Hi everyone, I’m looking for a plugin or a method to measure the actual loudness of tracks.

I’m not losing my mind, but let me give you some examples of why standard 'Integrated' or 'Short Term' values don't feel right to me. If you look at the Global Top 50 playlist, you’ll find tracks ranging from ultra-bright pink noise to tracks with a much darker 4.5 or 5.5 slope. Even with normalization active, these tracks sound very consistent when played back-to-back.

However, in my DAW, even if two tracks have the exact same LUFS values, they often don’t sound like they’re at the same perceived volume. This happens even when using meters with K-weighting or measuring short ranges.

I’ve noticed that K-weighting isn't perfectly precise; perhaps a meter based on the Fletcher-Munson curves would be more accurate? Does a meter with that specific sensitivity exist?

I suspect streaming platforms have some 'under-the-hood' sorcery to make everything—from amateur tracks to ultra-pro productions—sound so coherent.

I’m opening this up for discussion. If you have any plugin recommendations for measuring loudness more accurately, I’d appreciate it!

 

The Fletcher-Munson curve does not measure with dbFs but with db SPL.
You are not able to measure db SPL (the real loudness) objectively, because every listening environment is different.

And this is one proof for why a flat frequency response in your monitoring system is important - because only then no signal is emphasized.
Only then you are able to adjust the volume of your signals in relation to each other in the most senseful way.

Mixing becomes easy then.

 

Last edited: Apr 5, 2026

It just gets more complicated the more data you have...!

The best approach is to go back to the time before the DAW: how did they measure loudness in recording studios?

1. VU meter (Volume Unit Meter)
2. PPM (Peak Program Meter)
3. Experience and ear training

You're getting more and more lost by constantly overthinking it. First, you look at it, then use a microscope, then even more refined measurement methods... until you can't see the forest for the trees.

I think you're more likely a math genius, but you're lacking the basic understanding. Trust your ears and forget about Spotify and the like, because you have no control over whether they cut or boost the volume. As a sound designer, I only look at the 0dB limit in my DAW.

 

A-weighted F-M is older research (1930s), K-weighted LUFS EU (== LKFS International) is newer (2010s) and likely more accurate under all circumstances.

 

For adding examples: I have two tracks same integrated LUFS in the Hook (few seconds). But one is definetely louder because has more energy in the 1-5khz area. So a meter that indicate this kind of perceived loudness should be a great thing!

 

That's strange, because frequencies from 1k and above contribute more to percieved loudness and LUFS value. Maybe the problem is in the gating part of LUFS normalization. This video has some useful insights:

 

Maybe:

mastering the mix expose-2
https://www.masteringthemix.com/products/expose-2

F9 Spectral Match
https://f9-master-match.pages.dev/

Youlean Loudness Meter
https://youlean.co/youlean-loudness-meter/

Sonible
https://www.sonible.com/truebalance/
https://www.sonible.com/truelevel/

 

Last edited: Apr 3, 2026

pause for a moment and take that as point 1: it obviously works for stuff you hear on the net, but it does work for your own tracks.

so you might want to consider that it is not caused by the measuring technique itself - or at kleast not only.

no it won´t.

the filtering used in LUFS metering, RLB and k-filter shelf, might be only an approximation based on assumptions (it is!), but using such simple setups is the only way to make sure that it can become a standard.

some (even more inaccurate) measuring experiment from 1931 will not only not fix your issue, it is not even possible to implement such curves into a digital filter because it only a list of values and a handdrawn curve.

the latest scientific research in optimising that kind of stuff mostly delas with checking if certain spectral content appears together - but not with more complicated filters.

most of them are probably older than the specs behind LUFS. and there is a reason why you have never seen one anywhere.

they use the same stuff as everyone does. that is what standards are for and that is how and why they work.

if you record acoustic jazz guitar or of you make experimental electronica, you might end up with strange results when following the rule that you should master an album paying attention to the relative loudness bewteen tracks.

but seriosly, you just said you would hear it (that is sometimes not right sounding) - so the solution is to change the gain by hand, isnt it? that is, after all, how everyone had to do it before 2006, too, and it seems to have worked.

i belong to those who still differentiate between "audio mastering" and "technical mastering" - and see no reason to measure LUFS to apply gain to tracks for the former. you have ears for that.
the "LUFS" thing is for the whole album or the whole radio show and done a second step.

 

Last edited: Apr 3, 2026

10 years, 3000 posts, and you still don´t understand what loudness is.

 

There are two loudness prediction models that I know of that are way more accurate than LUFS.
However, I'm not aware of any plugin that uses one of them for loudness measurement.

 

Hmm... TVL and DLM use significantly more complex filter banks than LUFS. And they represent perceived loudness much more accurately than LUFS, although these models also have their own weaknesses, such as the standardization you mentioned. The problem with TVL and DLM compared to LUFS is the need to calibrate the system level to an absolute reference value (dB SPL / Sone). LUFS is relative, but unfortunately also relatively useless for measuring loudness.

 

My interpretation of this is that its the songs themselves / your mix. The two songs are not spectrally the same. If one has more energy in the upper mids, its going to be perceived as louder.

 

well, LUFS measures LUFS.

if you want to measure perceived loudness differences between track A and track B, you have to use your ears.

the problem is that people listen to youtube tutorials or software advertisement instead of learning how to research and how to work.

the main weakness of LUFS (as opposed to TVL) is probably that it uses the same algorithm for the three different things it can measure.

imho we would be misleadning people by recommending more and newer technology for or fix perception. mixing desks have gainsliders! use them!

 

F the LUFS. F the mix