Recording at 24-bit/96 kHz

If I recorded at this higher quality, will there be any problems later when I have to mix everything down to CD-quality? Will the CD-quality song sound worse, better, or unchanged compared to if I recorded in CD-quality to begin with? What other benefits or disadvantages are there to recording in this higher quality? For me, the large amount of hard drive space that is required is not an issue. Thanks for the help.

Aloha p,
For me it all depends on the type of music.

I can notice a difference if I am recording a single flute
or piano or violin etc.
Also I have also recorded string quartets and
(light) jazz trios and I can notice even then.

But If I am recording a pop or reggae or rock band or
even a ‘full out jazz trio’ etc I can’t really ‘hear’ the difference.

Seems to me, the more dynamic the music, the more
I can notice the difference.

As for recording at higher rates and then putting that work on CD,
Now we enter into the world of ‘dithering’.

Most folks dither when rendering files to be put on CD.
That is to add a ‘pleasing noise’ to mask the sound of the artifacts
made when changing sound files from a higher to lower bit rate

IMHO it best to use your ears and see if the dithering noise
sounds better than the artifacts.

HTH (hope this helps)
{’-’}

Do vst’s sound better in the higher sample rate?

I would imagine it depends on at what rate the sample was made.

If you have a VST playing a 16 bit sample, no matter what you
do, that sample is not going to sound any better at a higher rate.

Because it was recorded at 16 bit.

{’-’}

There are advantages on using higher sample rates (96 vs 44.1 kHz) and bit-depths (24 vs 16).

Sample rate:

Higher sample rate increases frequency response. While nobody can directly benefit on this (as we don’t hear these frequencies) using it moves artifacts of digital processing (aliasing) into those inaudible frequencies. This is most noticable when using some not so well coded plugins.

Also some prosumer-grade A/D converters do perform better on higher sample rates.

Bit depth:

While our final product is 16 bits, recording at 24 bits guarantees our final product taking advantage of all those 16. Whenever we do any processing on audio, we loose at least one bit of its accuracy due to rounding errors. Also compression by default throws away bits (1 bit/6dB of compression). If we start with 16 bits, export couple of times with some processing and compress the final product by 12 dB, our audio is effectly 12-bit.

Conclusions:

With high-quality hardware (high-end converters), well-coded software (plug-ins with transparent anti-alias filters) and careful working methods (recording in right levels, no intermediate exports etc), it should be impossible to distinguish a pop recording made in 24/96 from the one made in 16/44.1, but it’s sometimes impossible to meet those conditions. So if you are able to record 24/96 (24/88.2 does the job as well … even 20/60 would if it existed) just do it.

Conversion to CD-format:

When converting from 24/96 to 16/44.1, you may want to use best tools possible. Cubase’s sample rate converter isn’t really a top-notch one. But there’s free alternatives like r8brain or SoX. Now, can you hear the difference? Probably not.

Who says?

How does compressing 24-bit audio and then dithering to 16-bit compare (in principle) with dithering or truncating to 16-bit and then compressing (by the same amount)?

All the schoolkids around here know so. :mrgreen: And they know everything. :laughing:

There can be a few benefits at the cost of much larger files to save or more work dithering down to 44.1/24.

There is “some” difference between 44.1 and 96, but not enough to warrant the system demands of working at 96 imo.

If anyone can tell me the difference between 44.1 and 48 , i’d be happy to literally hear the difference. I’ll be damned if I can hear anything.

If you’ll take a 16-bit file, the digital artifacts (distortions) are at -96dBfs. Now if you compress by 12dB and apply make-up gain of 12dB to bring back the peaks at where they were, the artifacts will be at -84dBfs.

If you’ll take a 24-bit file the artifacts are at -144dBfs before and -132dBfs after compression. Now truncate to 16 bits and those artifacts are masked away by the new artifacts of truncating (at -96dBfs).

So doing all your processing in 24-bit domain virtually guarantees your final product takes full advantage of those remaining 16 bits.

Now can you hear this difference? Depends on music. -84dBfs is still 0.0006% of 0dBfs. Can you hear 0.0006% distortion? Definitely not! But what if you music is very dynamical … say the dynamic range is 50 dB. Now -84dBfs is 34dB below your soft passages … meaning 2% distortion. This is audible. It’s around the same values your speakers might introduce or even more. You really want to drop the distortions back to -96dBfs … meaning 0.5% distortion during soft passages (which is acceptable figure).

If your music is modern pop music, your dynamic range hardly exceeds 20dB. (If you’re playing in/producing Metallica it’s around 6dB). Let’s try to keep distortion down at or below 0.1%. What it means: distortion must be 60dB below your softest passage which is -80dBfs.

Again: in modern pop music, it hardly matters. But using 24 bits instead of 16 bits don’t take hit on your CPU and only consumes 1.5 times disk space. So I see no reason to go to 16 bits before creating final master.

I left dithering out on purpose in these examples, because it makes math behind those dBs more complicated, but principle is the same.

When I change the settings in my project from 16/44.1 to 24/96 it pitches all of my files and speeds everything up. How to I make it so that the stuff I have already recorded doesn’t get pitched up or sped up?

I love the friendliness coming off the “grown-ups”.It so inspires to ask questions and give tips and answers.

And I love the “We will not have any fun here.” attitude of those who know everything. :mrgreen:

Also if I need a doctor I visit my local bar. They know everything too. :mrgreen:
And, of course, so do I. :laughing: :laughing:
People who ask proper questions KNOW we’re only messing about. Put that ego in a bag and bury it. :smiley:

Do you use the “Projects Settings” for this?
If you do, do you not get a dialog box asking if you want to change the files’ sample rates? To match the Project.

When changing sample rate Cubase should ask you:
Convert audio files to new rate?
Into which you should answer: Convert

Yeah, so now the files aren’t being sped up, so thats good at least, but the VST’s are being sped up now and also the files aren’t being placed in there right positions. For example audio tracks occur sooner than they should, but play at the right speed and pitch.

Are these frozen VSTi:s. If so, unfreeze them before conversion.

You were asked during the conversion: Do you want to keep audio events at their sample positions?
into which you should have answered: No, because you don’t want to keep them in sample positions but time positions.

Thanks a lot for the help. However, how do you unfreeze VST’s?

Press Freeze button on the frozen instrument track.

THANKS! Everything worked perfectly and you saved me hours of messing around with cubase.