Are there any approaches to gauge how well a sound may sit together with other sounds in a mix?

I’m thinking about something similar to picking instruments that might go well together in a band. Probably that has been done subjectively throughout the history of music, but I’m wondering if there are any objective measurements that might provide an indicator of a potentially good arrangement?

Here has been posted a question on complimentary sounds, but probably I’m not only thinking about individual frequencies but rather overall timbres that don’t compete with each other but rather complement to form a more interesting whole.


1 Answer 1


I am not a big music guy but I understand audio signal processing and acoustics to a decent extent so I shall take a crack at it. There are no definitive rules or purely objective measures for predicting how well certain sounds will blend together in a mix, as much of it depends on stylistic choices, the specific musical context, and/or the intended emotional response. However, several principles and techniques can be used to craft a balanced and cohesive mix, which can be seen as a kind of "objective" approach in the sense that it's based on general principles of acoustics and psychoacoustics. Here are a few:

  1. Frequency Spectrum Management: Each instrument occupies a specific range in the frequency spectrum. Too many instruments in the same frequency range can lead to a muddy mix. Using an EQ (equalizer) to cut unnecessary frequencies from each instrument can help to make room for others and create a balanced and better sounding mix.

  2. Harmonic Compatibility: This is about how the harmonic content of different sounds interacts. Two sounds may be more likely to blend well if their harmonic series align in a way that is consonant rather than dissonant. This is a complex area that depends on the specifics of the sounds involved and the musical context.

  3. Dynamic Range Management: If two sounds have very different dynamic ranges (i.e., the difference between the quietest and loudest parts), they might not blend well. Compression can be used to control the dynamic range of different elements and help them to sit together better in the mix.

  4. Timbral Balance: This involves considering the timbral qualities of different sounds and how they can complement each other. For example, a bright, harsh sound might be balanced by a softer, warmer sound.

  5. Perceptual Loudness: Different frequencies are perceived as being different loudnesses at the same amplitude, due to the way human hearing works. Tools like LUFS meters can help to measure and manage perceptual loudness.

For a different approach, machine learning could potentially be used to analyze large numbers of successful mixes and identify patterns that could be used as a kind of objective measure for predicting how well certain sounds will blend.

Do keep in mind that these are just tools and should not override the most important aspect of mixing: the human ear. Mixing is an art form and while these tools can help, they should not dictate the creative process. The ultimate judge of a good mix is how it sounds to you and your intended audience.

  • $\begingroup$ Thanks for the summary, Ahsan, this will push me in the right direction. I’m not aiming to replicate the art form of mixing, but rather looking for ways to automatically identify sounds that might go well together. Do you know of any software implementations of the above measures? - such as of “Harmonic Compatibility” and “Timbral Balance”? So far I’ve found this: link.springer.com/chapter/10.1007/978-3-031-09917-5_37 which seems to deal with whole songs rather than individual sounds, but I’ll have a better look at this publication. $\endgroup$ Commented Jul 26, 2023 at 9:33
  • 3
    $\begingroup$ @BjornThorJonsson I think you might have missed Ahsan's central point: There's no "objective" good-sounding combinations. What sounds good to a middle-European in a classical piece of music might sound terrible to a Persian, and might sound strange in a Thrash metal piece. What sounds dissonant in pop music might be the fundamental stone from which you build an Opeth song. I'm not sure what you're hoping to find there – this is not about specifically about mixing music, but about the fact that music has perceptive qualities that arise from the cultural expectations and the musician's intent. $\endgroup$ Commented Jul 26, 2023 at 11:24
  • $\begingroup$ That’s useful to keep in mind, though I’m not aiming to find an automatic way of identifying sounds that universally sound good together, but rather to find some way of (pairwise) stacking sounds against each other. Ahsan’s points are useful in that regard. I might e.g. go for the measure of spectral centroid, which is supposed to be “highly correlated with the timbre dimension brightness or sharpness” [ISBN 9781119890942, p. 43], or spectral flux? $\endgroup$ Commented Jul 26, 2023 at 23:58
  • 1
    $\begingroup$ yes, I get that, but "pairwise stacking" qualities are still 100% dependent on cultural context. You are (probably) used to middle-European post-romantic era chords. Like, for you pressing two specific keys on a piano sounds good, others not so much. That does not transport to other people. The tone ladder (or circle, or how ever you represent it) with its seconds, thirds, fifths, octaves (which are not quite frequency doublings…) is a matter of being exposed to it since baby days, not a quality of the frequencies or sounds themselves. $\endgroup$ Commented Jul 27, 2023 at 12:09
  • 1
    $\begingroup$ For example, you might find traditional Chinese zither-related instrumental music to be "shrill" at first, maybe in a bit of getting used to, right? There's probably very significantly more people on this planet that are used to these stackings of two simple sounds than what you are used to! (same applies to actual medieval and antique European music, but it's harder to research examples for how a "pedestrian" early-hellenistc string instrument actually sounded when played in a contemporary manner on youtube… or a pre-viking flute instrument, for that matter) $\endgroup$ Commented Jul 27, 2023 at 12:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.