32-bit Float, Uninterleaved Format

All audio calls to and from Premiere use arrays of buffers of 32-bit floats to pass audio. Audio is not interleaved, rather separate channels are stored in separate buffers. So the structure for stereo audio looks like this:

float* audio[2];

where audio[0] is the address of a buffer N samples long, and audio[1] is the address of a second buffer N samples long. audio[0] contains the left channel, and audio[1] contains the right channel. N is the number of sample frames in the buffer.

Since Premiere uses 32-bit floats for each audio sample, it can represent values above 0 dB. 0 dB corresponds to +/- 1.0 in floating point. A floating point sample can be converted to a 16-bit short integer by multiplying by 32767.0 and casting the result to a short.


sample16bit[n] = (short int) (sample32bit[n] * 32767.0)

The plug-in is responsible for converting to and from the 32-bit uninterleaved format when reading a file that uses a different format. There are calls to convert between formats in the Audio Suite. For symmetry in the int <–> float conversions, we recommend you use the utility functions provided.

Audio Sample Types

Since 32-bit floats are the only audio format ever passed, there is no option of sample type or bit depth. However, file formats do use a variety of sample types and bit depths, so AudioSampleTypes define a variety of possible formats.

These formats are used to set members in structures passed to Premiere to define the user interface, and do not affect the format of the audio passed to and from Premiere.

PrAudioSampleType Description
kPrAudioSampleType_8BitInt 8-bit integer
kPrAudioSampleType_8BitTwosInt 8-bit integer, two’s complement
kPrAudioSampleType_16BitInt 16-bit integer
kPrAudioSampleType_24BitInt 24-bit integer
kPrAudioSampleType_32BitInt 32-bit integer
kPrAudioSampleType_32BitFloat 32-bit floating point
kPrAudioSampleType_64BitFloat 64-bit floating point
kPrAudioSampleType_16BitIntBigEndian 16-bit integer, big endian
kPrAudioSampleType_24BitIntBigEndian 24-bit integer, big endian
kPrAudioSampleType_32BitIntBigEndian 32-bit integer, big endian
kPrAudioSampleType_32BitFloatBigEndian 32-bit floating point, big endian
kPrAudioSampleType_Compressed Any non-PCM format
kPrAudioSampleType_Packed Any PCM format with mixed sample types
kPrAudioSampleType_Other A sample type not in this list
kPrAudioSampleType_Any Any available sample type (used by exporters)

Audio Sample Frames

A sample frame is a unit of measurement for audio. One audio sample frame describes all channels of one sample of audio. Each sample is a 32-bit float. Thus, the storage requirement of an audio sample frame in bytes is equal to 4 * number of channels.

Audio Sample Rate

PrAudioSample is a prInt64

Audio Channel Types

Premiere currently supports four different audio channel types: mono, stereo, 5.1, and max channel.

Greater than 5.1 channel support was originally added in Premiere Pro 4.0.1, with partial support for a 16 channel master audio track, only for importing OMFs and playing out to hardware.

In CS6, 16-channel audio export was added.

Starting in CC, the audio channel support is increased to 32 channels.

PrAudioChannelType Description
kPrAudioChannelType_Mono Mono

Stereo. The order of the stereo channels is:

  • kPrAudioChannelLabel_FrontLeft,
  • kPrAudioChannelLabel_FrontRight.

5.1 audio.

The order of the 5.1 channels is:

  • kPrAudioChannelLabel_FrontLeft,
  • kPrAudioChannelLabel_FrontRight,
  • kPrAudioChannelLabel_BackLeft,
  • kPrAudioChannelLabel_BackRight,
  • kPrAudioChannelLabel_FrontCenter,
  • kPrAudioChannelLabel_LowFrequency

New in CC.

kMaxAudioChannelCount, defined as 32 channels as of CC.

All channels use kPrAudioChannelLabel_Discrete.