Making DVD Audio Discs from CD's
I have seen several forum posts regarding making DVD Audio from CD’s and/or the value of increasing the bit depth and sampling rate of .wav or .aiff audio to make 24 bit versions of the files. The prevailing wisdom seems to be that even though the resulting files have more data representing the audio the apparent quality won’t be any better because you can never improve upon the source material. In a perfect world this would be true. Unfortunately when it comes to optical media like CD’s and DVD’s there are other factors at work.
Working in a Hi Fi stereo shop for some years I began to notice that inexpensive CD and DVD players seemed to be better at reading scratched media than their higher quality counterparts. Upon doing some research on the subject I realized that cheaper players employ more error correction. This is because they use cheap transports and other cheap components that don’t read the media and interpret the data as well as higher quality players, and so more error correction is needed. In other words whenever you play a disc some of the data that is encoded is lost and either replaced by silence when it’s just a sample or two or dither noise when more than a few samples are lost. It’s this dither noise that causes the “harshness” or “digital grit” that audiophiles complained about when the CD format was first introduced.
High end audio manufacturers soon realized that by making more rigid transports and dampening the vibration caused by the spinning disc the laser could to a better job of reading the disc. They also found that by reading the disc more than once and comparing the results(oversampling), and taking steps to ensure that the samples play back in the correct order(jitter correction), less error correction will be needed and the sound is dramatically improved.
When you take the 16 bit 44.1 Khz file you extracted from the CD, resample it up to a 24 bit 96 Khz file and burn it as a DVD Audio disc, you are not actually improving the quality of the audio. However by having more samples there to represent the audio even if twice the samples are lost, there are still tons more left available to recreate the original audio. If you are a person that listens to music as an activity, and by that I mean you’re sitting the correct distance from properly towed in speakers in a room that has been optimized for listening to music there will be quite a noticeable difference. If you’re listening to music as you do housework, or listening on plastic computer speakers this is not something you even need to consider.
Working in a Hi Fi stereo shop for some years I began to notice that inexpensive CD and DVD players seemed to be better at reading scratched media than their higher quality counterparts. Upon doing some research on the subject I realized that cheaper players employ more error correction. This is because they use cheap transports and other cheap components that don’t read the media and interpret the data as well as higher quality players, and so more error correction is needed. In other words whenever you play a disc some of the data that is encoded is lost and either replaced by silence when it’s just a sample or two or dither noise when more than a few samples are lost. It’s this dither noise that causes the “harshness” or “digital grit” that audiophiles complained about when the CD format was first introduced.
High end audio manufacturers soon realized that by making more rigid transports and dampening the vibration caused by the spinning disc the laser could to a better job of reading the disc. They also found that by reading the disc more than once and comparing the results(oversampling), and taking steps to ensure that the samples play back in the correct order(jitter correction), less error correction will be needed and the sound is dramatically improved.
When you take the 16 bit 44.1 Khz file you extracted from the CD, resample it up to a 24 bit 96 Khz file and burn it as a DVD Audio disc, you are not actually improving the quality of the audio. However by having more samples there to represent the audio even if twice the samples are lost, there are still tons more left available to recreate the original audio. If you are a person that listens to music as an activity, and by that I mean you’re sitting the correct distance from properly towed in speakers in a room that has been optimized for listening to music there will be quite a noticeable difference. If you’re listening to music as you do housework, or listening on plastic computer speakers this is not something you even need to consider.
Comments
Post a Comment