the first 3 to be concerned with:

Attack is how fast the unit engages. With a plug-in, it's possible for the computer to scan "the future", and look for big peaks before they actually reach the mix. Normally there is a slight delay, in milliseconds as to how fast the level reduction takes place, and how long it holds that reduction.

Ratio is the amount of reduction that takes place. If you have a 4:1 compression ratio, then for every 4dB the signal rises going in, the processed signal only rises 1 dB coming out.

Threshold is when the reduction starts to take place. For example, you can set the compressor not to change things when they are below a certain value, like maybe -12dB. What comes in goes out unimpeded as long as you are below that number. But when the signal goes above that threshold level, it compresses by the ratio. So a signal that peaks at -8dB will actually be processed so that the result is -11dB, -12dB + 1dB for the compressed signal using a 4:1 ratio. (-8dB is 4dB louder than -12).

There are other parameter, Hold, Release, Knee, Make Up Gain but you need to get control of the first 3 before you start playing with the other parameters.

You can get a complete explanation of all the compressor adjustments quite a few places on the internet. Just Google Compressor Parameters and you should find at least a half dozen articles. In the end, the numbers are critical, you need to listen. If things sound "squished" then you lower the ratio. If you hear the sound "pumping" then adjust ratio and threshold. There's no hard and fast rule, like drums must be 10ms attack, 6:1 ratio or -20 threshold. Eventually you will find a set of parameters that work for you and then make know of those settings.

 

[Simplex09]

The trick with a compressor it determining the parameters that produce the best sound for you. These are the first 3 to be concerned with:

Attack is how fast the unit engages. With a plug-in, it's possible for the computer to scan "the future", and look for big peaks before they actually reach the mix. Normally there is a slight delay, in milliseconds as to how fast the level reduction takes place, and how long it holds that reduction.

Ratio is the amount of reduction that takes place. If you have a 4:1 compression ratio, then for every 4dB the signal rises going in, the processed signal only rises 1 dB coming out.

Threshold is when the reduction starts to take place. For example, you can set the compressor not to change things when they are below a certain value, like maybe -12dB. What comes in goes out unimpeded as long as you are below that number. But when the signal goes above that threshold level, it compresses by the ratio. So a signal that peaks at -8dB will actually be processed so that the result is -11dB, -12dB + 1dB for the compressed signal using a 4:1 ratio. (-8dB is 4dB louder than -12).

There are other parameter, Hold, Release, Knee, Make Up Gain but you need to get control of the first 3 before you start playing with the other parameters.

You can get a complete explanation of all the compressor adjustments quite a few places on the internet. Just Google Compressor Parameters and you should find at least a half dozen articles. In the end, the numbers are critical, you need to listen. If things sound "squished" then you lower the ratio. If you hear the sound "pumping" then adjust ratio and threshold. There's no hard and fast rule, like drums must be 10ms attack, 6:1 ratio or -20 threshold. Eventually you will find a set of parameters that work for you and then make know of those settings.

Thanks for explaining it it can be a bit confusing at first. I downloaded the 1176 compressor because uaudio was giving it out for free. I have also been reading that adding a highpass filter on your tracks can help remove the lowend that you can't hear and free up headroom from my understanding. Thanks for all your help after joining this site I'm coming up with much better results than before.

 

[JamEZmusic]

Thanks for explaining it it can be a bit confusing at first. I downloaded the 1176 compressor because uaudio was giving it out for free. I have also been reading that adding a highpass filter on your tracks can help remove the lowend that you can't hear and free up headroom from my understanding. Thanks for all your help after joining this site I'm coming up with much better results than before.

Yeah be careful with that. It's not the magic bullet that people think it is.

But, if you are planning on creating a youtube channel that teaches people how to mix, then feel free to record a (Highpassing is your best friend) video to add to the 2000+ others that already sing it the highest praises.

 

[Simplex09]

Yeah be careful with that. It's not the magic bullet that people think it is.

But, if you are planning on creating a youtube channel that teaches people how to mix, then feel free to record a (Highpassing is your best friend) video to add to the 2000+ others that already sing it the highest praises.

Excellent thanks for your help! What would you suggest for a highpsss filter 90 to 200 is what i heard

 

[UltimateOutsidr]

@Simplex09 - There is a lot going on in this thread. There is a lot of art and science that goes into producing a final mixed track that sounds loud without being distorted or fatiguing. But your original post seems to be asking about loudness regarding individual tracks inside your DAW, rather than a final mixed/produced song.

When it comes to recording individual audio tracks, my philosophy is generally to record at the loudest volumes that give you the least amount of distortion and/or background noise, and favor the gain adjustments closest to the source audio (loudest signal with the lowest noise floor). (Example: If you have a volume control on your instrument and a record gain control on your audio interface, increase the instrument volume first when trying to set your recording levels. Only increase the interface or preamp gain if your instrument gets too noisy/distorted at high volumes and you don't think the recorded track will be loud enough without boosting the preamp gains.) Don't worry about headroom when recording; it's almost always best to record your multitracks at the loudest levels available (without artifically boosting them) without peaking. You achieve headroom by turning the faders down in your DAW during playback/mixing.

If you ARE indeed asking about how to produce a loud-but-still-good-sounding final/mixed track, I have a lot of tips to share in that regard, but I'm not sure if that's what you're looking for.

 

[Simplex09]

@Simplex09 - There is a lot going on in this thread. There is a lot of art and science that goes into producing a final mixed track that sounds loud without being distorted or fatiguing. But your original post seems to be asking about loudness regarding individual tracks inside your DAW, rather than a final mixed/produced song.

When it comes to recording individual audio tracks, my philosophy is generally to record at the loudest volumes that give you the least amount of distortion and/or background noise, and favor the gain adjustments closest to the source audio (loudest signal with the lowest noise floor). (Example: If you have a volume control on your instrument and a record gain control on your audio interface, increase the instrument volume first when trying to set your recording levels. Only increase the interface or preamp gain if your instrument gets too noisy/distorted at high volumes and you don't think the recorded track will be loud enough without boosting the preamp gains.) Don't worry about headroom when recording; it's almost always best to record your multitracks at the loudest levels available (without artifically boosting them) without peaking. You achieve headroom by turning the faders down in your DAW during playback/mixing.

If you ARE indeed asking about how to produce a loud-but-still-good-sounding final/mixed track, I have a lot of tips to share in that regard, but I'm not sure if that's what you're looking for.

Well thanks for the comment! I'm more comparing my audio to others audio and I'm seeing that I have to turn up my sound to more to hear it clearly if that makes sense.

 

[UltimateOutsidr]

Well thanks for the comment! I'm more comparing my audio to others audio and I'm seeing that I have to turn up my sound to more to hear it clearly if that makes sense.

But are you talking about your individual multitracks (guitars, vocals, drums, etc) or your final rendered stereo WAV file where everything's mixed together?