Sample and Bit Rate Confusion


Need some advice regarding to digitizing my tapes.

I’m using Cakewalk UA-1G and Audacity. Basically my questions are about Sample and Bit Rates. For best possible sound what rates should I import from the UA-IG and what rates should I capture the file with Audacity be? For example should the rates be the same from the UA-G as what Audacity captures it with? Should I have a higher rate coming into Audacity or the same rate that audacity is set to capture it with?


“Bit rate” refers to compressed file formats such as MP3, AC3, Ogg, and so on. For best quality, don’t export in a lossy format - stick with WAV, AIFF or FLAC. (“WAV” is the default, and is almost universally supported by other applications, players and so on).

CD quality is 16 bit, 44100 Hz. “44100” refers to the “sample rate”.
The sample rate governs the highest possible frequency that can be recorded. At 44100 Hz, the theoretical maximum is 22050 Hz, which is higher than anyone can hear. The practical frequency limit for 44100 H sample rate is a maximum frequency close to 20000 Hz, which is about the limit for children with exceptionally good hearing.

48000 Hz sample rate is standard for video. If you’re making films/DVDs, then 48000 Hz sample rate is recommended.
44100 Hz sample rate is standard for CD and most audio players and is recommended for such purposes.

Okay, so far so good.

But what takes place when my UA-1g digitizes my cassette music at 44.1, yet Audacity records it at 96000? Should I set Audacity to 44.1 also?



Is there any advantage digitizing the original music at 96000 then recorded by Audacity at 96000 when it will be eventually turned into an iTunes lossless file?


If you record a 44100 Hz sample rate source with Audacity set to 96000 Hz, Audacity will “resample” the audio to 96000 Hz. This does not improve the sound quality, it just means twice as much disk space requirement. Set the “Quality” settings in Audacity to 44100 Hz sample rate, 32-bit float sample format (these are the default settings).

So then I should also have the UA-1G also set to 44.1? Yes/NO?

Or what if I set both the UA-1G to 96000? What then?


If you are making ultrasonic recordings of bats or other ultra-high frequency sounds, then you will need to use a high sample rate.
For recording audio for humans (frequency range up to 20000 Hz max), then 44100 Hz sample rate is enough. Using 96000 Hz sample rate “may” be a very “marginal” improvement to the “extreme” end of the audible spectrum, but, depending on the hardware, there is no guarantee of any improvement, and the sound quality could even be worse at 96000 Hz than at 44100 Hz. What is certain is that when recording at 96000 Hz, your computer has to work harder to handle more than twice as much data, and projects will use more than double the disk space compared to 44100 Hz recording.

My advice would be to set everything at 44100 Hz.

Is there any advantage digitizing the original music at 96000 then recorded by Audacity at 96000 when it will be eventually turned into an iTunes lossless file?

If you record at 96kHz (if you are not up-sampling) the only disadvantage is a larger file. (Except as Steve suggested, some analog-to-digital converters perform worse at higher sample rates.)

A 24-bit, 96kHz losslessly compressed ALAC will be larger than an uncompressed 16-bit, 44.1kHz WAV file. (ALAC & FLAC give you a file that’s about 60% of the original uncompressed size.)

But, it’s kind-of overkill for analog cassettes. “CD quality” is better than human hearing* so it’s pretty-much good enough for anything. And obviously, cassettes are worse than human hearing. Copying a cassette to super high definition resolution audio is something like copying a VHS tape to the highest Blu-Ray resolution…


  • The folks on [u]HydrogenAudio Forum[/u] have demonstrated that you can downsample a high resolution file to CD quality and you can’t tell the difference in a scientific, level-matched, blind listening test. In fact, a good quality MP3 or AAC file is often Indistinguishable from the high resolution original. (It depends on the audio content and the ability of the listener to hear compression artifacts.)

I assume you mean Roland UA-1G.

It is always best to set sample rates the same “everywhere”.

If you are on Vista or later, sample rates are also set in Windows Sound and should match there too. Mismatched rates (high or low) have some potential for crackles/glitches or speed problems.

The “simplest” configuration is 44100 Hz on UA-1G, 44100 Hz Default Format for Roland in Windows Sound (on both the “Playback” and “Recording” tabs), and MME Host in Audacity with 44100 Hz project rate.

If you want to experiment with different rates and don’t want to keep checking what Default Format is, make sure (below where you change “Default Format”) that both “Exclusive Mode” boxes are selected in both the Playback and Recording tabs. Then use Windows DirectSound or Windows WASAPI host. Audacity “should” then ignore Default Format and you just have to match Audacity’s rate with the switch on UA-1G.

Note this from the UA-1G Owner’s Manual:

If the UA-1G’s [SAMPLE RATE] switch is set to 96 kHz and the [96 kHz MODE] switch is set to
REC, you won’t hear the audio playback from your software.
If the UA-1G’s [SAMPLE RATE] switch is set to 96 kHz and the [96 kHz MODE] switch is set to
PLAY, you won’t be able to record into your software from an instrument or audio device
connected to the UA-1G.


24-bit, 96kHz is the sound specification they would use for Adele 's next album at Metropolis Studio in London. Works are recorded like that so they can be extensively filtered, produced and mixed down to a final album without losing any of her original acoustic quality. They also have Digital Audio Workstations that can keep up with fast, dense digital audio bitstreams like that.

Many home machines can’t do that, so while those numbers look grand in the advertisements, you’re usually far ahead ignoring them and going with more normal sound specifications.

Another note. MP3 and other compressed sound formats cause damage while they work and if you edit an MP3 in Audacity, the damage will get worse. So do everything in uncompressed WAV including the archive. If you need an MP3 copy for your Personal Music Player, then make the MP3 with the idea you will never need to change it. If you do, throw it away and make a new one from the WAV archive.


Gale, yes, it’s the Roland UA-1G

Thanks for the help folks. I’m setting everything at 44.1 and be done with it.