Does Converter Affect Sound?

Converting between audio file formats is occasionally necessary. However, it tends to be an unfamiliar process. MP3 is a famous example of irreversible compression, leading to lost quality. But do similar conversions always affect sound quality?

Audio file converters will often affect the sound quality, but not always. Lossy versus lossless conversion targets are the primary factor. Encoding methods, sampling rates, sample depth, and bit rates are crucial. Unfortunately, audio fidelity can never be increased after recording.

Maintaining the fidelity of the sound you have is often a significant matter while making conversions as needed. All that’s necessary is understanding the encoding of the target format.

Do Audio File Converters Affect Sound?

Audio file converters will either preserve or discard the information already present. This maintains or reduces audio fidelity, respectively. Any added information is noise, which is usually undesirable.

Different audio formats have their own data coding and features. Still, they must always be interpreted to become a stream of PCM (pulse code modulation), which is then converted by digital-to-audio converter (DAC) circuitry. The quality of the sound is determined by how well the original audio data stream can be reconstructed.

Lossless codecs use compression methods that allow the digital audio input stream to be faithfully reconstructed as long as your sample frequency and depth are sufficient. Conversely, lossy codecs will always discard information to save space.

Converting Audio To Lossless Formats

Your target sample frequency and depth are crucial when converting to a lossless format. Hopefully, your source is lossless or high-quality lossy, but this is not always possible. Your source file will depend on the recording software or the download source. If you aim to preserve the sound of your audio, always convert from the original.

When it comes to lossless formats, converting to them should never degrade fidelity if your sample depth or sample rate is too low. While all lossless formats are technically equally averse to losses, some of them still have crucial advantages.

The Effect Of Sample Depth On Conversion

Sample depth is the number of bits per sample. When using more bits per sample, the number of different values increases. This allows for the amplitude of a waveform at every point to be encoded with less rounding error, also known as quantization error. In other words, high bit depth allows for better approximation.

This means that quiet components of sound or more nuanced inflections are more audible. Loud sounds have a large amplitude. Soft sounds have a small amplitude. Thus the rounding error that occurs when trying to approximate a sound wave will affect quiet sounds far more than loud sounds.

Depths of 8, 10, 12, 16, and 24 bits are most likely to be encountered in files. CD quality (defined by the Red Book standard) is 16 bits and plenty for most pedestrian purposes. 24 bits is more than enough; it’s beyond the Signal-To-Noise ratio of most audio equipment.

Converting to smaller sample sizes can easily save significant chunks of space, throwing out the information in the least-significant bits. Increasing the sample size will mean the least-significant bits are either zero or noise (random, meaningless, hard-to-compress data).

The Effect Of Sampling Frequency On Conversion

The sampling frequency reflects the rate at which the waveform has been sampled. Taking more samples per second gives a higher-resolution impression of the original wave. This affects both the audio fidelity and frequency range.

When PCM is translated by a sound card’s digital-to-analog converter, the samples are interpolated (mathematically guesstimated) to infer the continuous change in amplitude. Interpolation must generate an approximate value over time, given the surrounding data. If there is less time between samples, the interpolation error lessens.

For a given sampling frequency (Nyquist rate), the maximum frequency (Nyquist frequency) of sound that can be recorded is half the Nyquist rate. Informally: the frequency of sound is the oscillations of a medium per second. One oscillation involves moving to one extreme, the other, and back. Thus to record a complete oscillation discretely, both extremes must be captured.

CD quality is 44100 Hz, allowing you to hear the Nyquist frequency of 22050 Hz, which allows the range of human hearing (roughly 20-20000 Hz) to be covered. Many lossless audio standards use sampling frequencies of about 40-50 kHz for this reason. 96 kHz recordings are becoming more popular these days, but it’s often difficult to notice the difference.

Lowering the sampling frequency when converting will effectively re-sample the wave at a lower rate, “ignoring” higher frequency information and degrading the fidelity. On the other hand, you will save proportionally more space.

Best Lossless Formats For Conversion

The best lossless format you should use is FLAC, unless you’re an Apple user primarily, in which case ALAC might be more convenient. WAV is popular too, with AIFF being the Apple equivalent, but expect WAV/AIFF to be almost twice as large as FLAC/ALAC as the latter are compressed. On the other hand, WAV and AIFF are supported more universally.

With all this in mind, if you choose a format such as FLAC with support for the appropriate sample depth and frequency, you can mostly avoid degrading the audio upon conversion. Often, online tools will not allow you to control this and may lower the frequency or bit depth to save resources. You can use local software such as FFmpeg and Audacity to preserve fidelity when converting.

Converting Audio To Lossy Formats

When converting audio to lossy formats such as Opus, AAC, or MP3, the sound will always affect the sound waveform. Irreversible compression is inherent to lossy formats, meaning they discard information that cannot be recovered. However, the amount of data lost varies, and lossy may still be the ideal choice for good-quality playback.

A good lossy codec keeps essential frequencies while throwing out the stuff that doesn’t matter as much, achieving small file sizes while minimally degrading the audio. While encoders can make a considerable difference, bit rate matters most, and the codec chooses how best to use those bits.

The Effect Of Bit Rate On Conversion

The bit rate of a lossy-encoded data stream is the amount of information (in bits) stored for a duration of music (time in seconds). The typical unit used is kbit/s (1000 bits per second). Higher bitrate allows the codec to store more information about the original waveform, thus preserving the sound better and taking up more space.

While bitrates vary wildly, 196 kbit/s to 320 kbit/s is a good range for achieving or nearing “transparency.” Transparency is reached where the lossy compression artifacts (deformities of the sound wave) are indistinguishable. At 256kbit/s (a commonly supported setting), a 4-minute track would be about 5.86 MiB, which is quite reasonable for music nowadays.

For reference, we can calculate the data occupied by uncompressed CD-quality lossless audio, such as a 4-minute rip in WAV/AIFF format:

  • 16 bits per sample, 44100 samples per second, two channels, 4 x 60 seconds
  • 16 x 44100 x 2 x 4 x 60
  • Equals about 40.4MiB

Best Lossy Formats For Conversion

The best lossy format for preserving quality across the range of typically-useful bitrates is Ogg Opus. Opus beats other codecs like AAC, Vorbis, and MP3 in terms of quality for a given bitrate, as it will alter the decoded waveform the least.

However, you may prefer to prioritize support. Like MP3, AAC is almost universally supported but is superior to MP3 and Ogg Vorbis in terms of (blind-tested) perception of quality. In addition, modern codecs for the formats above (such as qaac, libvorbis, and LAME, respectively) aren’t far behind Ogg Opus’s, so going with AAC instead may make more sense, especially for Apple users.

Conclusion

Sometimes, you’ll want to preserve quality as much as possible: lossless formats with sufficient sampling will be the ideal solution. However, for people who aren’t working in audio production, engineering, or manipulation, you will find that high-bitrate lossy formats will make the quality-size tradeoff efficiently without you noticing the difference.

REFERENCES

Similar Posts