The questions people actually ask.

In Brad Huett’s view, after a lifetime of buying gear he did not need: start with what costs nothing.

• A free music program. Mac and Windows both have free ones, and the free ones finish songs.
• The headphones already in the house.
• The four free tools on this site, which show a mix what it is doing.

Buy nothing until one piece of music is finished start to end. A finished song is the proof that the next purchase will get used. That is an opinion, and it is labeled as one.

The tools on this site run without an account. What they do with the audio is in the rulebook, in numbered statements, and the build is checked against them:

Your audio never leaves your machine. It is decoded and analyzed in your browser, and it is never transmitted.

An account is optional. It asks for one email address and sends one link; there is no password, and signing up adds nobody to a mailing list. We cannot speak for other sites where music gets posted, only for this one.

If it is a laptop or a desktop, almost certainly, for the tools here: they run in a normal web browser, with no software to install. If it is only a phone or a tablet, no. Those can play the song back; making it needs a computer. On a phone the tools’ examples still run, so a young artist can see what they do, but the work happens at a desk.

In Brad Huett’s view, “this” is two different costs, and only one of them lasts.

• The gear. This is where the money pit is: equipment bought before the skill to use it, which ends up in the garage.
• The learning. This is what you are really paying for: a young artist who can walk into a friend’s studio and pull his weight, and who, the day he works with a real producer, can talk like a composer and a musician instead of a customer.

How the money pit usually happens:

• Buying on advice that is years out of date.
• Buying for the studio you picture, not the one your artist needs this year. For fun, for demos and for release are three different rooms at three different costs.
• Buying new gear before there is a finished song to justify it.

The one thing you cannot skip is a laptop. A phone or a tablet is enough to play the song back; it is not enough to make it. Your young artist is a creator now, not a viewer, and a creator needs a computer recent enough to run a free music program well.

Everything after that waits. Start with the free tools. Paid tools come once your artist has shown he can use the free ones, because a paid tool makes no sense to someone who has not yet learned what the free one does. So spend after a finished song, not before, and let each purchase answer a problem your artist has actually hit.

Headphones tell you one story about your mix; a car, a phone speaker and a room each tell a different one. Engineers call it translation. The usual causes are low end that only the headphones can show you, and a mix that leans on being loud.

• AudioVibes measures the mix: loudness, true peak, how the two sides relate, and what happens when the mix is folded to one speaker.
• ABA Compare plays two versions at the same loudness, so you hear the mix and not the volume.

Because the released song was mastered to a loudness, and yours has not been. The catch is that louder also sounds better to almost everyone, for a few seconds, which is why a louder mix usually wins a comparison it did not deserve.

Read the number first. AudioVibes reports your song’s loudness and true peak, so you know where it stands before you push anything. Then compare at the same loudness in ABA Compare and see whether the louder version still wins.

They are sharing the same space: the same frequencies at the same moments. Neither one is wrong on its own, which is why it only shows up together.

ABA Pocket Check shows where two to four performances sit in frequency and in time: where each one owns its space, and where they share it. Shared space is where one performance can hide the other. Your ears decide whether it does.

Usually because two recordings of the same sound are slightly out of step. Each one sounds right alone. Added together, parts of them cancel, and the low end is the first thing to go. Phones and small speakers play in mono, so that is where you hear it.

• ABA Phaze Check measures how two pickups of the same sound line up in time and polarity.
• AudioVibes shows what the whole mix does when it is folded to one speaker.

A tool cannot listen. It can measure, and tell you which measurements are out of the range a well-made record in your genre tends to sit in. AudioVibes does that and names the findings. For the worst ones, the free pages of the Repair Book say what to do about them.

The ranges are declared rules of thumb, written down from experience, and the site says so. The measurement is a measurement. The score is an opinion with the reasoning shown.

Your room and your monitors are telling you something the car cannot hear, usually in the low end, and the mix has been balanced to them. The fix starts with knowing what the mix is actually doing, not what the room says it is doing.

• AudioVibes measures loudness, true peak, the relationship between the two sides, and the fold to mono.
• ABA Compare holds two versions at the same loudness so a change is judged on the mix, not the level.

You have lived with this song for months. You can hear it finished. The recording still isn’t it, and each replay shows you the gap.

You can do it yourself, and for as long as the result is the one in your head, you should. A producer is for the day it isn’t: somebody whose whole job is the gap between the song in your head and the one on the recording, who hears it from outside, settles what the song needs before anyone records, and gets that out to the talent and the engineer. If the song is not settled in your head yet, he is guessing and you are paying for the guesses, so the song is settled first.

Here you see the price of each step before you pay for it, and you only ever pay for the next one. The first step, Milestone 0, settles what the song needs, how it will be heard and who plays, before anyone records a note; if you go on, it is credited in full. If the song isn’t settled yet, the Muse Chart Room is the step before it.

Do it yourself for as long as the result is the one you want. The day it is not, the question is what a studio charges for, and whether you can see it before you commit.

Here you see the price of each step before it starts, and you only pay for the next one. The price sheet shows where a record starts: the kickoff, the rhythm section and the finished-mix pass, priced line by line, three ways in. The arrangement, solos and vocals are the steps after it. Engineering that runs over is the studio’s cost; anything that comes in under comes off your bill. If the budget is short, work comes out. Nothing is ever discounted into something less. You decide what the record needs, and you can stop at any step and still leave with a finished mix.

Two mics on one source pick it up at slightly different times. Each sounds right solo; together, parts of the sound cancel and the kit goes thin or loses its punch. It is hard to hear and easy to measure.

ABA Phaze Check measures how two pickups of the same sound line up in time and polarity, from the raw recorded tracks, and shows what to move.

Most often the vocal and another performance are claiming the same frequencies at the same moments, and the louder one usually wins. Turning the vocal up makes it sit on top; it does not make it sit in.

ABA Pocket Check maps where the vocal and up to three other performances own their space and where they share it. The map says where to listen; your ears decide what to change.

Not when you run out of ideas, and not when the tweaking stops. In an ABA production, a mix is finished when the producer, the client and the engineer agree that it is, and sign.

Working alone, you are all three, which is why it is hard to feel finished. Before you start, decide what “done” means to you. When you feel like stopping, measure the mix against that, not against how tired you are of it.

Here, you do. The client owns the stems, and the delivery at the end of a production is a listed set of assets, not whatever happened to be on the drive. The list is public.

Everything that decides what “done” means, before a note is recorded: what the record needs, how it will be heard, who plays, and what each task costs. The engineers in the threads who complained about endless revisions mostly traced them to one thing: nobody had defined done at the start.

ABA makes that first conversation a milestone of its own, Milestone 0, with its own close: each task and its hours at the rate card, in front of the client, signed before Milestone 1 opens.

The loop is not caused by revisions. It is caused by a finish line nobody drew. When the targets are agreed before the work starts and each stage is signed at its close, a revision is a measured distance from a known target, not an argument.

That is what the milestone gates are: a signature and a reconciled number at the close of each stage, and the next stage does not open until the last one is signed.

Take the level out of the argument. Play the two versions at the same loudness, blind, and let the client pick. The louder mix usually wins a comparison; at matched loudness it has to win on the mix.

ABA Compare does exactly that, in a browser, with the audio staying on the computer.

We do not publish a target number, and we would rather say why than pick one. The services adjust playback loudness, the practice keeps changing, and a number printed here would be quoted back long after it stopped being true.

What we do publish: AudioVibes measures integrated loudness and true peak to the published standards, and shows them next to a genre range that is a declared rule of thumb, not a measurement. Know your number, then decide.

Get paid by the stage, not at the end. A production that closes each milestone with a signature and a reconciled bill does not leave a large unpaid balance sitting on the last delivery, and does not bill the client for work that was not agreed.