Spoken Word Archive Project for BoD Meeting 23Jan03 The Spoken Word computer (DADA) was equipped with a better soundcard (CardD+) and a Marantz dual tape deck is connected to it. There were some initial issues regarding how to monitor the sound quality throughout the encoding process, but new speakers and a set of professional headphones ($150 total) cleared that up. Some testing was undertaken to determine what encoding scheme to use. The factors considered were quality of sound; size of compressed files; speed and efficiency of program; and the ability to create files in batches - unattended. The following formats were considered: MP3, SpeeX, and OGGVorbis. Though OGGVorbis offered excellent sound quality and Speex was a very fast working program, we decided to go with the MP3 format for now due to its widespread use and resultant compatibility with many other programs. As for encoders, we tried several, but found that the MP3 encoder plug-in for WinAMP (OUT_ENC.DLL), running in tandem with the Disk Writer plug-in (OUT_DISK.DLL), enabled us to process files in batches, while at the same time using WinAMP's EQ function to boost (pre-emphasise) the treble (high frequency bands) of the sound samples. A program called mp3Trim is used to normalise the audio files after being encoded to MP3. mp3Trim also has the capacity to process files in batches, and normalising the files at the encoded MP3 stage was much faster than at the raw WAV stage. After much deliberation the ingest process is as follows: 1. Record the analogue signals as .wav files using Cool Edit Pro (sample rate 44100, 16-bit mono) 2. Encode the .wav files into .mp3 files (at 24 kBit/s, 22,050Hz, Mono) using WinAMP, while simultaneously boosting the treble slightly. 3. Normalise the .mp3 files using mp3trim. File creation continues as well as work on a finding aid - a searchable version of the SW archive data records, which can easily be accessed from Studios A or B. Nathan Rambukkana, Spoken Word Project Leader 22Jan03