24 Bit vs 16 Bit - digital transfer of an album

[UniSol GR77]

Hello,

I've been told that - if you must digitally transfer an old analog production master tape with only 2 stereo channels (right and left) - you must do that at 16 BIT, because 24 BIT would not make any difference.

On the other hand, if you're working with a multitrack analog master tape, you must transfer each single track at 24 Bit, remix that, and what you get at the very end is a TRUE remastered and improved digital master of the album.

True?

At the end of the day, what you get on CD is 16 BIT anyway.

[bigshot]

The difference between 16 bit and 24 bit is added detail in the quietest passages, down well below the noise floor of your living room. So you aren't going to hear a difference in normal playback. But there are times when music is being mixed where quiet details need to be boosted, so 24 bit is useful for professional sound mixers. Once it's mixed, it's bounced out to 16 bit for CD as you say.

[Fiffy]

Originally Posted by UniSol GR77

I've been told that - if you must digitally transfer an old analog production master tape with only 2 stereo channels (right and left) - you must do that at 16 BIT, because 24 BIT would not make any difference.

On the other hand, if you're working with a multitrack analog master tape, you must transfer each single track at 24 Bit, remix that, and what you get at the very end is a TRUE remastered and improved digital master of the album.

True?

There is some truth to it, but when digitizing an analog source it can make sense to do that in 24 bit as well even if you don't want to mix multiple tracks. The advantage of 24 bit format is that it gives you some headroom for both recording and mixing. This helps to keep the noise floor low without having to record and mix "hot" (i.e. close to the maximum level), which has a higher risk of clipping. When repeatedly performing arithmetic functions (e.g. summation when mixing) on the signal, the added precision can also reduce errors. But once all that is done, downmixing to 16 bit is perfectly fine.

[ZoetMB]

Whether to record or master at 16-bit or 24-bit has absolutely nothing to do with whether it's multi-track, two-track (stereo) or mono.

The first question is what is the source material? Is it analog or digital? If it's analog, mastering in 24-bit will theoretically give better quality for the reasons detailed below. If it's digital and originally 16-bit, mastering at 24-bit gains you nothing. If it's digital and originally recorded 24-bit, then you should remaster in 24-bit although whether it's audible to anyone is highly questionable.

24-bit isn't just for lower-level sounds. When an analog source is converted to digital, there's a process called quantizing. When mastered for CD, the sampling rate is 44.1KHz. That means, the voltage of the signal is sampled every 1/44,100th of a second. In 16 bit, you can have 65,535 different values. So for purposes of this discussion, let's say that 65,535 represents 65,535 microvolts. And a given sample measures 32,423.5 microvolts. That 32,423.5 microvolts gets changed to 32,424 microvolts. That's called quantization error.

But if you go to 24-bit, instead of 65,535 different values, you can have over 1.67 million values. So you have much higher resolution in the voltage domain and less quantization error.

Most CDs today are mastered at 24-bit and then down sampled to 16-bit for CD according to the RedBook standard. I happen to own a standalone CD-recorder that can record at 16, 20 or 24bit and at sampling rates of 44.1, 48, 88.2, or 96KHz. The Nyqvist theorem states that you must sample at twice the highest frequency you wish to record. So 44.1KHz gives you a frequency response to 22KHz. Except for some children, no one can hear beyond that. In any case, I was very excited when I bought this recorder because I thought I was going to record at 96/24 and have incredible quality. Couldn't hear any difference whatsoever.

There are some other factors, such as the addition of "dither", but these are the basics.

[Fiffy]

Originally Posted by ZoetMB

24-bit isn't just for lower-level sounds. When an analog source is converted to digital, there's a process called quantizing. When mastered for CD, the sampling rate is 44.1KHz. That means, the voltage of the signal is sampled every 1/44,100th of a second. In 16 bit, you can have 65,535 different values. So for purposes of this discussion, let's say that 65,535 represents 65,535 microvolts. And a given sample measures 32,423.5 microvolts. That 32,423.5 microvolts gets changed to 32,424 microvolts. That's called quantization error.

But if you go to 24-bit, instead of 65,535 different values, you can have over 1.67 million values. So you have much higher resolution in the voltage domain and less quantization error.

Which is just a complicated way of saying that the 24 bit signal has lower quantization noise and thus higher dynamic range. Higher bit depth is mainly about having some headroom for recording and mixing without sacrificing SNR.