20bit vs 24 bit, how big is the difference?

chambers

New member
Hi everyone,
I am just starting to enter the recording world. Right now I am looking at what DAW to purchase. I am debating on whether 24 bits is worth the price. What is the sound quality difference when going from 20 to 24 bits. I understand that 24 bits has 8 times as many possible values for a data point, but I'm not sure how much this affects the sound. Has anyone worked with both types, or upgraded from one to another? Would it be worth it for a beginner to invest twice as much to get the 24 bits?
Thank you all for your help,

Chambers
 
Twice as much?????
What exactly are you looking at? 24bit is pretty much the standard today for multitracking software or soundcards, unless you use a soundblaster which is 16bit...where have you seen 20bit soundcards or software??
 
dont forget to consider the quality of the converter...id rather an awesome 20 bit converter than a crappy 24 bit....
 
Thanks for your replies.

I am looking at a digital audio workstation (e.g. vs840gx, tascam 788, etc.). What I mean is that the 840 has 20 bit converters and the tascam has 24 bit converters. I was just wondering, from the number of bits in the converters, how big would the difference be in sound quality. Especially since this is my very first experience in recording, will a big price difference be worth to pay for the extra bits?

db,
The tascam with a CD-RW (so I can keep a master of the recordings) costs about twice as much as the vs840gx whose zip format in itself gives you a master.

Gidge,
How can I find out the quality of the converters?

Chambers
 
Quote: "What is the sound quality difference when going from 20 to 24 bits. I understand that 24 bits has 8 times as many possible values for a data point, but I'm not sure how much this affects the sound. Has anyone worked with both types, or upgraded from one to another? Would it be worth it for a beginner to invest twice as much to get the 24 bits?"

First, "8 times etc" is not correct. Its way way more than that. If you, for instance, go from 16 to 17 bits you virtually double the sampling points - 17 to 18 again, etc. At 24 you arrive at a sampling rate which is an accurare representation of the analogue equivalent, providing you record at a high clock rate.

20 bit was a more or less stop-gap sample rate, heavily promoted by Alesis. Its on its way out, there are very few manufacturers catering for 20 bit (converters etc). 24 bit will be the standard for some consideral time, as development will concentrate on clock rather than bit rates. All media is also more and more focussed on 24 bit, whichwill replace 16 bit CD's.

It is perfectly feasible to record good sounding audio at 20 bit. The difference however becomes very apparent once you start processing audio, as this is where you need high quality in order for a digital signal to retain its quality.

Finally, I don't think its correct that a piece of 24 bit gear costs 2 x its 20 bit counterpart?

Conclusion - if you can affort it, go for 24 bit gear.

I did a bit on how to evaluate converters, but I don't remember where it is. Do you Blue Bear?? I think you've got a better memory - all those drugs in the '70's :)
 
The price difference may not be solely due to the converters. For instance I would consider the ZIP format less desireable than the internal hard drive. ZIP offers fairly limited storage space at relatively high price per disk. Read the fine print in the specs for total track minutes per disk. I think you can only get acouple of 8 track songs on a 250mb disk.

Twist
 
Thanks for the info. This is helping me a lot in making up my mind. Now I have more factors to weigh in. Hopefully, it will all eventually make sense to me.

Sjoko,
The 24 bit doesn't really cost twice as much, but I think (and I may very well be wrong) that the only way to get your recording in multi-track form out of the internal hard drive is through a CD-RW which costs an additional amount. I really know little about recording techniques: what do you mean by processing audio? Is this after the recording? Or before? What type of proccessing? Please forgive my ignorance in this subject.

Twist,
With the internal hard drive unit, how do you make a copy of the song in its multi-track form so that you can edit later on without having to keep it on the hard drive? (hope that made sense)

Chambers
 
Yeah, that makes sense. What you want to do is save it as data on a cd. Then you can reload it later and it will still be 8 (or whatever) tracks so you can edit, remix, or rerecord any track.

Twist
 
chambers said:
Sjoko,
The 24 bit doesn't really cost twice as much, but I think (and I may very well be wrong) that the only way to get your recording in multi-track form out of the internal hard drive is through a CD-RW which costs an additional amount. I really know little about recording techniques: what do you mean by processing audio? Is this after the recording? Or before? What type of proccessing? Please forgive my ignorance in this subject.

Chambers, just about ever digital audio system, be it a workstation or software, has the ability to record at different clock speeds and bit rates. Included with this are programs which allow you to convert the audio files from a high to a lower bitrate, a process called dithering. CD's are 16 bit, everything recorded at higher bit rates has been through a process like this, be it 20, 24 bit.

What I mean with processing audio is the manipulation of a recorded audio file. be it adding an effect like reverb, compression, whatever, to individual tracks, or to a whole mix.
This manipulation requires high resolution in order to retain sonic quality. The lower bit rate, the worse it sounds. That is why high end equipment often processes at, for instance, 40 bit floating point, which is way over the 'normal' quality printed to disk.

and don't worry, we're here to try and help, certainly not to think of your questions as ignorant
 
Hey bear , I know why I couldn't find it - I didn't post it.
Here is part of a rough transcript of a presentation I did a while back, hope its of use: (I didn't check it - no time!)


As some of you already know, I have been / am involved in testing and development of some recording gear for various companies. My main interest is digital stuff, ‘coz that’s the future. By now I’ve got a pretty good idea what you can and cannot do with analogue, so that has lost its immediate appeal for me, I rather concentrate on making the future sound gooooood!


What makes digital audio sound good?


The basics:

Digital audio quality depends on bit rate, clock speed, quality of conversion and clock accuracy. Which one comes first? What is the most important? No choice, really. That must be one of the only choices in life that offers no choice, they are all equally important.
I won’t go into all the boring technical stuff here, but CD quality, 16 / 44.1, sucks major and sounds like sandpaper, it never was an alternative to analogue. Consider that every extra bit almost doubles the quality of audio, and you’ll start to realize that 24 / 48 is a pretty cool improvement. Now we are back to, potentially, good sound quality.
I won’t bother to go into the 24 / 96 stuff, despite all the shouting you hear about it. People say its great, wonderful etc? Cool! I’ll be the first one to use it – properly – once all the technical details have been solved, and there are plenty to be solve before you can complete a project to perfection at 24 / 96. Lets leave it at this; if you listen to a DVD A, with L,C,R at 24 and RL, RR at 16 – you are listening to a project which has been mastered TO 24 / 96, not likely one recorded IN 24 / 96.

So, that’s the first thing over with. To get a good quality recording you have to think 24 / 48. And don’t forget to dither it down properly to go down to CD quality, or you’ll just fuck all your good work up again.

You have to get your analogue signal to digital, and back again. In other words, you need converters. Not just converters, you need good converters. Let me help you out of your misery here. Good converters are getting cheaper. Now for the depressive bit, a converter is circuitry around a chip, or chipset, audio quality depends on the design and inter-relationship of these 2 elements. I find it pretty amazing therefore that some companies are now offering “high quality” converters at a retail price marginally above the wholesale purchase price of one single decent chip?
Unfortunately, it is possible to build a converter cheaply, with an apparent ‘decent’ sound, against what seems to be a good spec. Unfortunately you could also design a converter at a much higher spec, which sounds like total crap.
More about converters later.

Next, the clock. Another bit of free news for you. 1. If your heart doesn’t beat in a regular pattern, you’re in deep shit. 2. If you’re a drummer and you can’t keep a regular pattern, you’re likely to cause #1 to the other band members. 3. If your music doesn’t keep a regular pattern, it will sound like crap.
Same goes for all your digital recording gear. If it doesn’t ‘keep the beat’, it will discombobulate.
Let me spell it out once and for all:
THE QUALITY OF YOUR EQUIPMENT’S DIGITAL CLOCK RELATES DIRECT TO THE QUALITY OF YOUR SOUND. Not very often understood this, most people cannot see that a digital clock effects the quality of sound. To explain this as simple as possible, please note that you are NOT recording sound, you are recording DATA. This data needs to be kept in-line, with every sample accurately in its place. If not, when this data is turned back into music………………. Get the picture?

Lets put another ugly baby to permanent rest. NO! The clock in your gear is not good. Don’t bother coming back at me with “but my…..” It won’t wash. As in converters, some good news. You can now get a highly accurate external clock for less than a third the cheapest ones did cost a couple of months ago, the quality is going up and the prices are going down.

Next fairytale. You distribute clock with your audio signal, just daisy-chain from your machine down. Yes, that is what everyone does. And no, it does not work, it sorta barely scrapes its act together. If you hear a clock pulse without audio coming out of your generic build-in generator you’d be in shock about how uneven it is. If you’d listen further “down the line” – you’d call 911 and would try to revive it with an electric shock.

In order to get an accurate clock you need a quality external sync generator. In order to ensure all equipment runs to the same clock at the same time, this sync generator need to feed each piece of gear individually. Another thing the industry has yet to address properly – you simply cannot transmit audio plus a clock through AESII, you really should keep the audio and clock separate.

PRIORITIES:
Now this bit is, I think, important.
I constantly hear “I’m saving up for a new microphone” and “Next on my list is a good tube preamp” etc.
Let me ask you one simple question. Would a top quality microphone and pre make a bad clock and bad converters sound better?
If you want to get something resembling “pro quality” (whatever the hell that means) sound, you will need to get your priorities right, address the quality of your recording system first, before spending anything on anything else.
If I had a DAW, my first priority would be a good AD converter. After that, a good DA converter (often forgotten). Next would be a good clock. Only after that would I consider improving other things.

Example. I could record a vocal through a $4000 mic and a $3000 pre into, for instance, Pro Tools, using a 001, or even a Digidesign 888.
I could record the same vocal again, this time using an $89 Shure, into a small Mackie mixer as a pre, into a good AD converter, with Pro Tools clocked by a quality external clock.
My choice would be the latter, I’d get a better overall result. On top of that, I have just saved myself about $4000. (OK – now I’ll go and buy a good mic).
Get my drift?

That’s the first bit. For those of you who think I’m just bashing gear manufacturers, no I am not. Everything is build at a cost, if it wasn’t most of you would never have the chance to record something sounding remotely as good as you are already doing. Apart from that, this is new technology, it is developing, and more things are becoming apparent all the time.
To illustrate the above, Yamaha’s brand new digital console, the PM-1D, costs $120.000
It needs to run with an external (non-Yammers) clock, or its sounds, well….. let’s say not very good to be polite.

Next. Back to converters. If you remember, its possible to build a converter with high specs on paper, but one that sounds like crap. This begs the question, how the heck can you test / tell a good one from a bad one?

Prior to listening to a converter, lets ask the question: “what are you listening for?”
A good quality converter will show improved debt and detail. You will hear increased clarity and apparent volume in the whole spectrum, and particularly in the low and high ends. A track will seem more transparent, stereo imaging seems improved, you can clearly identify every detail (which also means – you can clearly hear all the F’ups in your playing / recording / processing / mixing J
A note to the above: This is why D/A converters, often forgotten, are the most essential part of gear for a mixing engineer. Without a good D/A converter you can’t hear half as much.

(Just to show you how people who SUPPOSEDLY know about this stuff screw up – I couldn’t wait to get my monitors going in the new control room. Set everything up, tested, calibrated, but I’d lend my D/A converters out. Got them back – had to do the whole thing all over again. 2 days wasted.)

Use one of the tracks you have recorded and are familiar with. Play it, as stereo tracks, analogue out from your system, to the inputs of the converter you are testing, recording the results back to your system. Make sure you set your recording levels at the same volume ON YOUR METERS as the tracks you played into the converter.
Now you will have the song 2 x, and can ab the results (you will have incurred some latency).

Another way. Select a couple of nice, full synth patches and drum loops. Open a new session and record them on two tracks each, one through your stock converter, once through the one(s) you are testing. Make sure you pan every stereo track l/r when you play back your recordings.

If you are testing more than one make of converter, things get more complicated. Some converters use the same chips or chip sets, like for instance the Apogee Rosetta and the Lucid 2496. The same dB headroom, similar overall specifications, do the testing as above and the result will sound very nearly the same. Yet, there is a difference, and above all, one is much more expensive than the other. How can you make the difference audible?
One way only, start tracking. Stacking things like vocals for instance. The more you stack, the more tracks you build, the more the difference will become apparent. The better converter will show higher definition, dynamics and stereo image.

Finally – if you do all the above WITHOUT a high end D/A converter, remember that whatever you are recording will sound a lot better than you can actually hear. So make sure when you print a mix, you do so without going through another conversion.
 
Last edited:
Wow, great info. I just looked at the Lucid converters and was flabbergasted! $700/ $600 for A/D-D/A converters! That's plain and simple robbery. Sure maybe those are some kick-ass converters, but poor college guys like my-self can't even dream of buying over $1400 worth of converters plus a $400-$500 word clock.

Since there are great conversations going on about converters I must ask this question to keep my sanity! Here goes, I am thinking about buying the Digi 001, but I have heard that the converters suck. I also heard a contradicting statement saying that the converters are not bad at all but pretty good. Who is right? I know it will not compare to "LUCID" converters but I need to know if it will sound horrible or acceptable. I don't need the best possible converters; just need good ones for my own recordings and maybe in the future I will start to record local bands. I wish I could try it out but I would have to buy it and if I find it sucks then return it with in 30 days or I'm out of luck.

Thanks for your help!
 
Well... robbery it isn't. Consider its nearest competition, which uses the same chip but doesn't sound as good, is $1000 more expensive.

Please realise that everything is relative. I have used converters with a 157dB dynamic range and practically no noise-floor. Never hear better. But they cost as much as a new Jaguar. Then there is this new Euphonics console. Really cool, and a snap at $400.000 Does that mean you shouldn't record unless you get hold of all that stuff? Hell no!!!

Get the 001, learn to use it, enjoy recording with it, its a really cool, good tool. Then later, if you want to improve, save some cash and buy some stuff to make it sound even better. Compared to much other gear in the same, or higher, price range, its a total bargain.
 
Hey sjoko,

Just curious, but are you advocating 24bit/48k over both the higher sample rates (88.2k, 96k) and the "native" CD sample rate of 44.1kHz? If so, may I ask why? It seems that either a higher sample rate would be better, or at least 44.1k would be better than 48k in the lower range to avoid doing SRC.
 
If I had it my way, I'd record at 192! The higher the better, no question about that whatever. 44.1 has really never been an acceptable rate, it simply is to low for good audio quality. 48 is good, a huge improvement over 44.1. 96 is simply the shit! BUT.... there are still a lot of problems associated with 24 / 96.
Without going into the technical detail, just look at the hard-disk recorders, Mackie, Tascam, Radar, etc. Everyone advertising 24 tracks at 24bits, 96 kHz, but the fact is, if you want to USE the systems at 96 ---- you only have a maximum of 12 tracks available, as non of the systems have the processing power required to run at 24 / 96. There are a millon reasons why doing a project (properly) start to finish at that speed is still a dream.
On top of that - ever calculated the disk space required to record at that rate?

Next, there really is nothing wrong with sample rate conversion, providing its done right. There are many reasons to record and process at the highest bitrate and clockspeed possible. Especially processing NEEDS to be done at a high rate to retain audio quality.
That is, for instance, why mainframe digital consoles often have specs like 40 bit floating point processing.

Some other things, 88.2 is not a commonly used audio clock speed. I run a SRC192 in my studio, which converts anything into anything (up to 192kHz!) instantaniously, without quality loss (other than that naturally incurred by reducing things to a lower rate / speed). It even allows me to work on a project with material in different bits / speeds simultaniously.
 
Clock, outboard converters, sync generator. Sjoko, this thread is gonna cost me a lot of money. That's the bad news. Still, I'm gonna give this thread five stars. I suppose that's good news. :)
 
Okay, let's see if I've got this straight. Generic clocks are shite. You in a position to name some reliable clock brands?

High quality external sync generator is necessary for good clock function, but the industry doesn't make any that address each piece of individual gear, which is what you really need. So why buy a piece of gear that doesn't do what it should?

About audible differences in converters: you talk about stacking up multiple tracks, and then the difference becomes apparent when you listen to the accumulated effect. But if you're *not* stacking up tracks, then it wouldn't make so much difference, right? (I go live to two all the time - guitar and voice - big deal...)

Finally, the 24/96 thing. The box studios they're selling these days that claim to track at 24/96 actually *do* track at that, right? And computer soundcards also, right?

I'm happy with my mics and pres right now. You've opened up a whole new area for me to come to terms with. I feel that all of us here at homerec.com are poised on the verge of greatness... :D
 
Back
Top