Audacity Forum

I recently bought a turntable with internal AtoD converter output on a USB 1.1 port.
It samples 16 bits at 44.1 kHz or 48 kHz.

I want to keep this simple. I would like to store the Deck’s USB output.
Just the data it kicks out. No editing, no further sampling.
Ideally I would like to do this mostly in hardware.

What is the format of the USB data, and can I look at it?

How do I store the exact USB data?
- is there a half-length audio card which will do this mostly in hardware?
- how do I configure the audio card and Audacity to use Windows Audio Session API (WASAPI) to do this?


Kit:
Turntable: Reloop RP-2000 USB.
Computer: Asus MC51BC of 2013, 64-bit, Windows 10.
Audio Card: Realtek “High Definition Audio”, version unknown, a half-length card bundled with computer. Probably needs replacing.

turntable with internal AtoD converter

That's the first problem, right there. The turntable has preamplification, RIAA compensation and then only the USB sampling. So you're going through quite a bit of processing before the sound even leaves the turntable.

There's a trick you can do with Audacity. Go into Preferences > Quality and turn the dithering off. Then capture in 48000 (the better of the two) and Export WAV (Microsoft), 16-bit, Stereo. That should give you a bit-for-bit copy of what came in, but in common sound file format.

You are warned that you can't do anything to the work while it's in Audacity. Just in the door and out the door. Audacity works internally in 32-floating, not 16-bit. As long as you don't mess with it, the two conversions should cancel out. In normal entertainment sound editing, Exporting the work can cause 32-bit to 16-bit errors which the Dithering system hides.

The flavor of your posting tells me you want to do scientific or data quality work with Audacity which is an Audio Editor. That almost never goes well.

The work as it comes from the turntable is in "wav" format (sampling) but streaming, similar to digital broadcast. The Audacity Exported files will have the same format, but with data headers, tags, tails, file management, etc.

What's the goal? I don't mean "capture data accurate, yadda, yadda." I mean "This is a project for college digitization perception management class."

Koz

Under your sample rate under the Quality settings.

Leave the bit depth at 32-bit float. There are "technical" advantages to floating-point and converting to/from floating point is lossless. When you export to WAV (or FLAC) you can choose 16-bit.

Just the data it kicks out. No editing, no further sampling.

Windows drivers can re-sample without you knowing. It's done for compatibility so you can play/record at almost any resolution setting regardless of your hardware capabilities (just as you can display a high-resolution image on a low-resolution monitor without Windows complaining).

Don't worry too much about digital resampling. As along as you leave it a "CD quality" or better nothing and you don't clip (try to go over 0dB) nothing bad will happen to the sound. All of the "weakness" is on the analog side.

What is the format of the USB data, and can I look at it?

Digital Audio Fundamentals. You generally "look at" the waveform. i.e. At 48,000 samples per second (x2 for stereo) it's not helpful to look at the numbers.

How do I store the exact USB data?

You can save-as an Audacity Project and/or as WAV, FLAC, MP3, etc. A WAV file is simply a series of audio samples, preceded by a header that that gives the sample rate, bit depth, number of channels, etc., so the software knows how to reassemble the bytes as samples.

What format do you want? WAV & FLAC are lossless. Tagging (embedded artist/title/album/artwork information) is not well-supported for WAV, and not everybody can play FLAC. MP3 (and AAC) is universal, but it's lossy compression. But, a high-quality MP3 or AAC is often "transparent" (it can sound identical to the lossless original in a blind listening test).

- is there a half-length audio card which will do this mostly in hardware?
- how do I configure the audio card and Audacity to use Windows Audio Session API (WASAPI) to do this?

With a USB turntable you're not using your soundcard.

Thanks to you both for your replies.
I have been testing and may have found some losses through digital format conversion.

Setting Sample Format to 32-bit float rather than 16-bit worked a lot better than before.

Then I tried a simple test. Whilst recording in Audacity, I pressed the Record button On and Off repeatedly.
I found the sound played back with the Record button On was really not as good as that with the Record button Off.
How do we explain this?

Perhaps on my system the playback of 32-bit float is not as good as that of 16-bit unsigned.
But still I think this indicates that the 32-bit float recording may not playback as well as the directly stored 16-bit input from USB.
This is a small loss which I would prefer to avoid.

Perhaps my Windows platform is at fault, but it is only four years old.
Are my Audacity settings still not the best? I list them below.

I want to store my 16-bit USB input as exactly as possible, split into labelled tracks as 16-bit WAV files.
I would prefer to avoid format conversion.

Is there any way I can do this with current Audacity?
Perhaps a later Audacity will allow us to choose a single format throughout.
Can you suggest a simple 16-bit editor to complete my simple task?




My Audacity Settings…
In Audacity > Edit > Preferences > Devices, I used settings…
Interface:
Host: Windows WASAPI
Using: PortAudio 19.5.0-devel, revision unknown

Playback:
Device: Speakers (Realtek High Definition Audio)

Recording:
Device: Line (3- USB AUDIO CODEC)
Channels: 2 (Stereo)

Latency:
Buffer length: 1000 mS
Track shift after record: -130 mS



In Audacity > Edit > Preferences > Quality, I used settings…
Sampling:
Default Sample Rate: 48 kHz
Default Sample Format: 32-bit float

Real-time conversion:
Sample Rate Converter: Best Quality (slowest) – increased from Medium Quality (may I increase this on an unstressed 3.30 GHz PC?)
Dither: None

High-quality conversion: Best Quality (slowest)
Dither: None


My Kit:
Turntable: Reloop RP-2000 USB.
Computer: Asus MC51BC of 2013, 64-bit, Windows 10.

You can change the Quality Settings to 16-bit, but I would NOT recommend it.

I tried a simple test. Whilst recording in Audacity, I pressed the Record button On and Off repeatedly.
I found the sound played back with the Record button On was really not as good as that with the Record button Off.
How do we explain this?

I can't explain that, but like I said, The conversion from 16-bit (or 24-bit) integer to 32-bit floating-point and back is lossless. ...If you have $10 in the bank and you "convert" it to $10.00 or $10.00000 you haven't changed the dollar amount.* It's just math and I'm pretty sure your computer can do math!

Perhaps on my system the playback of 32-bit float is not as good as that of 16-bit unsigned.

I find that hard to believe and I doubt you'd hear a difference in a proper, scientific, blind ABX test. (It's easy to fool yourself in a sighted test.) But, that's NOT what you're doing anyway... You're not "playing" a floating-point file, you're allowing Audacity to temporarily save it in floating-point before you export it to WAV.

And, you're being over-paranoid about this... You are digitizing an analog record... 16-bits at 44.1kHz is far-far better than analog vinyl.** And, If you digitize it twice you'll get to digitally-different (but identical sounding) files. In other words, there is no such thing as "bit perfect" analog-to-digital conversion. There are two reasons for that - Every time you digitize you are sampling at different "random" (uncorrelated) points along the analog waveform. Plus, you've got analog noise and some if that is random and it changes the sample values by a random amount (the noise adds/subtracts from the sample value).


* Audio file conversion to floating point isn't exactly the same as adding a decimal point and zeros because 0dB is defined differently in floating-point (or in different bit-depths) and the numbers "inside" the computer are binary, not decimal.

** Some people prefer the sound of vinyl so we can't say that digital always "sounds better", but technically (noise, distortion, and frequency response) digital is superior.

The input to my PC from my USB record deck does not sound too bad.

I have tried Audacity and Vinyl Studio, which claims to capture USB 44.1 kHz 16-bit without alteration.
Yet the recordings of both apps do not sound as good and contain occasional judders which are not in the input and break up most of the bandwidth.

On my Windows 10 platform, I have not yet found a way of capturing the input, storing it on disk, and playing it back, to sound close enough to the original input.

Logically, it should not be difficult to store the USB input from the record deck.
Still, I have not found Windows software which can do it.
Perhaps this is deliberate.