Sound Logo Tested on a Phone Speaker Before a Studio

The brief we work to is a second and a half through a phone speaker. Everything else follows from that. We write the mnemonic, then test it on the worst playback device your audience owns before you ever hear it in a treated room, because a sound designed on studio monitors and approved on studio monitors disappears in a train carriage at half volume. What gets written: A mnemonic between 1.2 and 2.5 seconds, and three derivations built from the same intervals rather than composed separately. A truncated stinger of around 0.8 seconds for six-second bumpers and app open states. An extended bed for the fifteen and thirty second cuts. And a set of interface cuts for the moments in your product that deserve a sound: a match made, a shift filled, an offer accepted. Marketplaces and apps have those moments and mostly spend them on a system default. How it is tested before you approve it: Through a phone speaker with a 300 Hz high-pass applied, which is roughly what a small transducer gives back. Truncated, to check that the first half still identifies. Played immediately after three of your competitors' current spots, in sequence, because distinctiveness is a relative property and testing it in isolation tells you nothing you can act on. And at the tail of a real cut of your own video, where most mnemonics turn out to be one beat too long. Delivery: Stems and mixdowns per target. Streaming at roughly minus 14 LUFS integrated with a minus 1 dBTP ceiling. Canadian broadcast at minus 24 LKFS, following the CRTC's adoption of ATSC A/85. A mono fold-down checked for phase cancellation, since a phone speaker is mono and a stereo trick that vanishes there is not a brand asset. WAV at 48 kHz and 24-bit, plus encoded versions at the file sizes your mobile team asked for. A short usage note covering where the stinger goes and, more usefully, where it does not. What we refuse: Jingles. A sung line with your name in it is a different craft, it dates faster, and it does not survive being cut to under a second. Library music dressed up as brand audio. If a competitor can license the same track next month, it was never an identity. Generative audio for the mnemonic itself. Commercial audio needs a documentable chain of authorship, and we will not hand you an asset whose provenance you cannot evidence if it is ever challenged. Rights are clean or we do not deliver. Composition and recording are assigned to you outright, the composer is credited and paid properly, and nothing leaves here carrying a residual claim. Not included: Voice casting and voice direction, full track composition for campaign films, podcast sponsorship packages, and audio for retail environments. A French voiced tag is quoted separately when it is needed, and often it is not, because a well-built mnemonic carries no language to translate. Who this is not for: Brands with no audio surface. If you do not run video, sell no app with sound events, and have no broadcast or podcast presence, you would be buying a file to keep in a folder. We would rather say that now than invoice for it. How the engagement runs: A listening session first, on your own inventory and your competitors', where the decision that matters gets made: what this brand must not sound like. Then one direction, developed properly, with the rejected sketches audible so you can hear the argument rather than take it on trust. Approval happens on a phone in a room with other noise in it. If it fails there, it fails, and we go again.