Editing Audio
.jpg) |
There are a number of recording software packages that are widely accepted as industry standards for digital recording. While their specific feature lists may vary, the majority of these software packages share the same basic functions.
In most professional applications, tracks are recorded as linear streams of recorded material. In recording jargon, these are called takes. During the course of a recording session, it may be necessary to record multiple takes of a prompt to get the best performance.
Incremental File Names
Some packages include a feature that can make prompt recording more efficient. They can automatically increment file names based on a root file name every time the system goes into record mode.
Normally, a script is recorded in order, beginning at the top with the first prompt and working down the script, recording as many takes as necessary of each individual prompt. Once the entire script has been recorded, a file map that notes which takes correspond to which recordings should be created. See the section headed File Map in the topic Recording Audio.
Creating Individual Files
Another common feature that is useful when dealing with large numbers of prompts allows for the batch conversion of these linear stretches of recorded material into a more manageable set of individual files. By using this type of "strip silence" feature, longer passages of audio can be edited automatically into shorter sections and saved as separate files.
Typical adjustments set the characteristics of strip silence, which is determined by a volume threshold. Anything under that threshold is then removed or flattened to zero. Additional parameters may include adjustments to the minimum duration accepted as silence. Any pause shorter than the threshold is ignored. In this way, several minutes of prompts that were recorded with enough space between individual phrases can be automatically separated into sequentially numbered files.
Professional audio editors are accustomed to trimming their audio files tightly; removing any unwanted silence from the beginning and the end of the file. This practice is advisable with stand-alone audio files. However, for files that will be included in a prompt database it is not recommended. Leave at least 0.2 second of silence before and after the audio material. The prompt database software occasionally has difficulty creating alignments when the sources are too tightly truncated. This space is also useful when sliding alignment markers to make fine adjustments to the timing of concatenated prompts.
Designers should remove any smart quotes that may have been inserted into prompt text as a result of passing through Microsoft Word in the preparation of transcriptions. The prompt text must use ordinary apostrophes.
When importing .wav files into the database, try to maintain the original filenames in the wave column. If it should ever be necessary to reimport the file, authors will know what the original source filename was. Under some conditions, the database will create a placeholder name of REC001.wav. When this happens, it may become difficult to reconstruct what the source files was without returning to the file map or recording script to look it up. If the original file name is retained in the database, it is easier to reimport the file based on the name.