Books could be written on the topic of the longstanding debate about which audio format is the best. Nowadays however, the average audiophile has specific preferences: a certain audio hardware configuration, combined with stable digital playback software. But is there an audio format that is the 'best', or is it after all just a matter of taste? Let's dive into the already hotly debated discussion: FLAC or DSD or WAV?
When we think back to the days of Napster, MP3 was the way to go when it came to digital audio distribution: small files tuned for minimum download time and a small footprint on hard drives. Compression was key, but the audio quality suffered greatly. MP3s with bit rates of 128 kbps or lower are today almost unacceptable, but around the year 2000, this was the consumer standard.
Music lovers who are aware of the effects of data compression on audio quality value lossless alternatives. With lossless files (versus 'lossy' MP3, AAC, OGG and others) there is no audible loss in audio quality, compared to CD resolution WAV files. The FLAC format (Free Lossless Audio Codec), for example, uses sophisticated techniques to pack and unpack audio. The result is a reduced file size, great for online distribution and storage on local hard drives, without suffering from compression artifacts.
Within FLAC, there are various possible sample rates and bit rates, from standard CD 44.1 kHz/16 bit resolution, all the way up to 192 kHz/24 bit audiophile heights. Classical music benefits from an extended dynamic range and a more accurate waveform. The higher the sample rate, the higher the amount of musical data in a given time slot. Small nuances and instrument attacks are perceived much more clearly with high resolution files, like FLAC 192/24. And like DSD.
Developed even before the year 2000 by Sony and Philips as the format for the Super Audio CD, DSD has now also become a digital audio file format, freed from the physical carrier. DSD has a sample frequency that is much higher than that of the highest FLAC or WAVformats, with 'single rate' DSD clocking at 2.8 MHz. That's MHz, not kHz. The quantization however is only 1 bit, compared to 16 or 24 bits with PCM alternatives. The result of this single bit system is quantization noise, inaudible for the human ear. Interesting is the quasi analogue character of DSD, which is said to be coming from the distortions in the frequencies higher than 20 kHz.
Since WAV isn’t compressed, one file will be significant larger than FLAC. The size of a WAV file is related to its length, even silences will be decoded to WAV while FLAC recognizes a period of silence. A minute of WAV audio (stereo, 16 bit, 44.1 kHz) will take up approximately 10 MB, about 11 times the size of 128 k/bit MP3. Because of its relatively simple structure, WAV is easy to process, making it the main format in recording studios.
Although the stereo 16 bit / 44.1 kHz quality is the most common WAV format, it is also possible to use higher bit rates and sample rates. However, there is one downside to using higher qualities: the WAV format is limited to a maximum size of 4 GB due to its 32 bit file header. When a better quality is needed, one can use the RF64 format. This is almost identical to WAV but it uses a 64 bit header.