Long post, please bear with me

I got this project from a client where the beat stems were at 32/48 but the vocals were at 24/44.1 .
So because of the pandemic situation, they recorded themselves . and it needed sone restoration. i use Izotope RX 7 for it. and it also has a decent sample rate conversation feature. i upsampled the recordings to 48khz. there was no audible distortions or artifacts. i was happy with it.
And opened a project with 48khz sample rate on Studio One .
But while mixing. i found it sounds harsher without really boosting that much high end. but i got around it. and made it sound smoother in the top end. Referenced it while mixing with ADPTR , comparing it with commercial releases . ( lossless files, also 24/48 )
it sounded decent. the top end sounded smooth.
but after exporting the track ( 24/48 ) , it feels like it sounds harsh on the top end. i didn't check if it sounds harsh if exported at 44.1 . i will tommorow. i am also thinking about changing the project sample rate to 44.1 . Studio one converts all the audio used in the project to the project sample if the audio sample rate and the project sample rate doesn't match . also I can't go back and use the original 44.1 files and convert the beat stems to 44.1 . because that would mean doing the restorations again on the vocals and i don't have time or will to do it again . but i am not really sure if its because of the sample rate conversation .
I also think it's because of the Studio One MixFX CTC-1 Console Shaper , that i use on every mix . while getting it, i saw some comments about it sounding bad in higher sample rates then 44.1 . didn't care back then.
i also need to double check if it sounds harsh in the daw or only when its exported. because pretty sure it sounded smooth before i was exporting .
i added the Mixdown that i exported. listen to it and say what you think. some might say it sounds fine in the top end. but i remember it sounding smoother in the daw and comparing it to the Commerical records that i admire the mixing , it sounds slightly harsher in the top end . and i personally like smoother top ends .
I don't usually export directly. i print the master out to another track inside the daw and export that . but it's still early on the mixing stage and i just wanted a quick demo audio and take it home and listen to it . so i exported it
Anyone know anything or had the same experience?

drive.google.com/file/d/1t6o9Q6G4WLwFtC7vrV7LHdlPFclyEP2_/view?usp=drivesdk

 

Last edited: Oct 3, 2020

So you didn't dither yet? You did a straight print?

 

i have a Pro L2 on the masterbus with dither on

 

i've always suspected the same thing as i work strictly 48khz , i'd usually spectral denoise in rx it sound much smoother on top end then when i imported back in Studio One it sounds harsher for some weird reason.

 

Do you use CTC or Console Shaper too?

 

Sounds Damn Fine here thru a yamaha o1v961 and 8 inch yamaha monitors

 

thanks alot. ❤️

 

thanks for the info, i am gonna try it for sure. and i found studio one works best on 44.1 all the time .

 

I had some time to test the SRC of Studio One 5.0.2 with or without CTC-1 enabled (also added Izotope RX8.1 test), i used a 96khz 24bit sweep test file downsampled to 44.1 at export, here are the results:

1. Izotope RX 8.1 SRC
2. Studio One 5.0.2 SRC
3. S1+CTC-1 Classic SRC
4. S1+CTC-1 Tube SRC
5. S1+CTC-1 Custom SRC

 

Plus two more tests of Reaper 6.13 SRC 96khz to 44.1khz, for comparison:

1. Reaper 6.13 SRC 192pt Sinc
2. Reaper 6.13 SRC 768pt HQ Sinc

 

Last edited: Oct 4, 2020

Decided to make 3 more tests, i resampled the 96khz 24bit sweep file to 44.1khz 24bit (In Izotope RX 8.1), and exported it through Studio One CTC-1 (project set at 44.1khz 24bit with the audio file 44.1khz 24bit):

1. S1+CTC-1 Classic 44.1khz
2. S1+CTC-1 Tube 44.1khz
3. S1+CTC-1 Custom 44.1khz

Conclusion: Yes CTC-1 does introduce aliasing, and the test is made only with a frequency sweep, not a complex audio file like a guitar recording for example.

 

as i suspected, it sounds better in 44.1 then 48 in studio one . now , should i stop using ctc-1? it's a part of my template now and i like it very much . i am pretty sure I'll get the same results without it, because in the end its my taste. but that's the whole idea of using something like that. to get the sound without doing to much

 

and thanks alot rootwits, thanks for taking the time to analyze this and help us that uses SO

 

This is how it should look like if the SRC filter is good:

1. Voxengo OVC-128 (Test file sweep at 44.1khz 24bit)

As you can see the harmonics introduced by the plugin stop at 20khz.

 

Last edited: Oct 4, 2020

You can use CTC-1 if you like the sound, now you know why the sound is harsher, this is not just for CTC-1, all plugins have diferent internal processing and they upsample/downsample the sound internally, some have better SRC filters then others it all depends on the manufacturer.

The bad news is that almost every plugin is like this so if the internal SRC filter is bad in every one of them the aliasing will be worse if you combine multiple bad plugins in the same project.

If you like the end result of what you are doing that's all that counts. It dosent matter if the signal is aliased or not.
I would recommend that your project sample rate is the same as the files, if you need to downsample or upsample use a good SRC like Izotope RX or Voxengo r8brain which is free.

 

Last edited: Oct 4, 2020

Good SRCs: Izotope RX, Voxengo R8Brain Pro, Voxengo R8Brain, Reaper, Ableton Live 9 and 10 [surprise!] ...

One can check up the SRC quality of the tools you use at https://src.infinitewave.ca/

However, as rootwits noted, sometimes aliasing can sound fun or good, but you have to know your tools, what they're doing. Not all sample rate converters are equal in all plugins. That's one of the reasons why I simply use mostly Voxengo plugins and Reaper at 192pt sinc conversion by default. I work @48/24, and doing the same thing nevermind the Reaper's very good SRC - converting all samples to 48/24 [or 48/32 even better] before using them in the song.

Such shame that Kontakt's SRC sucks pretty much... one has to watch out for aliasing when using NI Kontakt and Battery with 44.1 or 96kHz sample libraries. It would be good to convert them to the frequency you're working at.

Interestingly, speaking of SRC in samplers, my hardware ones all produce more or less aliasing... Sometimes it gives the samples that special sound you were looking for, but generally it doesn't sound nice. TAL Sampler plugin sounds most aliasing free using "variable" mode [like Akai S950 or Prophet 3000] of all the sampler plugins I tried, so I often use samples from Kontakt libraries with it, and the filter sounds better, too...

p.s. if someone has Prophet 3000 for sale, or even 2000, please PM me.

 

I don't understard how a company like PreSonus can't implement a good SRC in there software, and a company like Tokyo Dawn Records can and offer it for free, just look a this:

TDR Kotelnikov Compressor

Shame on you PreSonus

 

Especially since Aleksey Vaneev - Voxengo open sourced his SRC years ago. I don't understand why some companies just don't use his open sourced code for SRC? I'm sure he wouldn't mind. That's why he open sourced it, in the first place ffs.

SoX SRC is also very nice and open sourced.

 

Thank you for all the info guys.