After recording a track at 192kHz I downsampled it to 96kHz using the Tracks > Resample menu option. I then exported the file as a 24 bit WAV. When I open the file in another program, it still shows the sample rate as 192kHz, and the file size is nearly identical to the original 192 WAV. I’ve opened the supposedly 96k file in Spek and dbPoweramp, both report 192kHz. Am I doing something wrong? Using the latest version of Audacity (2.1.0) on Windows 10.
The sample rate of exported files is set by the “Project Rate” (lower left corner of the main Audacity window).
Just wondering, why don’t you record at 96 kHz?
I’m recording vinyl rips, I would like to keep the original 24/192 files just to have the highest quality. For most purposes, 92 (and below) is going to be fine, but if I can record at 192, I figure why not. Hard drives are cheap. You can always compress and downsample, but you can’t add information. Figured I’d keep 24/192 “master” files.
When I tried to change the Project Rate sample rate, it ended up slowing down the audio and lowering the pitch. How do I avoid that? Also, Spek shows a difference between the 2 files, even though it says both are 192k. The file sizes are almost exactly the same (he 96k file is only 1kb bigger).
Scratch that, I try changing the Project Rate to 96k again and it worked (not sure what happened with the pitch changing earlier). I’m all set now, thanks.
if I can record at 192, I figure why not.
One problem people run into is the machine not being able to keep up. Those numbers produce a very serious data rate, and besides, where did you find an Analog to Digital converter which would do that with an RIAA preamplifier built-in?
44100, 16-bit, Stereo is Audio CD quality and will carry you beyond audibility in all four directions. People capture musical performances at 96000, 24 bit Stereo, not to get the higher quality (nobody can hear that, either), but to make the music more robust in post production. Anything above that is pretty much overkill and can lead to data transmission problems.
It’s a given that all these rates are very much wider, deeper, taller and better quality than any vinyl record ever made.
Who made your cartridge, arm, turntable, and cleaning system? Sheckles poured into those areas will yield significant benefits way before increasing the sample rate.
I have a Grado cartridge mounted in a Shure SME arm (the one with the ropes) mounted on an Empire belt-drive turntable on a custom shock base. The preamp is the one built into a Hafler HD101 preamp.
How did you do?
Koz
Well, to be honest my setup is quite budget. Turntable is an Audio Technica AT-LP-120-USB (about $250 on Amazon) with everything stock (cartridge is AT95E). I use the built-in preamp (there is a switch that goes between phono and line). The analog to digital converter is simply the one in my on-board sound card, a Via HD VT2020 (hey, it’s called Via Vinyl, it must be good for recording vinyl, right? 
). As for a “cleaning system” I use an RCA brush and washing solution sold cheaply on Amazon.
My CPU and RAM are more than enough to handle recording at 24/192 (an AMD 8-core 4GHz job with 8 GB RAM), while doing multiple other tasks.
I’m perfectly willing to accept that the cheapness of my setup makes recording at 24/192 overkill (and possibly detrimental). So the question is, what isn’t overkill? My turntable has a USB output that has a limit of 16/48, but I decided to bypass that and use the line out of the turntable connected to the line-in of my soundcard (which supports up to 24/192). I figured, the higher the numbers the better. Guess I was wrong? How do you recommend recording vinyl with this setup?
Thanks for taking the time to edit my pic with that info. That’s a lot of noise and distortion. So, you’re saying 192 is overkill? Which part(s) of my setup (as mentioned above) do you think are most responsible for the noise and distortion? How do you recommend recording vinyl with my setup? I guess the good thing about the noise is that it’s mostly in frequency ranges that aren’t audible, to humans at least. 
Play that for a dog, and I wonder if he would go crazy. Unfortunately, I don’t have one to experiment with. 
Have a read if this workflow from the Audacity Manual: http://manual.audacityteam.org/o/man/sample_workflow_for_lp_digitization.html
it’s part of this tutorial set: http://manual.audacityteam.org/o/man/tutorial_copying_tapes_lps_or_minidiscs_to_cd.html
When I transcribed my LPs I recorded with Audacity set at the default project rate of 44100 Hz and 32-bit float sample format and exported to 44100hZ 16-bit PCM stereo WAVs. This produced excellent results and I listen on fairly high-end (albeit antique) QUAD ELS-57 electrostatic speakers which are very detailed.
The best addition to Audacity I used following a steer from Koz was to purchase Brian Davies’ excellent ClickRepair - it cost me 40 bucks but saved me loads of time and produced excellent results. See this sticky thread on the Forum: https://forum.audacityteam.org/t/click-pop-removal-clickrepair-software/1933/1
WC
Or if you crank up the volume - sufficient to blow your tweeters 
It’s “a lot” in terms of the amount of inaudible bandwidth that you are recording, but not really “a lot” in absolute terms because it is all fairly low level, and as you say, “inaudible”. Yes, 192 kHz sample rate is overkill for normal audio (someone wrote into the forum a while ago who was making marine recordings where the some of the sounds were going up to extremely high frequencies, so not overkill in his case).
The thing is that audio recordings are always “band limited” (the frequency range is limited) in some way. Usually audio recordings are band limited to about 20 kHz (20,000 Hz), which is the extreme high end for humans with “perfect” hearing. The original vinyl recording would have been band limited, either intentionally with filters, or for very old recordings by the limitations of the equipment. Even the microphones used to record the music don’t go much above 20 kHz, so even if we disregard the question of whether anyone can hear frequencies of 30 kHz or more (which they can’t), any “signal” in the ultrasonic range cannot be said to represent the “sound” that was being recorded.
Normal audio playback systems are not be capable of reproducing frequencies up to 96 kHz. The upper limit for the playback system is likely to be around 20 kHz. The question is, if you feed frequencies in the 20 tp 96 kHz range into a domestic audio system, what happens to those frequencies? In practice, they mostly “disappear” - being filtered out and converted to heat. If the system does that “cleanly”, then it’s not a problem, but for normal audio equipment, 20 to 96 kHz is beyond the design range - the equipment is not expecting the input to go much above 20 kHz, so there is no guarantee that the equipment will handle these signals very well. Feeding ultra-high frequencies into amplification equipment that is not designed to handle such high frequencies can cause distortion frequencies that “bounce back” to within the audio range, and that is clearly not a good thing. Better to feed the equipment with signals in the range that it is designed to handle.