I recently bought a Sony PCM-D50 to record live shows (replacing my Sony MiniDisc recorder after many years of faithful service), and have captured some great stuff in the past few months. With the MiniDisc recorder, I never gave much thought to sample rates or bit depth, but now I’m trying to educate myself. I’ve been doing all of my recent recording at 96khz/24 bit (which some would say is overkill, I know), with the goal of using Audacity to do some basic editing and then burning the shows to CD at the highest audio quality possible.
I don’t want to destroy the quality of the files during the conversion from 96khz/24 bit to a lower sample rate/bit depth. I’ve slogged through tutorial after tutorial and understand that converting from a higher bit depth to a lower bit depth requires dithering. But I’m still a little foggy on the whole thing. Can someone walk me through the process of converting the 96khz/24 bit files to those that can be burned to a CD, explaining which settings should be used and when the settings should be applied?
For example, I know that a CD’s sample rate is 44.1khz. But at which bit depth should the files be exported for use on a CD? 16, 24 or 32-bit float?
On another note, the 96khz/24 bit files tend to be at a low volume level when transferred to the computer and need some boosting. Any thoughts on ‘Amplify’ vs. ‘Normalize’? What’s the best way to increase the volume without altering the quality of the audio?
Sorry if all of these questions have already been discussed and I’ve missed it. Some of the tutorials are a little over my head. Any help you can offer to a novice in these areas would be greatly appreciated!
Audacity will automatically provide dither when converting from 24-bit to 16-bit. You can select this in Preferences > Quality > High Quality Conversion. Choose “High Quality Sinc Conversion”, and for Dither, anything other than “None”.
For burning a CD, the files must be in 16-bit format. Some CD burning software will convert other formats to 16-bit on the fly (for example, Toast accepts MP3 files).
This wiki page explains in detail the difference between Amplify and Normalize: http://wiki.audacityteam.org/wiki/Amplify_and_Normalize
As for workflow:
- Go to Preferecnes > Quality and set the default sample rate to 44100
- Import your 96k/24-bit tracks into Audacity and edit them (including amplification).
- When you are ready to export the edited version, just select File > Export, then in the Export dialog choose “WAV (Microsoft) signed 16-bit PCM”. Audacity will do the conversion from 24-bit to 16-bit, applying the high-quality conversion with dither, and also do the sample-rate conversion.
Now on to personal opinions …
In my view the damage done by conversion from 96 kHz to 44.1 kHz is greater than the advantage gained by recording in the higher sample rate to begin with. If you could record at a sample rate of 88.2 kHz (twice 44.1) then the conversion is much simpler. When converting from 96 to 44.1 the software must interpolate samples, and this must necessarily introduce errors.
OTOH, recording at 24-bit depth is something I’d highly recommend. Since your final destination is a 16-bit CD, this gives you 8 bits of “headroom”. As you’ve noted, your recordings are quiet - this is good. When you amplify them (after you’ve done your editing), you will only be amplifying the (negligible) 24-bit quantization noise. When you convert to 16-bit this noise will be swamped by the 16-bit dither noise, thus giving you the highest-quality output. That said, it is almost a certainty that the “environmental noise” (crowd, air conditioners, street noise, etc.) as well as the microphone pre-amp noise in your recordings is much higher than the inherent digital (quantization or dither) noise in the system.
IMO if you can’t hear it (and especially if you can’t measure it), it’s not worth worrying about.
Thanks so much for the quick reply, Bill! Your comments are very helpful.
So what you’re saying is that, in the end, I would be better off recording at 44.1/24 bit? That would definitely give me more recording time . . . but then what, if any, advantage is there in recording at 96 kHz if it always has to be converted to be usable?
In theory recording at 96kHz allows “finer” analogue to digital conversion - in particular, more accurate recording of very high audio frequencies. Whether or not this produces a practical benefit to outweigh the disadvantages of 96kHz recording is a matter of much debate.