Due

Objectives

• to read structured text files
• to write binary files

Reading

Introduction
Sound is created by air (or whatever) pressure changing over time. If you graph pressure versus time for a given sound, you may come up with a graph that looks something like this:

To represent a continuous curve digitally, we can sample it at regular intervals. At each sample point, we measure the pressure to the desired resolution and record the result as an integer. As we increase the sample frequency and resolution, we will digitize the sound more accurately. In picture above, the blue curve represents the original sound. The tick marks along the bottom show the times at which we take samples. The tick marks along the side reflect the resolution of our samples; the value recorded will be that represented by the closest tick to the actual curve. The red dots represent the samples and the white curve represents the approximation of the original curve that our samples give us.

If we want to create an audio file without an original sound source, we can synthesize sounds by coming up with what the pressure vs. time graph would be if the sound were actually played. For example, a pure tone has a sine curve for a graph. The shorter the period of the sine curve, the higher the tone. Higher amplitudes represent louder pure tones. For example, if the range of our samples is -128 to 127, then

0 61 108 127 115 76 18 -45 -96 -124 -122 -90 -35 27 83 119 126

Audio file formats
An audio file must contain information about the sample range and frequency in order to be played back correctly. Different formats specify different ways of recording this information. For this lab, we will read .pas files and write .au files.

.pas files are a text-based file format similar to .pnm image files (I made up .pas specifically for this lab). Each .pas file starts with a 5 line header. The five lines contain

• the string "S1" to identify the file as a .pas file,
• the number of samples in the file (counting a left and right value together as one sample),
• the sample frequency,
• the number of channels per sample (2 for stereo for the left and right channels), and
• the maximum value for a sample (32767 for 16-bit audio); the minimum value is assumed to be one less than the additive inverse of the maximum.

.au files are binary files that begin with 6 pieces of information, each written as a 32-bit integer. Common values are given below.

Assignment

You must complete the part of the AudioClip class that reads .pas files and writes .au files. AudioClip already has fields for the sample rate, format (which should be set to AU_ENCODING_PCM_16), and number of channels declared, as well as the declaration (but not instantiation) of a 2-D array to hold the samples. The rows of the array correspond to the time intervals and the columns correspond to the channels, so samples[10][0] would be the 11th sample of the left channel and samples[10][1] would be the 11th sample on the right channel. There are also some methods that manipulate the arrays to change the sample rate and resolution of the file.

Your code to read .pas files will go in the AudioClip constructor and the code to write .au files will go in the write method, which will make use of the RandomAccessFile class. In particular, you will need to use the following methods and constructor.

• RandomAccessFile(String name, String mode), which opens a file for reading or writing depending on the mode string. mode should be "r" for an input file and "rw" for an output file.
• void writeShort(int v) and void writeInt(int v), which write 16- and 32-bit integers.
• long getFilePosition() and void seek(long pos), which get and sets the current position in the file. This is what makes RandomAccessFiles random access -- you don't have to read or write data in sequential order. If you want to overwrite data already written to the file, you can use getFilePosition before you write the data the first time, write some more data after than, and then seek to the position where the original data was first written and then rewrite it.

In the constructor, you should

• open the file for reading by creating an instance of BufferedReader,
• read and ignore the first line of the header (or perhaps check that it is indeed "S1" and throw an exception if it is not),
• read the information from rest the header and save it in the appropriate fields or local variables (if the maximum sample value is 32767 then the format field should be 3),
• instantiate the samples array,
• use a loop that uses a StringTokenizer to break the lines read from the file into the samples for the left and right channels, and saves those values in the samples array.

In write, you should

• open the output file by creating an instance of RandomAccessFile,
• write the 6 values in the header with writeInt (the only ones that need to be different for each file are the 5th entry for the sample rate and the 6th entry for the number of channels; all the other values can be as in the table above),
• use a loop that invokes writeShort to write the data in the sample array to the file (each sample should have its left value written and then its right value), and
• close the file.

Neither the constructor nor the write method should throw checked exceptions.

Extra credit

• writes a message using writeBytes,
• saves the current position in the file as reported by getFilePointer,
• seeks to the 4th byte in the file to position the file pointer just before the data offset field,
• write the saved file pointer using writeInt, and
• seek back to the position reported by getFilePointer using seek.

Files

Exercises

1. Change AudioClip's write method to use writeInt instead of writeShort to write the sample data. Change main so it writes an .au file to posse_clip.au before processing the clip with resample or changeResolution. Listen to the resulting file in Windows Media Player. Explain what is going on.

Submissions