does any1 hear record at 16 bit so they do not have to dither?

[beatzz]

Hey guys, i'm curious here if anyone hear records their material at 16 bit, therefore does not have to dither their material at the end.

I have heard people on hear say that even recording at 16 bits, that gives one enough head room to fully capture a signal and u don't have to worry about dithering noise because u will not have to dither your project.

I'm curious if this is a route anybody here takes as opposed to recording at 24 bit and then dithering?


Why or why not?

[BrianT]

Originally Posted by beatzz

Hey guys, i'm curious here if anyone hear records their material at 16 bit, therefore does not have to dither their material at the end.

I have heard people on hear say that even recording at 16 bits, that gives one enough head room to fully capture a signal and u don't have to worry about dithering noise because u will not have to dither your project.

I'm curious if this is a route anybody here takes as opposed to recording at 24 bit and then dithering?


Why or why not?

I record rock/pop music at 16 bits, but let me give you some advice.

If you use your own ears to decide instead of just reading about the numbers (like most people do), don't tell anybody, because then, you are going to get an earful about how wrong that is.

You ask, "Why?" I made that decision after comparing basic tracks I recorded at both 16 and 24 bits with some of the best players on the planet on the best gear made. A bunch of us set around on more than one occasion, under instant A/B comparison of the same basic tracks cut at 16 vs 24 bits. That's how I decided to record 16 bits.

But to be fair, it is probably very dependent on the particulars of the setup being used. The A/D, D/A , soundcard and internal aspects of the software all contribute.

I would listen to what other people have to say, but ultimately I would use your own ears and compare for yourself, because if you can't trust your own ears, you're already in trouble, eh?

[AlexLakis]

I record at 24 bit because I use plugins which process at 24 bit. That, and higher bit depth in general is very important to me, much more so than samplerate. From my understanding, higher bitrate vs. lower bitrate is like Renoir vs. Van Gogh. Sometimes Van Gogh is what you're looking for, a lot of people do prefer 16 bit for hard rock. To my ears, it sounds a little bit more gritty. Sometimes I can't tell the difference.

[Youn]

...u don't have to worry about dithering noise because u will not have to dither your project.

You don't have to do anything you don't want, but recording 16-bits does not mean dither is any less important then if recording at 24-bit. Once you do any sort of mixing, you should want to properly dither because those 16-bit tracks are being upsampled and all kinds of math is being done to it.

Now, if you just mean recording a stereo 16-bit track and then burning it straight to CD, with no editing or mixing or nothin', then you probably will not need to dither.

[dynomike]

Originally Posted by Youn

You don't have to do anything you don't want, but recording 16-bits does not mean dither is any less important then if recording at 24-bit. Once you do any sort of mixing, you should want to properly dither because those 16-bit tracks are being upsampled and all kinds of math is being done to it.

Now, if you just mean recording a stereo 16-bit track and then burning it straight to CD, with no editing or mixing or nothin', then you probably will not need to dither.

Very good point, I was about to say this myself. Dithering or not dithering is not the factor here. What is important is headroom and low level detail. If the material you're recording does not have an enormous dynamic range or very critical reverb tails or something, you may very well not be able to tell the difference between tracks recorded at 16bit vs 24bit. Both have a great dynamic range compared to the useable range of most analogue recorders.

As Youn said, whatever program or digital mixer you are using probably has a much higher internal bit depth than 16 or 24.. (protools is 48bit, sonar 5 is 64 bit, soundcraft digital mixers are 56 bit, etc) This is very important when doing math on the signal, and often dither is applied when the wordlength is reduced to 24 bit or 16 bit (whatever you specify for the output I guess). There's usually an option within the software to turn this on or off. If its UV22 or PowR, I'd just recommend leaving it on. Dither isn't really a bad thing! Too much dither is, but just on your 2 track, it generally provides extra dynamic range and smooths out the sound at very low levels (for more, see Ken Pohlmann's "Principles of Digital Audio").

Right, and in response to Alex who said he records in 24 bit because his plugins are 24 bit... the internal bit depth of software (plugins, daw software, whatever) should operate independently of what word length you recorded at. Its gonna float your 16 or 24 bits within its internal bit depth, which is often much higher. Lots of reverbs have 64 bit processing. For more on this, there's a whitepaper on the internal mixer on digidesign's website, don't remember the link.

Mike

[AlexLakis]

Originally Posted by dynomike

Right, and in response to Alex who said he records in 24 bit because his plugins are 24 bit... the internal bit depth of software (plugins, daw software, whatever) should operate independently of what word length you recorded at. Its gonna float your 16 or 24 bits within its internal bit depth, which is often much higher. Lots of reverbs have 64 bit processing. For more on this, there's a whitepaper on the internal mixer on digidesign's website, don't remember the link.

Mike

Correct, I just meant it isn't limited by the medium it passes thru. If I had the means to record at 32bit, or even higher, I would! Mmmm, 64-bit A/D-D/A converterrrrs...(insert Homer drooling noise)

[BrianT]

If you want to do a test, it's pretty easy.

Record a multitrack at 24 bits. Finish and mix the project.

Now, rerender all the individual tracks to 16 bits, in place in the project. Most, if not al DAWs should allow you to do this. You can dither, or not, as you render to compare that as well.

Print the same mix again now, with no other change except for going from 24 to 16 bits on the multitrack files. Have a listen to compare the two mixes and see what you think.

If more people did this one simple test, this subject would have a lot more validity when discussed, IMO. As it is now, mosty people just quote numbers and talk about mathematics, never having done a real A/B for themselves.

There is no such thing as "Right" or "Wrong" when it comes to music. There is only "Feeling" or "No Feeling".

[Les Ismore]

That's not quite correct. If you dither a 24 bit file to 16 bits you end up with considerably more than 16 bit resolution but less than 24. I forget the exact number but I think it's around 20 bits of resolution.
I feel that to do this properly for a fair evaluation you should realy record the tracks seperately into both formats.
As many people have stated, to really judge you must do a completely blind test not knowing which is which. It's suprising the number of times people choose 16 over 24 when they don't know which is which. People read thier preconditioning from articles and ads into the equation subconciously. You'd be amazed at what people "think" they hear because of the power of suggestion.

[aussie_techie]

Originally Posted by beatzz

I'm curious if this is a route anybody here takes as opposed to recording at 24 bit and then dithering?

if you record at 16bit you proberly should dither before going to disk, many converters and digital desks can do this and if done correctly there should be very little audible difference to recording at true 24bit. truncating to 16bit can be usefully in certain situations but i will never record anything at 16bit again. IMO recording at 16bit with no dithering lacks deepth. i usualy work at 24/44.1 but i hardly ever look at sample rate or bit deepth untill exporting or if i feel the song lacks deepth ill check and sure enough i find its been recordered at 16bit. to me it just sounds better but use your own ears and decide for your self what you like with your equipment
why do some people go to 32bit float? is there realy a difference, use what works for you

[cdog]

Originally Posted by beatzz

Hey guys, i'm curious here if anyone hear records their material at 16 bit, therefore does not have to dither their material at the end.

I have heard people on hear say that even recording at 16 bits, that gives one enough head room to fully capture a signal and u don't have to worry about dithering noise because u will not have to dither your project.

I'm curious if this is a route anybody here takes as opposed to recording at 24 bit and then dithering?


Why or why not?

Dither is you friend.

Do not fear dither.

It preserves the detail of your 24 bit recordings.

You really, really, want to record at 24 bits.

With close micing, you can tell much less of a difference 16 vs 24.

Those extra bits are at the bottom of the audibility scale, not the top.

But distance micing is all in the low level detail.

Drums tracked @ 16 vs 24 bit is the most obvious to me.

Electronic music, dance, rap.... its less dynamic so 16 bits is proabably fine.

Sample rate? ehhh. Doesn't matter nearly as much.

[JP11]

Originally Posted by BrianT

I record rock/pop music at 16 bits, but let me give you some advice.

If you use your own ears to decide instead of just reading about the numbers (like most people do), don't tell anybody, because then, you are going to get an earful about how wrong that is.

You ask, "Why?" I made that decision after comparing basic tracks I recorded at both 16 and 24 bits with some of the best players on the planet on the best gear made. A bunch of us set around on more than one occasion, under instant A/B comparison of the same basic tracks cut at 16 vs 24 bits. That's how I decided to record 16 bits.

But to be fair, it is probably very dependent on the particulars of the setup being used. The A/D, D/A , soundcard and internal aspects of the software all contribute.

I would listen to what other people have to say, but ultimately I would use your own ears and compare for yourself, because if you can't trust your own ears, you're already in trouble, eh?

Hey BrianT...not to hijack the thread, but quick question....are you still using Paris in any way....if not, what have you found to replace it....

there are many discussions here about summing, summing boxes, keeping your levels low, and all that...Paris was known for summing and spanking levels, against all logic and all other DAWs...

would be interested in your thoughts these days...

[genericperson]

if you are not doing any processing, then keep it at the bit rate it came in on. (for example if you are just chopping the ends off of something).

but as soon as you apply a plugin like an eq, or change the level on the fader, then it gets bumped into the computer's 32 bit processing (or 40 bit or whatever the DAW uses). in this situation, you will want to be working at a higher bit rate, and then dither at the final stage.

[dasbin]

If you are recording at 16-bit, you are dithering anyway.
Most modern converters these days are 24-bit converters, and when they are in 16-bit mode, they are actually still first converting to 24-bit and then just dithering the tracks down to 16-bit on their own.

I don't see why on earth a person would want to avoid dither. It helps a make a painful process much less painless.

[Les Ismore]

Originally Posted by dasbin

If you are recording at 16-bit, you are dithering anyway.
Most modern converters these days are 24-bit converters, and when they are in 16-bit mode, they are actually still first converting to 24-bit and then just dithering the tracks down to 16-bit on their own.

I don't see why on earth a person would want to avoid dither. It helps a make a painful process much less painless.

Protools 5 did not automaticaly dither to 16 bit when recording 16 bit files, it truncated. (I had many long talks with digi on this). I'm not sure about PT6 but I would err on the side of caution. When recording 16bit files thru 24 bit converters, you must create an aux channel as the input of each track and insert a dither plug, then bus that to an audio channel and record that. Otherwise you are being truncated.

[dasbin]

I'm talking about the hardware, not the software.
The A/D chips themselves will dither when they are "speaking" 16-bit.

Read datasheets from modern AKM or Cirrus ADC's to see.

[moops]

I think I see what you are asking: Is it better to record in 16bit in the first place since the output medium (CD) would be the same bit depth, and you could then presume that whatever you hear when you're recording would be the same as you'd hear on the final medium...?

Lots of good advice given here on the forum so far.

In all seriousness, recording at 24bit gives you a much greater headroom when recording.

Unlike the analog days when you used to record things as hot as you could to get as large a signal to tape as possible and keep noise as low as possible; with 24bit digital you can effectively record at a reduced level as the noise floor is so much lower, and that translates to giving you the headroom you need to cope with peaks, especially if you don't want to use compression when tracking...and also means that you when mixing you aren't constantly backing things off with the sum of the tracks being too hot...or alternatively, a final mix of quietly recorded tracks at 24bit will have a lower noise floor than a final mix of quietly recorded tracks at 16bit.

Dither IS your friend, and there are many dithering algorithms around that are thought to be goodly and musical. You will certainly get a greater perceived bit depth from recording at higher bit levels and dithering down than recording at 16bit alone. For example, Sonar has a choice of algorithms depending upon the type of recorded material:

POW-R #1: This is optimized for simple programs of average dynamic range, such as spoken word.
POW-R #2: This is optimized for low dynamic range programs, such as rock music.
POW-R #3: This is optimized for more complex, high dynamic range program, such as orchestral music.


Hope that's been of some use.
Paul

[theblue1]

Originally Posted by BrianT

[snip]...
If you use your own ears to decide instead of just reading about the numbers (like most people do), don't tell anybody, because then, you are going to get an earful about how wrong that is.
...[snip]

I'm probably someone who has garnered a rep as a guy who will argue at length against what he feels is an illogical or scientifically untennable statement.

But I want be very clear that I think ears are the second most important tool an engineer, producer, or musician has.

The only thing more important, seems to me, is the brain.

I use my ears all the time. I don't engineer by number or eye -- although when things go wrong I use all the information at my disposal to try to figure out why and make sure I can avoid the issue in the future.


The only real problem I have is when someone confuses his own subjective evaluation with objective reality or when he makes public statements which are manifestly not true and/or are contradicted by the body of scientific data.

Ears... where would we be without them?

More important than any test gear... but when things go wrong, it sure helps to have that sensitive test gear because the ears are amazing and they can tell you an extraordinary amount -- but they are far from calibrated, standardized test gear. And that's a good thing.

But don't get your ears confused with test gear. And don't get a personal opinion not backed up by by facts confused with objective reality.

:-)
__________________



Anyhow, on the OP's question, but first: to dither is to add noise; the OP probably meant reducing word length, ie, reducing bit depth.

Using longer wordlength (24 bit instead of 16) buys you greater dynamic range above the digital noise floor.

If you record hit a converter set to 16 bit AD with a well optimized signal, you probably won't suffer any significant degradation of your signal.

That said, your DAW is likely going to process the data internally at 32 bit float (for most modern DAWs) and save its work files to that format, anyway, no matter what AD bit depth you use.

And using 24 bit AD can give you enough headroom that you don't have to sweat getting your gain up at the AD input -- you can worry about getting an gain structure in your analog input chain that's optimal for your analog chain -- which can be critical to getting the best sound out of your preamp.

[jmarkham]

There mere fact that in 2008 we have to dither for a format
created 25 years ago pisses met off. The 44.1 / 16 bit CD audio
format should die now.

jeff

[Dean Roddey]

It's a crazy tradeoff to record in 16 bits just to avoid dither. The practical benefits of recording at 24 bits (in terms of being able to get good dynamics without worries of overs and processing the content with huge dynamic resolution) far outweights the tiny down side of dithering to 16 bits.

[JPZ]

Concerning dither you all should take a look at Steven Massey's blog on his website about dither. It's kind of how I've felt for years. As to 24bit vs 16 bit I will say that recording at higher sample rates made less of a difference to me than when recording at 24 bit was finally an option. That's what made leaving analog for good possible.