How loud is loud? How quiet is quiet?

[JPGarbarini]

Something I've noticed during my first attempts at mixing... different volumes sound different depending on where they're played...

Obviously something everyone experiences I'm sure [:)] so is there any convention out there people use while mixing to determine out loud/quiet a music piece should be? Or should I just set my computer/speaker volumes to medium and take my best guess at what sounds right?

[clarkcontrol]

Something I like to do is comparative mixing. I always A/B my mix against something similar from a major label that was (undoubtedly) mixed on a first rate system. This also helps me get balance, reverb levels, e.q., and panning without re-inventing the wheel, so to speak.

If I'm doing a film score, I use a CD soundtrack of the same style. Pop, Jazz, etc. all get A/B'd as well with CDs from their respective genre.

This is very important to me as my setup is not $$$ so this method really helps to make up for my budget speakers. Without exception I am always pleased with the results.

Of course you don't HAVE to mix things the same as a John Williams soundtrack, for instance, but if those people could afford the very best in the world, then I would say its worth considering very seriously.

Clark

[Jonathan Mitchell]

Clark is absolutely right in my view. I always mock-up a favoured recording to use as a template to mix my own stuff (particularly reverb). When I try and mix on my own without a mock-up, I think it sounds better and better, but when I come back to it the following day - ugh! However, when I mock up something to sound like Karajan, it sounds fantastics when I apply it to my work. By mock-up I'm talking about getting the gigastudio reverb and placement similar and also taking note of how the instruments are used, after using VSL to copy a classical recording. It's not quite the same as having the great man conduct it for you, but it's the next best thing.

Notes do sound different at different times. Remember that when you are orchestrating, you get a composite blast of everything, rather than a set of separate noises. I seem to recall people warning of mixing trombones and violins which apparently sound awful. Follow the advice you are getting - there are some seriously cool people on this forum - listen, use feel and be careful about using formulaic approaches.... oh, and do read Beat's tutorials.

[jbm]

...just a quick thought. Is it that you're getting different levels on different systems at certain frequencies (instruments, if you like) in your mix, or is it the entire mix that seems to change in level?
I've been setting up a little studio in my new place and have become marginally obsessed with room acoustics. From this I'm wondering whether it's a possibility that your room is doing this to you(?). (When you listen on different systems, I'm assuming they're not in your studio...) What you might try is to see whether different CDs played in your studio have similar balance problems, focused around similar frequencies. If so, then the problem could very well be your room. If you Google around a bit on the subject of room tuning or acoustic treatment you'll be amazed at some of the stories -- in some cases, even the smallest investment (DIY) and attention paid to room treatment seems to make all the difference in the world.

Just crossed my mind, maybe meaningless.

The other possibility is orchestration, as mentioned above -- and quite possibly a result of perceptual masking. It can be easy to do this with brass, in particular, because brass instruments have a very "full" spectrum. For this reason they can easily mask other instruments, basically by "chewing up" the spectral content in the listener's ears (okay, so that's a strange way of describing masking, but not too far off! Check out this link for a better description: http://www.zainea.com/psymusic.htm -- about half-way down the page. You'll notice that this also explains the "violins and trombone" problem.) This phenomenon can be even more tricky when dealing with samples, because there is a sort of acoustical mis-match between the spectral content and the perceived loudness (caused by the normalization process in recording/editing the samples -- two notes, performed in the studio at dramatically different levels, are tweaked into a 0 db peak). In certain cases, you might get a considerable improvement by simply knocking your brass down a dynamic or two, so that it uses a p or mp sample. The lower dynamics have less energy in the high-frequency range (i.e., less "brassy"), and will have less of a tendency to mask out other instruments, particularly those with "thinner" spectral content. If, however, you want the power from the brass, then you could support it with winds... and even _more_ brass!

You can actually check out perceptual masking yourself:

1) make two long notes (tied whole-notes, quarter = 60) in both VI-14 and PO-3 or TP (solo trombone).
2) place the violins above the trombone (a pretty natural decision!)
3) pick a basic "long-note" for the VI-14, and a "pfp" for the TP
4) make the two notes match in dynamic at the _start_ (so that the TP does a cresc/dim "around" the VI-14)

What you'll probably notice is a strange sensation of "dipping" in the violins as the trombone reaches a certain point in its cresc/dim. This is the masking taking effect! Cool... And what's even cooler about it is that this is happening in _your_ ears! (Don't know about you, but I find this pretty fascinating!)

cheers,

J.

[clarkcontrol]

J,

great points. I find that with samples even more than in an acoustic ensemble it is very easy to get that "shrinking sound" mix when just piling on the orchestration. Of course, the effect is different but the law of diminishing returns rears its ugly head quite early in the sampler world.

As a result, I find that I need to reduce any doubling to the absolute minimum, concentrate on distilling the arrangement to reveal linear movement, and really try to find the right sample to fill the intended "spectral niche."

Often this means re-orchestrating only because it doesn't mix well and is not believable. In Jonathan's case it could mean more closely following the choice of instruments (and the roles they play) in the mock-up. I rarely mock-up a "target recording" but I'm sure this method leaves a lot less to chance.

Another one of my tricks is to have a set of budget speakers in every room of the house, all linked to my studio. I can check my mix in the kitchen on a pair of $80 speakers, my bedroom on a set of computer speakers with a sub, and on a twenty year old portable stereo with 5'' speakers, etc.

The key to this being that they are in DIFFERENT ROOMS as well as being different speakers. I listen to my test stuff right alongside the original material, paying particular attention to the mid to bass frequencies, as acoustical balance problems that happen in these frequencies are the hardest to correct with room treatment. My solution was to have different rooms.

Clark

[Angelo Clematide]

@JPGIII said:

so is there any convention out there people use while mixing to determine out loud/quiet a music piece should be?

There are standards!

For a start four basic things have to be taken care of:

1) Calibrated monitoring in the mixing studio.

2) The level standard for a particular medium.
Media are: Music for Broadcast Video, Music for Compact Discs, Music for surround cinema formats etc.. Different formats have different level standards, and therefore are mixed different level wise, respectivelly have different standards for maximum Peak/RMS levels.

3) The level and balance withhin composition itself. The relativ dynamics of a single composition may stay the same for all medium.

4) Calibrated playback.

For visual comfort I would also recommend to use specialized level meters for the different formats, i.e. K-14 for pop music, K-12 for broadcast etc.. An Integrated Approach to Metering, Monitoring, and Levelling Practices: http://www.aes.org/technical/documentDownloads.cfm?docID=65

All this doesn't fully explain what is very loud and what is very soft in a piece of music played back from a media, let us look at that later...

.