Hello there so super new to this I’m sampling off PC loopback for making sample based music and have a pretty simple question or request for advice might be better lol, when I’m recording say a drum break or a melodic lead what dbfs level on the meter (and by extension the linear scale on the side of the waveform) should I aim for ? I’ve been recording them to peak at -6dbfs
The second part of my question is if I plan to chop these drums up into individual one shots does that advice still stand , since for example the snap of the snare would most likely be the -6dbfs and everything else would be lower by a good amount as well as most of the track when the other drums are playing, I’m guessing that is the desired outcome anyway since they were mixed to that level by someone when the original recording was done anyway right ? ( Probably answered my own question that one lol) because if I tried to record each sound to peak at -6dbfs then I’d be battling the dynamics and volume level in regards to each other later anyway, probably better to leave them as is with the snare or highest part of the song being -6dbfs and adjust volume levels via faders in a DAW when mixing the final track ?.
Digital recording levels aren’t critical at all as long as you don’t “try” to go over 0dB and clip (distort).
-6dB is often recommended to leave headroom for unexpected peaks. With real drums the peaks aren’t very predictable and you might have to record lower. Pros typically record everything around -12 to -18dB.
But if you are mixing, mixing is done by summation. Analog mixers are built-around summing amplifiers, but analog mixers have a level control for each input channel plus a master control. Audacity doesn’t have a master-mix control.
-6dB is 50% so 2-channels can mix to 0dB depending on how the peaks line-up. With more tracks you’ll have to reduce the tracks more.
Internally, Audacity uses floating-point so it can go over 0dB without clipping, but regular WAV files, your analog-to-digital converter, and digital-to-analog converter is hard-limited to 0dB and 0dB should be considered the “digital maximum”.
One solution is to can export temporarily as 32-bit floating-point WAV without worrying about the mix-level. Then re-import and run the Amplify or Normalize effect to bring the level down (or up if it’s low).
Alternatively, you can use Track → Mix and then Amplify or Normalize the mixed track before exporting.
Awesome I appreciate your advice so basically just don’t hit 0dbfs and distort and I’m more or less basically fine ? I suppose with summation and mixing later I could always just adjust volume levels with faders later in a DAW if all the recordings are -6dbfs?
These recordings im doing are of already recorded drum breaks off of vinyl or streaming audio so I’m able to use the silent monitoring to set the volume level and limit it to -6db in this case would it be beneficial to record them at -12dbdfs to -18dbfs or just leave them at -6dbfs and adjust in mixing software later ?
One other question when pitching down i try to keep samples at 48khz rather than 44.1khz samplerate so I have a little more room above 20khz before I start getting aliasing. Say I was recording audio that was 16/44 would there be any benefit to recording it at 16/48 or 24/48 or would it be irrelevant due to the original quality when manipulating it later?
Same question if the original material samplerate was OVER 48khz if I recorded that at 48khz would there a huge noticeable difference in sound quality of the file ?
Right. You are fine. And if you are using a regular DAW you’ll have a master level control.
No. Digital levels are not critical. (Sometimes low analog or acoustic levels are a problem because with a weak signal you get a worse signal-to-noise ratio.)
Technically… you do lose resolution at lower levels (you aren’t using all of the bits) but it only becomes an issue at very-low levels and it’s less of an issue if you’re recording at 24-bits. (Low resolution shows-up as quantization noise and you’ll hear it if you make an 8-bit file.)
The only downside to high resolution (higher bit depth and/or sample rate) is bigger files. If your hardware supports it, there’s not much reason not to record at higher resolution (even if you are going to down-sample later).
…It can be hard to tell what your hardware actually supports because with any-old soundcard you can set Audacity to record at 192kHz and the drivers will happily and secretly up-sample. (And Audacity always converts to 32-bit floating-point because it makes DSP easier & better…) It’s the same with playback - You can play a high-resolution file on any-old sondcard.
In a proper blind ABX test, you generally can’t hear any difference between a high resolution original and a copy down-sampled to “CD quality” (16-bit /44.1kHz).
Every soundcard & audio interface has an anti-aliasing filter so you should never get aliasing.
BTW - Since mixing is summation, mixing effectively increases bit-depth resolution!
Two side questions if the source material I’m recording via loopback is 16bit/44.1 would I gain any benefit from recording that at 24/48? Any help with the noise floor (which I don’t think would be much of an issue at normal levels anyway) or any help when using pitch shifting down without shifting tempo which I think introduces artifacts or aliasing in samples ? Like would I gain any more room from 44.1 to 48khz for pitch shifting down before degradation began If I recorded a 16/44 signal at 24/48?
Also if I want to make sure I’m recording those at the specified bit depth and samplerate I have to set my DAC or soundcard in Windows device manager to the desired rate 24/48 In this case as well as setting the export and project depth and rate in audacity ?