I see that the Audacity Manual suggests keeping the sampling rate at 44.1. I am wondering if any people here have experience recording and mixing at 88.2 or 96, with the bit depth at 32 bit float. Did things run smoothly, or did the higher sampling rate lead to glitches?
Bit depth at 32-bit float is highly recommended. This doesn’t really affect recording, but it does improve efficiency, quality and safety when processing.
Usually I record at 44100 Hz sample rate, which avoids unnecessary resampling later in the production process. I have used higher sample rates on occasion, and although I’ve not experienced any problems doing so, I’ve not experienced any benefits for ordinary audio recording either.
When working with very large projects, my computer is more responsive and processing is noticeably quicker when working at 44.1 kHz than at high sample rates.
One notable exception was when I was building a pre-amp for my electric violin. I made test recordings at 192 kHz sample rate so that I could measure the upper frequency limit and high frequency EMR interference rejection of the pre-amp. There was no problem recording at 192 kHz. (The pre-amp’s upper frequency limit turned out to be around 35 kHz, which I then reduced to 22 kHz so as to improve interference rejection.)
I wouldn’t recommend recording at 88.2 kHz. As far as I’m aware, there aren’t any A/D or D/A converters that operate internally at 882000 Hz, which means that the audio data has to be resampled on the fly when recording or playing. This may reduce the maximum practical track count for the project, and increases processing time, for no benefit.
For “audio” recording (sound for humans to hear), there is no advantage in using sample rates above 96 kHz, in fact very high sample rates reduce sound quality due to increasing non-linearities and increasing jitter. Here are some good papers (PDF format) by Dan Lavry about sampling theory:
Other papers by Dan Lavry can be found on his pro-audio website: http://www.lavryengineering.com/lavry-white-papers/
Thanks for the info, Steve.
I’m finishing up an album that I started on my Tascam DP-24SD, so I will be using 44.1 for all of those songs. I was thinking after that of experimenting with higher sample rates. 24/96 seems to be common for hi-res files on Bandcamp, which is where I post my music.
The main “benefit” of 24/96 is that a lot of people think that it is higher quality, so that’s good for your marketing.
In reality, the difference in sound quality between a well mastered recording in 24/96 and 16/44.1 is almost undetectable but files are 3 bigger, and may need to be re-encoded to a lower quality format for playing on mobile devices.
When streaming the audio from bandcamp, the audio is encoded in a lossy compressed format, so the sound quality is much less (in absolute terms) than 16-bit WAV (but should still sound very good).
Even if you final export format is 24-bit, it is still highly recommended to record as “32-bit float” as this offers quality and safety benefits during the production process.
My 10 year old daughter says that I can’t call myself a “rock star” until I get 10,000 downloads/streams, so marketing is very important. I’m currently just over 200, and my music is free! So I have a ways to go, obviously.
I understand the benefit of 32 bit float during the production process, so I will downsample to 24 bits as the final step.
Even if my music does miraculously get a decent following, I know that very few people these days are downloading files, and only a small proportion of those have audiophile quality systems where 24 bit vs 16 bit will make any difference. I’m mostly interested in producing music with high sound quality for my own personal satisfaction.