16 or 24 bit records

Last days I made several records with keyboardvoices like piano, synthpad, pad with lot of effects.
First track in 16 bit / 44.1 khz
Second track in 24 bit / 44.1 khz
Then cut the two tracks in several pieces and muted them alternately to compare at a stretch.
I did not here the 24bit track was better :astonished:
On the contrary, even the 16 bit track is maybe … slightly better.
Do I have to record in 48 or 96 khz ???
Tried to find it out in the manual, but couldn’t find something as “24bit recording settings”.
My soundcard is a roland ua-25ex (24/96)
Hope you can give your ideas / suggestions.
Thanks a lot! :wink:
Big Bear

As I explained in my last post, If all your sounds are coming from (internal) VSTi’s, then the project bit depth setting will make no difference.

As you don’t explain where the sounds are coming from, external or internal it’s impossible to make a judgment.

Although when recording external sound sources, you may hear a slight difference between 16bit and 24 bit, the main improvement is in the available headroom when using 24 bit recording. (again this only applies for external sound recording) This allows you to record at a lesser “level” with corresponding advantage of not having to “push” the analog side (converter input) so hard, if that’s important. It also IMO makes mixing better as not everything is pegging the meters. Although (mostly) once converted into digital the entire signal path in Cubase is 32bit FP and becomes almost impossible to clip, some plugins like to operate at a nominal level and the output should be kept below the clip mark.

I do believe some VSTi’s can benefit from high sample rates, but beware that this comes at the expense of processing power. Again experimentation is always a god way of convincing yourself.

Split, I appreciatie your comments very much, and that you don’t refer to the manual!
Indeed as you mentioned, the tracks are recorded by external sound sources: Korg Kronos / Tyros3.
One question I still have:
Does it make sense to try out records in 48kHz or 96kHz? (24bit).
Should this result in better soundquality IN THEORY!?
Again thanks for your answering!

It does make sense to try differing samplerates, if only to satisfy the curiosity in yourself :stuck_out_tongue:

Most of us came to the conclusion that the samplerate should be set to fit with the end product rate needed, so for CD 44.1, film/video 48 and if you can afford the power and disk space 96! although the differences are usually very small and most modern music is fine at 24bit 44.1KHz

But give it a go and see what you think :sunglasses:

In theory: no! If you have good A/D converter and all digital processing is made with well-programmed tools, you have absolutely no benefit on going to higher sampling rate than 44.1k.

In practise: Your A/D converter may have better performance with higher sampling rate. And same goes with plugins you are using: they may introduce artifacts (aliasing), of which most are moved into inaudible frequencies when using higher sample rates.

Conclusion: If you have enough CPU power and disk space, you may want to use higher sampling rate for just “being safe”.

Isn’t the max. bitrate of a CD 16bit/44.1khz … Therefore useless to record in 24 bit ? (so it has to be converted to 16bit).

Of face value you would think so but… no

The extra headroom and higher representation allowed in 24bit is well worth it when dealing with multiple tracks and FX etc, and as an added advantage, it doesn’t stress your system like higher samplerates do. Also the conversion from 24 bit to 16 bit is far simpler and produces far less artifacts that a samplerate conversion.

Stick with 24bit.

OK … 24bit … CD or whatever … 24bit :smiley:
Thanks for your patience and very quick and clear answers!

Though I know, that you (split) know that also: recording in 24 bit will not give “extra” or more “headroom”.

Just to add a bit more confusion to the topic.

In most cases once the above is done, more ‘noise’ is actually
added to your work in the final mastering process.

We give that noise a nice name called:

Dithering.

But make no mistake; like it or not, it is still ‘noise’.

Pick yer poision.
{’-’}

On my ua-25ex is written 24/96 . . . sounds to be nice, but . . .
there IS a clear difference between the record (extern soundsources/keyboards) and live sound.

  1. Are there higher qulity cards in recording audio to digital? / I cold try out another card . . . ?
    Are you satisfied with the quality of your cards?
  2. Is it reachable to “repair” the loss of sound by using effects etc.
    Hope you understand my inadequate way of describing :wink:
    Thanks for helping.
    I keep trying out all different recordingmethods :smiley:

Aloha B
For discussion on your question, try here:
http://www.steinberg.net/forum/viewforum.php?f=81

{’-’}

Little more practical:
You may react on an example of one of those records:

http://www.youtube.com/watch?v=m9mGrrd2W2c&feature=context-chv
Recorded as 16bit / 44.1 khz

Thanks!

Are you comparing like with like?

Are you recording the keyboards in stereo, and make sure you have the limiter/compressor switch of the UA-25ex switched OFF (Bypass) and make sure the advanced driver switch is switched ON. When comparing sounds it’s essential that the listening volumes are matched as the louder one will always sound better.

It is nigh on impossible to repair the recorded sound with FX.

Indeed … settings are just as you described … point about volumn is clear:
I will again make some records exactly in same settings, only change bit and khz.
Thanks again !!!
In a way I start really like the research, first take a coffee.