Is anything to be gained by converting sample sizes 16/24?

I notice on the import dialog there is an option to convert sample size on import. I cant see why this should be done as it cant (I think) improve the quality of the original sample, but will result in a bigger file size.

I must be missing something?


There’s an advantage when you use offline processing. A 16 bit file/event will result in a 16 bit file after offline processing though the calculation process is 32 bit float. Half of the bit resolution doesn’t make it to the audio file. Which you might consider a problem in some cases and in other you won’t, depending on the sample.

Without offline processing it’s indeed just a bigger file size, no improvements to be expected.

Are you sure about that? is offline processing bitdepth final result not tied into the project bitdepth setting?

You have to be careful when changing sample rates up and down. The higher the bit rate, the more detail is in the sound. If you covert a 24 bit sample to 16 bit, it will loose the last 8 bits of the wordlength and it can’t be replaced, even if you convert it back to 24 bit, you’d notice a lot of difference or loos of definition (mostly) in the low end.

Not an expert on this, just some info I came across reading up on the dithering section of mixdown in a Waves Multimaxer/Compressor. There’s quite a good, simple explanation of what happens during changing the sample rates in the Waves L1/2/3 etc mastering plug in manuals which you can download for free. There’ a wealth of other information available all over the web, probably even some in Steinbergs if you search it.

Hope this helps, if this is the sort of info yor asking for :sunglasses:

Here we are talking asbout 16 bit being converted to 24 bit and then never back to 16 bit, as the project is sert to 24

Bitdepth (wordsize) and samplerate should not be confused!

OK so I am now totally fogged. It’s easy to do this as my brain is made of mushy stuff which resembles Smash - if you remmber the advert. Anyway I am not hard wired like some of you earthlings, all I want to know is: use it yes or no?

Read marQs´ post and decide yourself if this applies to you. If it does: Yes, if you want
If you don´t understand what he´s talking about: No.

Also read my first post as I’m pretty sure that it’s the project bitdepth that determines offline processing bitdepth, thus the answer would be no.

Sorry, it´s not, it´s the file´s original bitdepth. (I tried it before posting, since I first also thought it was the way you suggest, but the project bit depth only applies to recording with FX)

Oh… thats interesting :stuck_out_tongue:

I could almost of bet that was not the case :blush:

That would go some way to explain the convert bit depth on import :exclamation:

wtf? are we talking sample rate here or bit depth? some confusing folks here. Sample rate id like to use these days is 96khz when many plugins are used due the aliasing. Bit depth im not very clear as 24 bit all the way done me good so far.

This topic has nothing to do with samplerate… I hope thats clear :slight_smile:

The bitdepth of 24 bit gives you more headroom, audio clips at higher DBs.

By the way the bussing in Cubase occurs at a bitdepth of 32 bit at any samplerate you may want.

24 Bit will not give you more headroom!!!

Was waiting for that :laughing:

Of course you were :laughing:

Oops! Sorry :laughing: But think both are important and part of the op’s question to get a full understanding bewteen the two and how they are both important factors in any equations of recording and mixing/mastering.

Here’s a fairly straight forward explanation the in and outs (off wikipedia) which puts it in a nutashell as both are important in context of releases and intended market/audience from MP3 through CD, DVD and Blue Ray

Word Size

The number of bits used to represent a single audio wave (the word size) directly affects the achievable noise level of a signal recorded with added dither, or the distortion of an undithered signal. Increasing a sample’s word length by one bit doubles its possible values, likewise increasing the potential accuracy of each sample and the fidelity of the recording to the original. 24-bit recording is generally considered a current practical limit as this word length allows a signal-to-noise ratio exceeding that of most analog circuitry, which by necessity must be used in at least two points in the recording/playback chain.

Sample rate

The sample rate is even more important a consideration than the word size. If the sample rate is too low, the sampled signal cannot be reconstructed to the original sound signal. Hence the output will be different from the input. The process of under sampling results in aliasing whereby the high frequency components of the sound wave are represented as being lower than they should be. This causes the output wave shape to be severely altered.

To overcome aliasing, the sound signal (or other signal) must be sampled at a rate at least twice that of the highest frequency component in the signal. This is known as the Nyquist-Shannon sampling theorem.

For recording music-quality audio the following PCM sampling rates are the most common:

44.1 kHz 48 kHz 88.2 kHz 96 kHz 176.4 kHz 192 kHz

When making a recording, experienced audio recording and mastering engineers will normally do a master recording at a higher sampling rate (i.e. 88.2, 96, 176.4 or 192 kHz) and then do any editing or mixing at that same higher frequency. High resolution PCM recordings have been released on DVD-Audio (also known as DVD-A), DAD (Digital Audio Disc—which utilizes the stereo PCM audio tracks of a regular DVD), DualDisc (utilizing the DVD-Audio layer), or Blu-ray (Profile 3.0 is the Blu-ray audio standard, although as of mid-2009 it is unclear whether this will ever really be used as an audio-only format). In addition it is nowadays also possible and common to release a high resolution recording directly as either an uncompressed WAV or lossless compressed FLAC file[8] (usually at 24 bits) without down-converting it .

However if a CD (the CD Red Book standard is 44.1 kHz 16 bit) is to be made from a recording, then doing the initial recording using a sampling rate of 44.1 kHz is obviously one approach. Another approach that is usually preferred is to use a higher sample rate and then downsample to the final format’s sample rate. This is usually done as part of the mastering process. One advantage to the latter approach is that way a high resolution recording can be released, as well as a CD and/or lossy compressed file such as mp3—all from the same master recording.

Hope this helps make it all a bit more of real world examples :slight_smile:

16/24 bit at 44.1 id fine for CD, can be lowered then for web/mp3 but for DVD/Games/Scoring, the obviously 96-192Khz is more the norm (as long as you’ve got racks of disks to store the projects and back them up onto)

Funny discussion on internal ‘mysteries’. I still believe it’s as simple as I wrote. I draw my conclusions from what’s displayed in the pool - 16 bit files remain 16 bit files. No matter how much processing is done to them.

It’d be consequent if Cubase turned files of lower bit depth to 32 bit float automatically when they get processed down on file level (‘offline processed’), just to maintain the highest possible quality. However - it doesn’t. With many samples it’s not an issue. Will you be able to detect any audible difference in that heavily processed noisy snare sample at 16/24/32 bit?!? Guess you won’t (as long as it doesn’t clip)…

Increase of bit depth really just makes sense for offline stuff. In case of 24 bit the resulting file contains the original 16 bit of true audio information plus 8 empty bits. It’s identical but has 8 extra bits of ‘room’ now for more detail that will happen to be there if you process it with plugins (as this processing will be done in 32 bit float internally).

Makes no sense if that 16 bit file simply plays on a track. Any operation, the most complex plugin in a slot as well as a simple gain change on the fader is performed with the resolution of the audio engine anyway - again 32 bit float in Cubase.


Somewhat surprising but indeed ofline processing does yeald a bit depth of the original files bit depth.
Seems like I was assuming that the result would be tied into the project bit depth setting, like it is for bounce audio!