My master mix is always quiet!

My master mix is always quiet!
0

I record and, mix, and master everything in reaper. I record guitar and bass into it, program midi drums and midi orchestra and I love making orchestral/symphonic metal. The thing is, i’ve been trying to improve my mix and mastering skills over the years, and many people have told me to use YouLean Loudness Meter to watch my levels. I usually master to around -2.0 True Peak, and around -11 to -9 integrated LUFS. I don’t smash my stuff through limiters, and all I have on the master is a very very slight EQ, widening, and limiting with Ozone. I have 30 ohm headphones and find myself having to boost my volume all the way to 100 to hear it at an acceptably loud level. But once I do all this and finish up my song, and then distribute it to spotify, it’s SO quiet compared to professional mixes! For most songs I listen to I keep the volume around 44-50 (on the windows volume buttons), anything louder than 58 hurts. But when I play my song from spotify I have to boost it to at least 60 or 70 to get the same level! Why? I’m following all the advice and mastering tips I was given, but it’s still so quiet! What’s up with that?

Well, this is sure to open a can of worms! :roll_eyes:

The first point I should mention is that this statement presents somewhat of a paradox:

Have you measured the stuff you are comparing your mixes to? I don’t think I have ever found a pro mix that masters to -2dB true peak. Whether doing that is good practice or not is beside the point. Since probably virtually every major label mix is mastered to peak at 0dB full scale, you are already putting yourself at a 2dB disadvantage.

My guess is that the mixes you are referencing are at least 3dB hotter than that… but as you say…

My guess is that your reference material has some pretty assertive limiting going on… but that is not the full story either…

… It comes down to this… The louder your mix is, the easier it is to get a loud master.

Getting a loud - sounding mix is a lot about how you handle the midrange, as it is about LUFs measurements. It’s all about the frequencies you cut and the frequencies you emphasise.

For example, you might push up the high mids in your mix to make it sound loud to the ear (because the ear is most sensitive to midrange frequencies) , but if your low end and your low mids are too full, then although your mix might measure quite loudly in terms of LUFs, comparing it to another mix where the low mids and lows are handled by a pro, the imbalance in the low to high frequencies in your mix will cause the other mix to sound louder and punchier, even when it measures close to the same LUFs.

This is something that takes a long time to get a handle on, so you have to be patient and experiment with how sculpting the frequencies of the individual elements in your mix can make them sound loud and punchy long before they hit the final limiter… This is part of the art of getting a professional mix.

…but there is more to it than that. The other aspect is transient control on individual elements. The following thread suffers from from the aforementioned “can of worms” syndrome, but it outlines the basic concept.

Now I’ll just sit back and eat some :popcorn:

1 Like

Hi, can you provide a Spotify link to one of your songs that you think is too quiet, and another link of a reference song that you think is well mixed and mastered and sounds louder?

Also, how are you listening to Spotify exactly? Smartphone? Computer? App or browser? And do you have the default settings or did you turn normalization off?

These are all important things to check before answering your initial question.

I don’t know why that’s happening with your loudness levels. I use the same things as you, Reaper, Youlean Loudness meter and set many of my heavier songs to similar Lufs levels as you do (between -9 and -13, depending on the song). I’m always happy with my volume levels and they have no difficulty keeping up to the professional’s levels. I don’t use Spotify but my guess is that they are somehow attenuating your music for some unknown reason. This is just a wild and somewhat inebriated guess. :wine_glass::wine_glass::shushing_face:

https://open.spotify.com/track/2Z0XzYbWxrO6jXnqVzJxxZ?si=717f7366c7f347cb My track.
https://open.spotify.com/track/4Dsxp3FZYCLUxYjQpQo1BP?si=7841b22c75034976 Comparison.

I’m listening to spotify on the computer with the same headphones, desktop app. And yes I do have normalization turned off.

Thanks for all the detail! That’s odd though cause most times I ask about what LUFS to master to for metal, people say -2dB true peak and -9 integrated.

Very strange - I don’t think I have even come across relatively tame pop/rock songs being mastered at such levels…and certainly not heavier stuff!

These videos may be helpful too…

Thanks. I analyzed both tracks with normalization on and off and here is the result, I think it is pretty self explanatory…

Top left is the original reference track, top right is your original track.
Bottom left is the normalized reference track, bottom right is your normalized track.

Here are the measurements.

Fight to Forget (reference)
Normalized: LUFSi -14.0 / PLR 5.9 LU / TP Max -8.1 dB
Original: LUFSi -7.0 / PLR 7.4 / TP Max +0.4 dB

The Battle (your track)
Normalized: LUFSi -15.0 / PLR 14.1 / TP Max -0.8 dB
Original: LUFSi -15.2 / PLR 14.1 / TP Max -1.1 dB

So what you are comparing here is a song that has been heavily limited and pushed to a high loudness (the reference track) with a song that is unusually soft for this music genre (your track).

In fact, your track is so soft that it has been turned UP by the algorithm by approx. 0.3 dB, while the reference track has been turned DOWN by a whooping 8.5 dB.

It is interesting to note that your dynamics are fully retained after normalization, while the loud track’s lack of dynamics is exposed even more since the total dynamic range from noise floor to highest peak is smaller after normalization.

So the logical conclusion here would be that you need to make your mix/master louder. However, I personally don’t recommend making it as loud as the reference track, which in my opinion doesn’t sound great. It is important to note that last year, 85% of the music that has been streamed was normalized, and this number is constantly increasing. There are very few people who turn the normalization off. So in my opinion, unless you are distributing to a physical medium (CD…) you should care about how your track sounds when it is normalized more than anything else.

How does your track compare to the reference when it is normalized? Well, I urge you to do the test… Your track actually sounds louder and fuller. Note that if you’re using the app, you need to close it and restart it after turning normalization on or off, the change only applies after restarting.

How loud should your track sound? In my opinion the answer isn’t a fixed number. You should just mix and master it so that it sounds the best it can. This is very subjective of course, and there isn’t just one ideal loudness/dynamics combination that is right.

I am currently mastering a metal EP and because the producer trusted me to set the dynamics and loudness as I see fit, the songs are somewhere between -9 and -10.5 LUFS. This is conservative compared to the common practice in this genre, however I am sure that his EP will sound louder and fuller normalized than a lot of other metal songs because it will have more dynamics and its low end won’t be restrained so much by limiting.

I am an advocate for dynamics, so I usually master with more dynamics in mind than loudness, but this is purely a matter of taste. Some people like the sound of heavily compressed material. If you’re one of them, then by all means go for it. Just remember than the vast majority of your listeners will listen to the normalized version of your song.

2 Likes

Not sure I understand here, Andrew. Are you saying that you haven’t heard pop/rock songs that are softer than -9 LUFSi?

Id’ start off by saying to ignore that advice. There are a lot of people who go around talking about the benefits of decreasing the level of your masters. The benefit is more dynamics. The downside is quieter songs. There’s a pretty strong discrepancy between what mixing engineers say should be done vs what mastering engineers are actually doing.

Also, people are saying it doesn’t matter the loudness because Spotify will compensate. Based on your example, this is obviously untrue, so you can ignore that advice as well.

Remember, you are not mixing music to impress other engineers, you are mixing music to impress your listeners. They don’t care about LUFS and true peak. They care that it sounds like the other music they listen to.

3 Likes

The context of the comment is referring to the general loudness levels of CURRENT major label music, much of which is mastered at least to -8LUFs…Sometimes much louder (as your analysis above proves)…

…I’m 52 years old - of course I have listened to music that was mastered softer than that!..Context, context, context!..:upside_down_face:

It is true that a lot of the current major label releases is mastered loud, but “at least -8LUFS” is not correct in my experience. I have been researching this topic for a few years, analyzed a ton of releases and I think you might be overestimating the average loudness. I believe there is more label music mastered at reasonable levels than you seem to think. Also, there is a steady decreasing trend in this regard: more and more people in the industry are realizing that loud masters end up being played back softer than dynamic ones. The decline of the compact disc and other unnormalized media is no stranger to this trend, of course.

Here are a few examples that include some heavy or EDM music and are all recent (and in the millions of streams):
Daft Punk - Beyond: -11,7 LUFSi (https://open.spotify.com/track/0k1xMUwn9sb7bZiqdT9ygx?si=2ee0871edf4b4cf2)
Bombay Dub Orchestra - Strange Constellations: -16,9 LUFSi (https://open.spotify.com/track/7pdfCkCQsxrJru6cMdvD6w?si=6f679f68bcea49a9)
Leave the Door Open - Bruno Mars: -11,1 LUFSi (https://open.spotify.com/track/7MAibcTli4IisCtbHKrGMh?si=966fce0a3f42441d)
Katatonia - The Winter of Our Passing: -11,2 LUFSi (https://open.spotify.com/track/2mM8dxT3Ql6WIgVpE6iXN7?si=6d3cd68f0ab14729)
Tool - Pneuma: -12,1 LUFSi (https://open.spotify.com/track/03sEzk1VyrUZSgyhoQR0LZ?si=3e2aae8dab4d4e69)
Tycho - Outer Sunset: -10,9 LUFSi (https://open.spotify.com/track/6MUuMaiREbhGc2HvvBYFik?si=666ab5905de24a95)
Katie Melua - Joy: -14,3 LUFSi (https://open.spotify.com/track/7GxxuoeZoMszrcGKL79Arg?si=7505d72eaed04028)
Khruangbin - Time: -16,6 LUFSi (https://open.spotify.com/track/4fMIe4htuvA8RoygYDFqO5?si=c5fd28c20365445b)
Hyperion - Gesaffelstein: -14,0 LUFSi (https://open.spotify.com/track/3jRl9iaoq6BmkkDbFmndpi?si=d514349fcffc481c)
Jon Hopkins - Singularity: -12,7 LUFSi (https://open.spotify.com/track/5S3F5GL8LASIjPg2PMiU1Q?si=5de01ed5c9344c39)

I have tons of other examples, I could go on and on…

All of the above examples sound way better, louder and fuller than most “loud” masters in my opinion. But please don’t take my word for it: go listen through the app with normalization on.

Several of these songs have been mastered by some of the greatest mastering engineers in the world: Bob Ludwig, Bob Katz, Antoine Chabert… and they are all advocating for less loud and more dynamics, and have been for a long time.

Here’s an interesting blog post about this topic: https://www.sonametro.com/blog/mastering-loudness-dynamics-and-eq-in-todays-music

Just to finish on a light note:

Jean Marc, I’m not arguing with you, and I have no interest in doing so. You asked me a question - I answered it, truthful to MY experience and the music I have in my collection - Simple as that.

Once again, the thread has gone way off topic, as I predicted it would.

I don’t see that as arguing… it is a discussion, or maybe a debate. And I’m not sure how a discussion on loudness is “way off-topic” when the OP asks a question about exactly that…?

You were the one starting this, by implying that the vast majority of current releases are louder than “at least -8 LUFSi”.

I think this is incorrect and can have an adverse effect on someone reading this, who might think that the right way to master is to master louder than “at least -8 LUFSi”. So I am just giving my opinion, which indeed is different from yours. I hope there’s nothing wrong with that.

I think that the ideal loudness is specific to each individual song, and is also a subjective choice that the producer or artist should be able to make freely. There is a lot of advice on the internet about specific loudness targets and because of that, some people think they “have to” reach such and such target, but this is not the case. Everything ends up at the same average loudness anyway, in most cases. So the only thing that should matter is to get your music to sound the best it can. If this means loud, fine. If it means dynamic, fine as well. But dynamic doesn’t really exist above -8 LUFSi.

You see, this is what you have wrong. You’re assuming, and you continue to assume my opinion is different to yours on the subject of loudness. I have never said that I prefer loud mixes, yet you continue to imply that.

I never said the OP should or shouldn’t do anything. I simply explained why they were experiencing a differential in loudness.

Ok this is starting to look like an argument now and I’m not comfortable with that, especially when I have so much respect for your musical talent and contributions to this forum, so I’ll leave it at that. I’m sorry if I made wrong assumptions about you.

1 Like

That’s great, 'cause neither am I.

Well, the feeling is mutual - right back at you, my friend! :slightly_smiling_face:

No worries - it happens… Peace, love and rock’n’roll! :beerbanger:

Dang what on earth happened here hahahah

Xenon, I’ll give you my take for what it’s worth. I’ve been a mix and mastering engineer for what feels like 100 years. All the people chiming in here to you, are correct. You may ask “how are they correct if they all have different opinions”?

The answer is simple. Over time, things have been written (not by anyone here, just in general) to totally confuse people to where they try to make this look like a dark art. It’s all horseshit. You find your own way and the people here have read between the lines and found what has worked for them. You have to do the same.

Your song has to be mixed “for loud” if you want a loud master. So what Andrew told you was absolutely correct about eq and transients.

To me, transients are more the deciding factor unless you have a mix that has loads of sub low or piercing highs that appear as frequency boosts.

But most of the time it will be kick drums, snares, hats, a vocal lacking enough compression and bass guitar pick or pop with transients. Here’s a cool example of what I’m talking about.

You have a mix that averages -4dB. If a snare drum cracks to -2 you’re already topped out with your method while using a limiter. The mix won’t go over -2. If that snare hit happens one time, that transient stopped your entire mix from being louder. So you’re really almost 2dB under your target -2.

As far as prep, all my mixes are exported at -3dB. My rule of thumb for my clients is “-3dB or don’t send it to me.”

This gives me room to manually level a mix by hand while also removing all spiking transients so I can make everything louder the right way gradually. By the time I get done, all masters are processed at their highest bit and sample rate to end up at a final peak of -0.3dB. The reason for this is so mp3 conversion doesn’t clip. If you master at -0.0 the mp3 will clip and show errors. But for a wave file you can go to -0.0. Notice I say negative. Negative 0 will not show any clipping and if you do it right, minimum and maximum sample values should read exactly 32767. Only Bob Ludwig seems to nail that consistently. Lol!

Lufs should be between -8.5 to -9. With all that said, the loudness of a master depends on the song. If you listen to these crushed masters people are comparing, all of them have lost the crack and delivery of their snare drums when all the music is firing at full band levels. You don’t EVER want that. So always listen to the loss of snare crack as one of your determining factors as well as overall mix pumping and breathing.

Next, your -2 model is definitrly your problem. You can mix out at that level, but to master there is definitely too low. At the end of the day, here are my final thoughts.

We live in a world that listens to our 1000 hour mixes, tthrough esr buds. The only people with good listening gear are people like us. CD players are gone in new vehicles. Everyone plugs their phone into the aux in their car stereo and guess what…that destroys the audio even further. My fiance bought this thing for her phone that puts the voice through the speakers and allows you to play your iPod and phone into it to listen to music. It plays back in freakin mono!!!

So in closing, don’t stress about what others are doing. So what if you have to turn up your volume a little more. If you have retained dynamics and have shared your art with the world and you like your mix, win/win! But definitely think about some of the things discussed in this thread. They are all GOOD things to consider. Just remember though, don’t bank on numbers and graphs. Always let your ears be the judge and do what sounds best for YOUR music.

I’ve been at this 100 years and will be dropping a 15 song album with a credible label. They loved my production and don’t want to change a thing. That made me feel great. My songs aren’t as loud as major label releases. Why? Because I think major releases are horrible with artifacts that depreciate the music. I’ll never be a follower. I’ll always just be me and do what I feel is best for my music. :+1:

Danny

3 Likes

I get round that by sending the snare to an output other than the master 2 bus. It does clip but of course you can’t hear the clip due to the fact it’s…well, a snare, Slap a peak limiter on the resultant file after the fact and Robert’s your father’s brother.

1 Like