Hacker News Reader

Comment by anigbrowl | Hacker News Reader

Sorry to be pedantic, but Whiplash won the Academy award for best sound mixing; the award for best sound editing went to American Sniper. Mixing involves the spatial and tonal quality of the sound; editing the temporal and semantic quality.

But either way yes, it is very meticulous, and I think you'd be even more shocked to look at the project files, because each individual sound you hear in a film often involves multiple layers of sound or multiple treatments of a single source sound through different mix busses. As a rule of thumb there are bout 10 edits on the sound track for every cut you see on screen. Thankfully there are some tools available to partially automate the process, but in the it requires a huge amount of work and near-endless reserves of patience.

Source: movie sound is what I (mostly) do for a living, though I haven't won an Acamdemy award for it.

View on HN

Have there been blinded tests done to gauge audience impact? I hobbied a bit with sound and when I see, for instance, the result of "professional mastering" side by side, I've often thought the pre-mastering version was better. No doubt this isn't always the case, but it makes me wonder. Especially given how much myth and superstition seems to happen in audio, even coming from pros.

Not to come across as rude - I'm just ignorant and curious.

Not really. There are some loudness measuring standards and conventions about the 'right' level for dialog vs peak volume, but really it's down to the combined tastes of the mixing engineers, the director, and the producer, with the latter having priority. Thus there are fads about particular sounds (eg notice how every big exposion in a sci-fi film now is announced with the same falling sub bass sound...) and pressure to push the volume envelope on action pictures (Michael Bay is notorious for this sensory overload approach to both picture and sound).

"Mixing involves the spatial and tonal quality of the sound; editing the temporal and semantic quality."

What does that even mean? I thought sound editing was the creation of all the sounds and sound effects, and mixing was the mixing.

The same thing you said, though in a more abstract way. By temporal I mean where they go on the timeline, and by semantic I mean the content of the sound. Much of the time what you hear simply matches what you see on the screen, either from production sound or from a sound effects library or by a foley artist, eg if you see a glass bottle smash on the floor you will probably just use a natural sound for that. But then you have to look at what it means in the context of the story - suppose the bottle is being smashed on the floor by an alcoholic who has finally decided to fight his addiction? You migh choose to emphasize that with music, or you might emphasize it by layering in other sounds from elsewhere in the soundtrack or some purely expressionistic sound like a heavy lock opening. Also, you'll frequently use sound to speed up or slow down the action by transitioning into the next scene before or after the camera cut.

Mixing involves setting the levels of the different sounds, but it also includes some decisions about how the sound moves, what sort of reverb and coloration are used (spatial) and how it's balanced against the music, how it's EQed (tonal).

Hope that helps.

Sound editors are responsible what to include in the track and when, sometimes much more or less than what the mics on set picked up.

Sound mixers are like sound engineers in the rest of the business, crafting relative levels, dynamics, tonal transformations like EQ, and spatial transformations like reverb.

Wow! I was instantly sure that this came down to someone at some DAW making a few million little tweaks on individual drum hits... Thanks for the correction and the detail!

Oh I think it amounts to the same thing. Automation can help but it's fundamentally painstaking stuff. As a general example, imagine someone walking down the street, but you don't like the original recording for some reason and decide to replace it. All the footsteps have to be matched up and that's tedious. There are plugins that will match the peak sounds of one recording to another to save you aligning every footstep (or word if you're replacing dialog) but you often end up having to do things by hand. Then if it's one person alone the footsteps may sound too simple so you need to add some fabric swish for their clothes. Or if it's two people, you need to have two different sets of footsteps which is really really annoying. Three people walking is fine because nobody can keep up with that so you can completely fudge the timing and nobody will notice.

The big challenge in movie sound is not to distract from what's going on in the story, which often means throwing away sounds that are present in the real environment but with are excluded by the picture frame, and whose presence thus becomes confusing to the audience because they don't know where the sound is coming from. I hate shooting in restaurants for example, because most restaurants have very loud refrigerators and of course most of them can't be switched off for food safety reasons. When you eat dinner don't notice this because there's music, other diners, maybe the person(s) you went to have dinner with, and the sound of your own body chewing and swallowing - so your brain just filters all the background noise away and lets you focus on the conversation or whatever. But when you watch the same thing on a screen, your brain is like 'what's that big machine noise? why does the lead actress sound like a farm animal? why is it all so echoey?' Total nightmare :-)

I haven't seen Whiplash yet but I'm looking forward to it.

Reply on news.ycombinator.com