Confused about DAW's sampling rate/bit depth

[Dave G 6]

I have several projects in progress, and suddenly I'm concerned about the corrsampling rate and bit depth. (FYI: I use soft synths and no external hardware. All my projects are composed within the walls of my PC.)

I've been doing this for a very long time and learning things as I go. Before I go on further with my projects, I want to be sure these values are set correctly.

Windows default sample rate is 48,000. I've read that for my music tracks, I can change this to 44,000, so that's what I did (also, ensuring that the DAW's sample rate matches that). The DAW's values seem to remain in place, not reset per project.

My question is: do I need to change the sampling rate and bit depth at all, or leave them at default? I've done some research but I fail to understand what I'm reading about managing these values. I'd like to use 44,000 and 24-bit, but I don't want to mess up anything before I finish the projects.

Please advise! Thank you in advance.

[azslow3 324]

Sample Rate is important. Any project  has fixed sample rate and it can't be changed. So, if your project is already 44.1kHz, you can't change it later. That is why the setting is called "Default for New Project", that is one time decision at the project creation. Also Cakewalk use the audio interface with current project sample rate (only). For several reasons, if you always use the same Sample Rate in Cakewalk better set the same in Windows. 

Bit depth of everything can be changed any time. And it can be different for different settings at the same time. Obviously the change has no influence on already recorded material, if you have recorded in 16bit, the file will stay 16bit after you change the setting to 24bit.

Cakewalk internally process in 32bit (floating point) or 64bit (if 64bit engine setting is activated). Normally render bit depth should be 32bit. 32bit engine data will be saved "as is" and in case of 64bit engine, 32bits  is still sufficient (extreme and artificial processing is required to produce any difference in sound between 32bit and 64bit files, such processing is never used in real projects).

But record bit depth should be as high as your audio interface support. Most interfaces support 24bit. Note that 16/24bit are fixed point numbers, audio interfaces always sample into fixed point numbers (even so drivers in some modes can transfer the data as 32bit and some interfaces claim they are 32bit). Real bit depth is around 20bit on good interfaces, extra bits are just noise (if you know SNR or Dynamic Range of your interface, you can estimate meaningful bit depth by dividing this number by 6). But every single bit is around 6dB, which you can have as a "headroom" during recording to get equivalent precision.
Note that record bit depth is not used till you record audio (probably the case in your setup with Realtek).

[John Vere 1,783]

Actually midi only projects can be changed at any time. I do this a lot. I used to be using 44.1 because it was for CD and my Yamaha o1v only used it. Now because I work with video I’ve switched over to 48. So I open old projects and immediately I convert them all to 48. I have a tutorial about how to do this. But simply you export all 44.1 audio as 48 . Delete all audio from project, change the sample rate rate of the project that now is only midi instruments and bring the new 48 audio back in. Done this dozens of times now. 
Just last night I opened an older backing track project that had no audio tracks and converted it from 44.1 to 48. 
To me 44.1 is for people who make CD’s. These days everything is video which is 48. And even though my projects are 48 I can still export them at 44.1 if I need a CD. 

[scook 4,774]

14 hours ago, Dave G said:

My question is: do I need to change the sampling rate and bit depth at all, or leave them at default?

May want to change the Record Bit Depth (currently 16) to be the same as the Driver Bit Depth (currently 24).

 

[David Baay 1,067]

25 minutes ago, Cactus Music said:

Actually midi only projects can be changed at any time.

This is corrrect. The inability to change sample rate after the fact only applies to projects with clips referencing recorded/rendered/imported audio files.

Technically "MIDI-only" refers to a completely audio-free project (no soft synths, no active audio tracks, and MIDI metronome), using the 'Internal' system clock for MIDI timing with MIDI tracks driving direct-monitored hardware synths (or onboard 'Wavetable' synth), but it's often used to refer to any project that doesn't contain audio clips as John is in this case. FWIW, the last time I tried it, CbB wouldn't actually run a true MIDI-only project with clock set to Internal; the transport wouldn't budge.

[John Vere 1,783]

30 minutes ago, scook said:

May want to change the Record Bit Depth (currently 16) to be the same as the Driver Bit Depth (currently 24).

 

Good catch. But if the OP is not recording any audio then I’m not sure it would matter?

Do soft synths output audio at these settings or do they  care? Good question.  Never gave it any thought. Except in the case of freezing the midi instruments if need be I’m not dead sure it matters?  

[John Vere 1,783]

@David Baay  interesting point about midi only being like old school sequencer projects that drives hardware. But all the ones I’ve converted had at least 4 VST instruments and it worked fine. This is now something to ponder about what is it exactly happens with audio from soft synths?  I think all the ones I’m using are sample based. 

[scook 4,774]

3 minutes ago, Cactus Music said:

Good catch. But if the OP is not recording any audio then I’m not sure it would matter?

No, as long as there is no audio being recorded at all, it does not matter but it does not hurt to be prepared just in case they decide to record some audio.

3 minutes ago, Cactus Music said:

Do soft synths output audio at these settings or do they  care? Good question.  Never gave it any thought. Except in the case of freezing the midi instruments if need be I’m not dead sure it matters?  

Audio from a soft synth will be at the project sample rate at the bit depth determined by the plug-in. When frozen, bounced or exported the Render Bit depth is used as the default with an option in the case of export to override.

[David Baay 1,067]

49 minutes ago, Cactus Music said:

But all the ones I’ve converted had at least 4 VST instruments and it worked fine. This is now something to ponder about what is it exactly happens with audio from soft synths?  I think all the ones I’m using are sample based. 

Not to continue too far off-topic, but just having a soft synth in the rack will force the project to use the Audio clock. Same with enabing the audio metronome. You cannot set the clock to Internal and have it stick unless nothing in the project depends on the Audio clock.

[Dave G 6]

Thank you, all, for the concise and detailed answers! 

To reiterate, my projects are strictly instrumental MIDI / SoftSynth-driven tracks with no audio / recorded data whatsoever. However -- when it comes time to finalize and "preserve" my finished tracks, I may bounce them to audio so they won't have to rely on the necessary installed VST's years down the road...

When composing MIDI back in the mid 90's, I don't recall any knowledge or adjustment of sampling rate or bit depth. (From what I recall, Cakewalk's default is 44 kHz.) So, relating to what you've all said, I was going to ask earlier:

1)  Considering these tracks are already in progress, at what point during production does sampling rate become important? (I'm guessing this has been answered.)

2)  I feel confused that Figure 1 in my original post seems to represent the DAW's rate/depth settings, rather than the project's settings. Am I misunderstanding?

3)  Windows default 48 kHz sampling rate -- could I have left it alone, or as long as it's higher than my DAW's 44 kHz? Or must it match exactly?

Otherwise, is it safe to say (per the statement about halfway through this thread) that a MIDI-only project doesn't require these adjustments? Either way, I may set these anyway to 44 kHz/24-bit in a template so every project follows this setting.

I apologize for having to repeat myself in these questions, but I'm documenting what I've learned so it sinks in. Thank you again!

Edited September 15, 2022 by Dave G

[azslow3 324]

I was corrected. I forgot to mention "no audio" case.

Cakewalk always use audio as long as audio interface is configured and used, otherwise soft synths couldn't produce the sound. But till there is  something in the project with audio time reference, the sample rate can be switched.
Cakewalk is "sample accurate" DAW (not all DAWs are like that), all audio times in the project are saved as a sample number. If there are no such "markers" at all, there is no problem to change. But once there is at least one, it has to be recalculated on change and Cakewalk doesn't have the code to do this. That is not something easy to add. In fact "perfect" recalculation is not possible at all, till the sample rate is changing from 48kHz to 96kHz (or other "perfect" up-sampling).

@Dave G better keep Windows and Cakewalk at exactly the same sample rate. Set to whatever you like your final files should be.

[scook 4,774]

1 hour ago, Dave G said:

1)  Considering these tracks are already in progress, at what point during production does sampling rate become important? (I'm guessing this has been answered.)

The current project sample rate is determined one of two ways.

1. If there are audio clips in the project, the sample rate is read from one of the audio clip file headers. All audio clips in a project have the same sample rate.
2. If there are no audio clips in the project, the sample rate is read from the "Sample Rate" setting under "Default Settings for New Projects" in Preferences > Audio > Driver Settings

1 hour ago, Dave G said:

2)  I feel confused that Figure 1 in my original post seems to represent the DAW's rate/depth settings, rather than the project's settings. Am I misunderstanding?

The transport module shows the current project's sample rate and the Record Bit Depth that will be used the next time some audio is recorded. See the answer to 1 for where the sample rate comes from.

1 hour ago, Dave G said:

3)  Windows default 48 kHz sampling rate -- could I have left it alone, or as long as it's higher than my DAW's 44 kHz? Or must it match exactly?

They don't have to match but it may make things easier if they are set the same. Windows may be set to 44.1 or the DAW to 48 take your pick. There is practically no difference between the two.

1 hour ago, Dave G said:

Otherwise, is it safe to say (per the statement about halfway through this thread) that a MIDI-only project doesn't require these adjustments? Either way, I may set these anyway to 44 kHz/24-bit in a template so every project follows this setting.

As noted in the reply to 1, in projects that have no audio clips the sample rate is read from preferences every time the projects are opened.

In projects that have no audio clips, the Render Bit Depth is what matters as it serves as the default in the export dialog, however; the actual bit depth used may be set in the export dialog. What bit depth should be used for export depends on how the export file will be used. If the export file is for playback only, select a bit depth compatible with the playback device(s). If the export file is going to be processed further (mastering?) use 32bit (or 64bit if the 64-bit Double Precision Engine is enabled and the mastering software can load the file).  The current export dialog allows users to create tasks to export. Think of these tasks as the templates you need for projects going forward and update your custom project templates. Similar tasks may be added to existing projects too.

[Starship Krupa 4,059]

Okay, I also do In The Box (aka "ITB") projects, and I'll offer this advice:

Where sample rate matters is when you do your final export/render. Depending on your soft synths and FX, you may likely get better sound if you export at 88.1K or 96K and then use a converter (I use and recommend MediaHuman Audio Converter) for creating files for distribution.

(And BTW, I consider it best practice to make ONE master rendered file and convert it for distribution rather than doing a FLAC export, an MP3 export, etc. The reason for this is that soft synths and FX have randomized elements that will be different every time you render.)

Cakewalk has a built-in feature, plug-in upsampling. You enable this by clicking on the "2X" button in the Control Bar's Mix Module. For ITB projects, this will have the same effect as rendering at 88.1 or 96. If you enable it during mixing/composing, you'll hear any differences at that time (if you hear any differences). What it does is make your plug-ins "think" that they're running at double the sample rate. One caveat: I've gotten unexpected results from upsampling while having the 64-bit precision engine enabled. Since the word is that the 64-bit precision engine is kinda snake oil, I now leave it turned off. I hear no difference in playback or rendered audio. YMMV.

If you're interested in the reason for wanting to do this, it has to do with aliasing, which I won't go into here. It's a kind of distortion that's usually considered unpleasant (unlike the GOOD kind of distortion). Either just try it and listen for a difference in quality, or study up on aliasing in soft synths and FX. It can especially have an effect on reverbs and distortion FX.

HTH.

[Dave G 6]

Thank you again, everyone, for your valued responses and insightful information.

I always find myself second-guessing everything so this turns out right in the long term. After all, I'm no professional, so I'm not trying to pursue professional results...just quality.

Per what I've learned here, I'm going to set the DAW's sample rate and bit depth accordingly (44,100 kHz, 24-bit) and leave it there, build it into the template for forthcoming projects.

Thanks for your help!