This is something one would think I should post on Gearsluts/space or WeAreTheMusicMakers, but I have found those resources as well as Facebook music production groups over the years increasingly useless, but this group always comes through for any questions I’ve had in the past, so here it goes…
I mastered a song using a 24-bit/48-kHz stereo mixdown. When I rendered the track, I forgot to adjust the DAW’s (Bitwig Studio) preferences from 44k to 48k. So I re-rendered it selecting 48k and now I have 2 masters with same bit depth, but different sample rates. The 48k one sounds noticeable smoother whereas the 44k one sounds brighter and more fatiguing. Is this how it should be? I wasn’t expecting such a noticeable difference. I’m wondering if using a 48k file in a project initially set at 44k was introducing some distortion/aliasing making it sound less smooth.
My question is whether I should submit the 48k file to my digital distributor for streaming services (i.e. Spotify) and for my personal artist pages (Bandcamp, Soundcloud)? Or do I need to down-sample from 48k to 44k? I understand I don’t need to dither and can keep the file at 24 but as I’m not releasing it on CD (I hope I’m correct in this understanding as well). My digital distributor (Symphonic) has this info on uploading songs:
You may upload audio assets in aiff, flac, and wav @ 44khz / 16 bit stereo.
So If I need to down-sample, I see no option for that in my DAW. Can you recommend a free or cheap program for Mac? People mention Voxengo R8brain but that’s Windows only. I’m not looking for a complicated solution. I use MediaHuman Audio Converter, I’m wondering if that would suffice in converting a 48k wav file down to 44k? Please advise. Thanks!
Correct, but I already addressed this in my post. When I did select 44k render from a 48k audio file, it sounded noticeably worse. I’m unclear whether that option in the DAW means I can only use the same sample rate as what’s chosen in the drop-down menu or if it means I can down-sample with the option.
No I’ve never recorded in real time because often my computer isn’t fast enough after I select higher oversampling rates on some mastering plugins. Does recording in real time make a difference in sound quality?
Its marginal but yes, in my experience it does. Realtime recording sounds a little nicer. And you dont get any weird shit from conversion algorithms, Worth giving it a shot.
To be honest I would open the mastering project and simply re do it, in 16 / 44.1 . There’s really no benefit to using higher sample rates, once its published it is going to be played either as a 16/44.1 or converted by some horrible automatic thing anyway.
Yeah you’re right, but I just want someone to confirm if I down-sample using the above option in the image attached to my original post or if I need a 3rd party device.
my (fairly limited) understanding of these things is that if youre doing any downsampling then you’ll want to apply dither also to smooth over the difference. that may be why you can detect some difference between the two files. i dont use bitwig so dont know if it offers this as standard, but know that ableton does
not all sample rate converters/methods are created equally. some just truncate in a brutal way and others will dither with an algorithm/process which is sometimes proprietary.
most sample editors and DAWs have dither options.
i record at 24/44 and that’s what i send off to be mastered then it comes back as 44/16 and 24/44 and mp3.
edit: for clarity on a lot of digital audio questions you might as well get the book and forget the forums. you can find it used for wicked cheap and there’s probably a digital version/kindle/pdf etc
Welcome to the wonderful world of sample rates and all the hassle (and opinions) that come with it. Note that your soundcard / interface also has a samplerate so realtime resampling is needed to playback on different samplerates if the playback frequency differs from your source files.
You could try to compare Audacity and Ocenaudio (my favourite audio editor for MacOS) with MediaHuman Audio Converter: Both Audacity and Oceanaudio do resampling (and are free). Set your interface to 44.1khz too to make sure you’re not listening to the real-time resampling algoritm of your OS, DSP or driver.
Thanks. Ocean Audio is my default audio editor. It’s great. I’m wondering if it’s just better for me to use a 24/44 mix-down as opposed to 24/48. I chose a higher sampling rate because I am using so many plug-ins on the master bus.
discussed this with my publisher and my digital distributor…
and now, since i always produce in 48k anyways, all master files stay in 48k/24bit…
all streamingfarms have their own little procedure of converting…
and demand of lufs/loudness…so, in that jungle i always stay a little louder than they prefer and keep the overall dynamic range the way it suits the various trax best from case to case…
everybody has play level compensation turned on in their streaming farm apps anyways…
so let apple music, spotify and co do, what they do…
and whatever u use to dither down, they got THAT covered in some better way…
and my streaming tracks can all compare fine with the average others out there this way…
while yes…a 44k,16 bit master file is always a little more in ur face and also a little harsher…
so it really depends on ur free gusto and what genre u deliver…
if u still wanna win battles in the loudnesswar, u need damned great mastering and x tra dithering…and ending on 44k/16bit for sure…
whatsoever…48k/24bit worx as the basic source file for all streaming farms…
and on the last record store of the planet, bandcamp, it’s really all left to ur taste and what genre ur working on and in…
and there’s no real ear to ear, next to each other in direct continous chain comparison thing going on that much…
while usually, the average bandcamp consumer likes higher bandwidth…
and in a modern producing process, i hardly reccommend to work in 48k/24bit anyways, since with all that bounce in place and resampling procedures along the way, it’s the best compromise between cpu usage and that constant “little loss” of sonic redundance…
cause yes…in digital domain any copy of a copy of copy is like the original…in theory…but in real sonic wilderness, meh, not really…end of the day…
it’s truu for the finished thing…but on the way to that finish, many many many generations can flow down south on the digital river…
and 48k/24bit is covering that pretty decent…
A general rule of thumb mostly is to stay in at least 24/48 and do the conversion as a last step on your final mixdown when needed. This should give you the best results but it depends on the software and plugins you use and the source material (and samplerate of the source material). Some plugins also do internal up- and downsampling, or sound nicer on specific samplerates.
You were talking about some really noticeable differences between the 48khz version and the 44khz version. My experience is that most of the time it’s hardly audible, so I think when you resample your 48khz master with Ocenaudio (or tool of choice) to 44khz (instead of selecting 44khz in Bitwig) the difference will be minimal.
