24bit 96Khz is it worth it

[chemn]

Larger sample rates do capture more information (they take up more data storage space also).

They also help to reduce noise when layering and processing multiple takes - goes back to what I was saying earlier about how the noise shaped dithering is spread more thinly across the whole available range meaning there is less dither noise in the audio range.

If the processing and layering you are doing is minimal then I would think that 16-bit 44.1 would be adequate (though there would be no harm in using 24 bit for the sake of having more headroom for recording). But if your processing and layering lots of recorded audio (say 20-30 tracks plus) then the layered noise would become more apparent.

If you have a computer with enough oomph and enough storage space to handle greater bit-depths and higher sample rates, then why not use them? It won't do any harm

 

[Xanadu]

Larger sample rates do capture more information (they take up more data storage space also).

They also help to reduce noise when layering and processing multiple takes - goes back to what I was saying earlier about how the noise shaped dithering is spread more thinly across the whole available range meaning there is less dither noise in the audio range.

If the processing and layering you are doing is minimal then I would think that 16-bit 44.1 would be adequate (though there would be no harm in using 24 bit for the sake of having more headroom for recording). But if your processing and layering lots of recorded audio (say 20-30 tracks plus) then the layered noise would become more apparent.

If you have a computer with enough oomph and enough storage space to handle greater bit-depths and higher sample rates, then why not use them? It won't do any harm

I think you're mixing up sampling rates with bit depths.

 

[southside]

Yeah I think you are right although when I record my own stuff I rarely exceed 12 tracks but when I record a band 16 tracks is usually enough. I might aswell go with the 24/96 as my box has enough power.

 

[chemn]

I'm studying a degree in acoustics, and part of that is digital audio. I've just been learning about how higher sample rates mean that dither noise is reduced in the audio range - giving a greater dynamic range, and reducing quantisation distortion.

Higher sample rates increase data storage size, as do higher bit depths.

 

[chemn]

FWIW I generally use 24 48 - and my audio interface is capable of 192kHz sample rates... But I don't do a lot of recording, when I use 24 48 it's for bouncing loops or edits which I know I will be processing again or that I might use at another time.

Though bees has the most experience with this and he uses 24 44.1 so... yeah.

 

[southside]

No degree for me I'm afraid but I have been recording digitally via pc for 10 years so I have quite a lot of experience with it, it's just some of the dull technical stuf needs looking at now and again, and other peoples views and experiences are always worth checking out.

 

[Xanadu]

I'm studying a degree in acoustics, and part of that is digital audio. I've just been learning about how higher sample rates mean that dither noise is reduced in the audio range - giving a greater dynamic range, and reducing quantisation distortion.

Higher sample rates increase data storage size, as do higher bit depths.

I had a think about it, and I think I can see what you're talking about. But it only affects certain situations.

As part of converting audio from 24-bit to 16-bit (for example), you introduce dithering (i.e. noise) to reduce the effect of quantisation noise. If you are working with a higher sampling rate, the dithering(/noise) can be applied at the higher sampling rate (thus the noise will be at a higher/non-audible frequency).

If you layer lots of tracks together which have been downsampled and dithered from 24-bit to 16-bit (like many commercial recordings), you're layering up the noise, thus increasing the volume of the noise.

If you are keeping all of your audio in the 24-bit domain, you won't have to worry about higher sampling rates, since you'll only be dithering when you're publishing to CD/MP3.

 

[chemn]

What you're saying is ringing true. However, I believe dithering occurs when the recording is first taken, it is part of the quantisation (digitizing) of the audio and is, as such a standard process to reduce quantisation distortion.

End users only come into contact with dithering as part of downsampling because there is usually a choice of what sort of noise shaping to use.

 

[Xanadu]

So going back to your previous post, if you're recording at 24-bit, 44.1kHz, wouldn't it take something in the magnitude of 2^8 (256) tracks before it became remotely audible when mixing down to 16-bit?

Also, will all of the plug-ins you use support 96kHz? I think most support 24-bit, but if it's having to convert down to 44.1kHz and back up to 96kHz, you're probably not going to be doing yourself any favours on the sound quality front.

24 tracks at 24-bit/96kHz is going to be around 6.6Mbytes per second (compared to 3Mbytes/sec for 24-bit/44.1kHz). Processing that amount of data (e.g. with effects) in real-time is quite taxing.

 

[beesonthewhatnow]

However, I believe dithering occurs when the recording is first taken

Not true AFAIK...

<goes off to read a few books>

 

[DarthSydodyas]

One positive outcome is all that wonderful 48khz supporting mixers and pro-gear dropped down in price.

 

[newbie]

FWIW I generally use 24 48 - and my audio interface is capable of 192kHz sample rates... But I don't do a lot of recording, when I use 24 48 it's for bouncing loops or edits which I know I will be processing again or that I might use at another time.

Though bees has the most experience with this and he uses 24 44.1 so... yeah.

One positive outcome is all that wonderful 48khz supporting mixers and pro-gear dropped down in price.

if you're producing for the world of CDs, rather than television, there's no advantage to 48k and probably some disadvantages, whether or not they're audible to real ears with real equipment in real rooms. However good sample rate conversion is, I don't think there's any positives to a 1.088435 transformation.

The same must be true of 96k. I've never heard of anyone using 88.2k but I presume they do.

 

[DarthSydodyas]

if you're producing for the world of CDs, rather than television, there's no advantage to 48k and probably some disadvantages

Perhaps, but it will support 44khz in any event. A win-win situation, imo.

 

[Danalogue]

I had a big technique gibberish answer for you and then I re read your original post and realised its all bollocks.

I am a rock musician so I always record live instruments and timestreeatching and tempo adjustment is not necessary. I thought that the sample rate was akin to tape speed adjustment like 15ips or 30ips for example.

this is gold dust. Record in basic 16 bit 44.1 - if you play your instruments over and record each one until you have the take 'your' happy with for each intrument, you are on a winner, your ahead of a big percentage of the game. A song is a song, my guess is your looking to send a demo out. 24 bit 16 bit is fuckall, mate if a record company likes your track they will have you re record it with their people. Given the software around these days you could sit a bunch of monkeys in a room full of Bannanas injected with speed and a lap top set with three VSTs and Quantise switched on and it wouldnt sound out of place. What bit rates, quantise, melodyne correction , omnisphere psycho accoustics, dblue glitch randomness, convelution reverb modelled on my fucking arse crevis mixied with a tripple tap delay with oochy coochy reverb, dont make a difference. If you can create something different that has 'your sound' then your on a winner. My adivce is to concentrate on writing and recording your songs into the computer with whatever hardware you have avialable right now. Then stick through Izatope Ozone 4 or T Racks Mastering software. PM me if you want any help on mixdown.

 

[southside]

I dont get what the flying fuck you are on about?

 

[Xanadu]

this is gold dust. Record in basic 16 bit 44.1 - if you play your instruments over and record each one until you have the take 'your' happy with for each intrument, you are on a winner, your ahead of a big percentage of the game. A song is a song, my guess is your looking to send a demo out. 24 bit 16 bit is fuckall, mate if a record company likes your track they will have you re record it with their people. Given the software around these days you could sit a bunch of monkeys in a room full of Bannanas injected with speed and a lap top set with three VSTs and Quantise switched on and it wouldnt sound out of place. What bit rates, quantise, melodyne correction , omnisphere psycho accoustics, dblue glitch randomness, convelution reverb modelled on my fucking arse crevis mixied with a tripple tap delay with oochy coochy reverb, dont make a difference. If you can create something different that has 'your sound' then your on a winner. My adivce is to concentrate on writing and recording your songs into the computer with whatever hardware you have avialable right now. Then stick through Izatope Ozone 4 or T Racks Mastering software. PM me if you want any help on mixdown.

Almost agree. The extra headroom of recording in 24-bit makes a big difference when recording. It means you can set the volume level of your source to 1/256th of the maximum that your audio interface can handle, and you've still got as much detail as a 16-bit waveform. It means you can do so much more messing about before it sounds shit.

 

[southside]

Rubbish in rubbish out in my book.

You cant polish a turd.

But you can varnish one thou.

 

[beesonthewhatnow]

Or roll it in glitter.

 

[southside]

Lol.....

 

[audiotech]

Sound advice literally from Danalogue.

 

[southside]

I had a big technique gibberish answer for you and then I re read your original post and realised its all bollocks.



this is gold dust. Record in basic 16 bit 44.1 - if you play your instruments over and record each one until you have the take 'your' happy with for each intrument, you are on a winner, your ahead of a big percentage of the game. A song is a song, my guess is your looking to send a demo out. 24 bit 16 bit is fuckall, mate if a record company likes your track they will have you re record it with their people. Given the software around these days you could sit a bunch of monkeys in a room full of Bannanas injected with speed and a lap top set with three VSTs and Quantise switched on and it wouldnt sound out of place. What bit rates, quantise, melodyne correction , omnisphere psycho accoustics, dblue glitch randomness, convelution reverb modelled on my fucking arse crevis mixied with a tripple tap delay with oochy coochy reverb, dont make a difference. If you can create something different that has 'your sound' then your on a winner. My adivce is to concentrate on writing and recording your songs into the computer with whatever hardware you have avialable right now. Then stick through Izatope Ozone 4 or T Racks Mastering software. PM me if you want any help on mixdown.

Im a wrong side of 40 Guitarist I'm never going to send anything I do to a record company, I only do it for fun really a lot of trial and error. Sometimes stuff works othertimes I sound like a dog fucking a bag of nails but you have to try these things for your self and most of the things I record are experiments.

 

[Erich Zann]

I had a big technique gibberish answer for you and then I re read your original post and realised its all bollocks.



this is gold dust. Record in basic 16 bit 44.1 - if you play your instruments over and record each one until you have the take 'your' happy with for each intrument, you are on a winner, your ahead of a big percentage of the game. A song is a song, my guess is your looking to send a demo out. 24 bit 16 bit is fuckall, mate if a record company likes your track they will have you re record it with their people. Given the software around these days you could sit a bunch of monkeys in a room full of Bannanas injected with speed and a lap top set with three VSTs and Quantise switched on and it wouldnt sound out of place. What bit rates, quantise, melodyne correction , omnisphere psycho accoustics, dblue glitch randomness, convelution reverb modelled on my fucking arse crevis mixied with a tripple tap delay with oochy coochy reverb, dont make a difference. If you can create something different that has 'your sound' then your on a winner. My adivce is to concentrate on writing and recording your songs into the computer with whatever hardware you have avialable right now. Then stick through Izatope Ozone 4 or T Racks Mastering software. PM me if you want any help on mixdown.

If I use 24bit depth, does the CPU of the machine work harder to use them?

 

[southside]

Yes because the app you are using will be capturing more information.