Octatrack: 16bit or 24bit Mode?

Do people tend to work in 24 or 16 bit or end up with a mixture? I have a stack of samples but of course at different bit rates. Is it better to stick to one bitrate, and batch convert before loading, or for example let flex format convert on loading/saving using the 16 bit setting here?

Best quality or maximum time? You have to choose. I chose maximum time, for longer recordings (3-5 mn) and most of my samples are 16 bits.

Max RAM time with 16 bits : 8m28s
Max RAM time with 24 bits : 5m39s

I tried 24 bit recording, I didnā€™t hear the difference. Same noise floor apparently. Probably better for mixing. Some people use 24 bits only. I need to make better tests to check if it worth it.

2 Likes

Actually higher bit rate is mostly useful for recording because of the much lower noise floor. So you can record at lower levels and thus less risk of clipping. For mixing it really doesnā€™t matter.

I wonder if the conversion within octa also uses dithering when converting. Otherwise I probably would batch convert with dithering before loading them into octa.

3 Likes

Would you test it with sine waves in 16/24 bits?

I donā€™t own an octatrack just yet :wink: just following the thread and reading the manual haha. Just thought Iā€™d chime in with some theory. But would love to know if octatrack does that :slight_smile:

4 Likes

Thanks @sezare56 @DaveMech
Octatrack arrives this afternoon so keen to start off on the right foot (or bit rate :slightly_smiling_face:)

2 Likes

If it doesnā€™t influence mixing, 16 bits should be better to save RAMā€¦Mono samples save RAM too. OT records in stereo only. Not possible to convert to mono files with.

I just played / recorded a sine followed by silence with OT, CUE out > Input A, in 16 and 24 bits. Apparently no sound / noise floor difference, played in 16 or 24 bits.

Iā€™m totally open to hear something thatā€™d prove (by ear) that 24 bit recording really worth it with OT.

1 Like

It really depends on how loud you are recording. 24bit in general (also in a daw environment) is useful for lower level recordings to avoid any clipping. If you would record at that same low level at 16bit and then turn up the gain after to make it louder, the noise might become more apparent. (noise floor difference is about 40db between 16 and 24bit).

So it wonā€™t matter much if you record loud enough indeed. And it is better to have as little convertion stages as possible. Therefore I think if I had an octa Iā€™d record at 16bit as well. Maybe the exception would be if I want to record a guitar or bass guitar directly into octa. Because transients will be less predictable

1 Like

I made the same test playing the sine - 24 db lower. Same noise level with 16 bit or 24 bit. :thinking:

Have you normalized both afterwards?

It is possible that you still would have to play back the recordings at a considerable level to hear the difference. Room noise, speaker noise etc. All have influence to and could mask it. Preamps / converters also have an influence even.
Hard to judge from here :wink: (also have to play back the 24bit recording as is and not have it be converted to 16. Same goes for the 16 bit recording, play as is witniut conversion).

What I said above is simply the theoretical difference between 16 and 24. If you donā€™t hear a difference donā€™t worry about it too much and just enjoy making music :wink:

1 Like

No I only used OT FX compressor gain, same settings for both (16/24 bits) on 2 tracks, headphones.

I compared them with 16 and 24 bits playback.
This deserves a topic btw!

1 Like

I wouldnā€™t use a compressor gain to boost the level. You want to have the signalpath as clean as possible to check this. So no eq /fx / Comp etc. I think something like time stretch should also be turned off etc. As I read it influences sound as well.

Try playing the sine at -40db and then some silence afterwards. . Then save the recordings to the flash card. Load them into a daw environment, normalize them both and then check with meters.

Edit: maybe a bit off topic indeed

1 Like

I have a few dragons to finish first!

7 Likes

Yes indeed, 24 vs 16 and mono vs stereo and how they relate to file size, etc would be a handy reference, with audio examples and screen shots etc. where applicable.

FWIW I mostly always use 16bit too, Iā€™m not particularly concerned about noise floor and most of my sampling is electronic sounds or pre-recorded sounds where the noise floor is often already higher than the difference of using 24bit would benefit, mostly.

2 Likes

is that per track or project?

total (per project)

my understanding is its about dynamic rangeā€¦ (something like ~50db difference?)

if your recording is just quiet, youā€™d turn up the gain (at the ADC level),
the issue comes when you have a recording where you have both quiet and loud,
at this point if you raised the gain, you would start clipping louder parts,
but keeping the gain level low, means due to the lower resolution, your SNR is higher.

soā€¦ if your signals are not too dynamic, then 16bit wonā€™t be an issue.

note: id assume the dac is always 24bit. then its reduced to 16bit when its placed into recording buffers - so at the point where amount of memory because the ā€˜issueā€™.

hard to know technically without having more specific about DAC, and how things are coded in the OT.

3 Likes

@thetechnobear Not quite. Dynamic range is the difference between the softest part of a recording and the loudest part of a recording. Or the range (in db) that a system can reliably play back audio at.

24bit vs 16bit means a difference in noise floor. 24bit has a noise floor of -144db. 16 bit has a noise floor of -96db. See my other posts where I explain this as well. Itā€™s about headroom. 24 bit gives you more headroom to play with and thus you can record audio on lower levels without it being an issue.

So if you play an instrument with big differences in loudness because of transients (like a picking guitar piece) or record something else that has very soft and very loud parts, then 24bit is very useful because you can record at volumes low enough to keep from clipping the signal without having to worry about noise when using compressors and what not to even out the dynamic range of a recording (or simply boosting the volume of softer parts for instance).

If however you play letā€™s say pad from a synth that has a constant volume all the way through (or another thing you record that has a small dynamic range <- ), recording at 16 bit wil be fine.

Cheers

5 Likes

I just flagged my own post as off-topic btw haha. Maybe best to move this to a dedicated topic.

2 Likes

This one seems appropriate.
Octatrack: 16bit or 24bit Mode?

@DaveMech So I recorded with a sine at - 36db, the difference was really noticeable. Barely noticeable with MIXER GAIN set to +63.
I need to investigate. :thinking:
Thanks for the insight!

3 Likes

Mixer gain is only +12db no?