Importing Raw or Headerless Audio

When an audio file's format header is missing, MP3 Encoder asks for the parameters it needs and imports the file anyway.

Why this exists

Most audio files start with a header that tells programs how to decode the samples that follow: sample rate, bit depth, channel count, byte order, sign convention. When that header is missing or stripped, the audio data on disk is still intact but programs refuse to open the file because they have no way to interpret the bytes. MP3 Encoder would normally have to skip such a file. Instead, it puts up the Raw Audio Import sheet, asks you to supply the parameters by hand, and imports the file once you confirm.

The most common case in the wild is a Sound Designer II (.sd2) file whose resource fork has been stripped. Sound Designer II is an old Mac audio format that splits the file across two parts: the audio samples live in the data fork and the format description (sample rate, bit depth, channel count) lives in the resource fork. Resource forks are a Mac-only concept and they don't survive a trip through most modern channels: email attachments, cloud storage, ZIP files made on a non-Mac, or any non-HFS filesystem all silently strip them. The audio is intact; the format header is gone. When MP3 Encoder spots SD2 files in this state, the Raw Audio Import sheet is what comes up.

MP3 Encoder also recognizes generic raw PCM dumps by extension. Drop a .raw or .pcm file and the same sheet appears. SD2 captures use big-endian signed integers (the format's fixed convention) so the sheet hides the byte-order and sign controls; generic raw drops show the full set, since those files can use either endianness and either sign convention.

The Raw Audio Import sheet

MP3 Encoder Raw Audio Import sheet, showing sample rate, bit depth, and channels popups, and a list of two raw-audio files needing import

If MP3 Encoder finds several files in the same drop with the same apparent shape, they're grouped into a single section and one set of pulldowns applies to all of them. Files that the heuristic disagrees about get their own section.

The controls

Sample rate - The original recording rate. The pulldown ranges from 8 kHz up to 96 kHz, covering the common rates: 8, 11.025, 16, 22.05, 32, 44.1, 48, 88.2, and 96 kHz. If playback comes out faster or slower than expected, try again with a different sample rate selected.

Bit depth - 8-bit, 16-bit, or 24-bit. MP3 Encoder pre-fills this by analyzing the audio data: it samples each frame's bytes and compares their relative magnitudes, since 8-bit data lives in a single byte and 16-/24-bit data spreads its energy across more bytes. The auto-detection is reliable on real recordings; trust it unless playback is obviously wrong.

Channels - Mono or Stereo. Pre-filled by another heuristic that compares the energy in odd and even bytes. If a stereo file is mostly center-panned, the heuristic can mistake it for mono - preview the file before encoding to confirm.

Endianness (generic raw drops only) - Big-endian or little-endian. Determines which byte of a multi-byte sample carries the high-order bits. Most modern raw PCM dumps are little-endian (the WAV-data convention), so the default is Little-endian. If a file plays back as a steady wash of static or radically wrong levels, flip this and listen again. Wrong endianness on 16-/24-bit data sounds dramatically broken, not just slightly off. Hidden for SD2 files, which are always big-endian.

Sign (generic raw drops only) - Signed or unsigned. 8-bit raw PCM is conventionally unsigned (sample values 0–255 with silence at 128); 16-bit and higher are conventionally signed (silence at 0). The default is Signed. If 8-bit playback comes out crackly or clipped, try Unsigned. Hidden for SD2 files, which are always signed.

Per-file overrides

Each row inside a section lists the file name and its duration. The gear button on the right lets you split that single file out into its own section if you need to apply different parameters to it than to the rest of the group. This is useful when one file in a batch has a different sample rate or channel count from the others.

Confirming the import

Click Import to accept the parameters and queue the files for encoding. MP3 Encoder wraps each accepted file on the fly into a temporary AIFF and feeds that AIFF into the normal encoding pipeline. The wrapped temp files are cleaned up automatically once encoding completes. Click Cancel to drop the raw files from the import; nothing is added to the queue.

Limitations

32-bit integer and floating-point PCM are not yet supported by the wrapper.