Is 16BIT enough vs 24BIT

=D

Knew you were joking too. What ullanta was explaining about bit depth is very correct.

Bit Depth, one of the most explained topic in Digital Audio, refers to the amount of resolution present within a passing signal that's being converted by the analogue to digital converter. For example, you record your signal into three separate recorders, one at 8 bit, 16 bit, and another in 24 bit. All these signals have the same level and came from the same source. Within a given second, which audio is "finer" in detail? The 24-bit because there are more "steps" and "resolution" in the same matter of time and loudness compared to 16 or 8. Each step represents 0 and 1. Same goes for walking a flight of stair with 8x 2' increments, 16x 1' increments, or 24x 2/3'. Which one is finer in detail in terms of shape? This is resolution.
 
Last edited:
Petrus said:
Study Fourier transformations; small amplitude changes, appearing alone or piggybacking a loud signal, are those quiet sounds that 24 bit system captures better (or at all) compared to 16 bit. That's dynamic range (call it resolution if you wish). This simple fact seems to escape you. It is not only the totally quiet parts, in AD conversion the small changes are low level signals, big, loud changes are high level signals. They are all present at all times in a natural, complex sounds we have.
Click to expand...


This is interesting... it agrees with everything I've been saying except for your interpretation of low-level and high-level signals. A "small amplitude change piggybacking a loud signal" is still represented in the high end of the dBFS scale - that is, if you have, say, a 400 Hz tone at, say, 80 dBSPL which the converter changes to -12dBFS, and a 1600 Hz tone at, say, 40dBSPL, the waveform will indeed be piggybacked...

so where the 400 Hz signal alone peaks at -12dBFS, the combined signal would have extra little "bumps" that peak at a higher level (something like -12dBFS + -24dBFS, or -10.1dBFS). Depending on how the waveforms line up, some of the small "bumps" will be on top of the "hills", some will be in the "valleys", near 0.

What you say applies to the bumps in the valleys. If the bumps are very small, then these fluctuations of signal level are indeed "low level signals" (compared to the hills).

But the bumps on the hills are small fluctuations in high-level signals. Not "quiet sounds", not "low level signals" - these bumps on the hills are fluctuations that take place above -12dBFS, in the top half of the dynamic range - relatively high signal levels.

"Small differences in signal level" is an issue of resolution. "Low signal level" is an issue of dynamic range.
 
(As an aside, I just want to say that while the principles of my math are correct here, the details may be off. Specifically, I don't know the exact details of the representation for any given system, and it likely has positive and negative range (that is, -0dBFS to 0dBFS). This doesn't change the resolution vs. dynamic range issues, and clutters up the examples a lot. If anyone really cares, I can clarify this with the EXACT representations used by different systems.)
 
Study Fourier transformations; small amplitude changes, appearing alone or piggybacking a loud signal, are those quiet sounds that 24 bit system captures better (or at all) compared to 16 bit. That's dynamic range (call it resolution if you wish). This simple fact seems to escape you. It is not only the totally quiet parts, in AD conversion the small changes are low level signals, big, loud changes are high level signals. They are all present at all times in a natural, complex sounds we have.
Click to expand...

Maybe if you spent less time studying FFT and a little more time actually dealing with sound your comments would be more useful.

Try maybe an audio reference book instead.

Dynamic Range:
The difference, in decibels, between the loudest and quietest portion of a program is known as its dynamic range. Sometimes, the quietest portion of a program will be obscured by ambient noise. In this case, the dynamic range is the difference, in dB, between the loudest part of the program and the noise floor. In other words, dynamic range defines the maximum change in audible program levels.
Dynamic range also applies to sound systems. Every sound system has an inherent noise floor, which is the residual electronic noise in the system. The dynamic range of a sound system is equal to the difference between the peak output level of the system and the electro-acoustic noise floor."
Yamaha Sound Reinforcement Handbook, pg33

THAT is a simple little fact that you can't seem to grasp.


One more thing, if you insist calling this "resolution", what is the measurement used? I have never seen anythig else but dB (as in dynamic range) connected with bit depth.
Click to expand...

Really? dB is used for a lot of things other than dynamic range and I have never seen a device describe it's bit depth in dB.
 
I do wonder why is it that a few theatricals were recorded at 16bit audio into camera!. This info defies what my instructor told me (i know we're not perfect),
But isnt 16bit roughly adds up to 96db? is this enough to represent dialogue in theatrical films?

This makes me wonder if double system is even necessary anymore
 
ullanta said:
Hey Alan! How ya doing?

Right on, except I'm not sure what you mean by:



Each step represents 1/2^(bit depth)dBFS!
Click to expand...

You're right, I'm trying to dumb it real down.

I'm not sure why Sound Devices mentioned that there's no difference between 16 and 24 when turned all the way up, but there's a HUGE noticeable difference between the 2 when turned all the way up and down. Decay of an instrument can be heard more smoothly, for example.

The MOST valuable source I have is this:

http://www.24bitfaq.org/
 
Each BIT represents a 0 or 1. But a bunch of n bits can represent 2^n STEPS. So, in a 4-bit system, there are 16 steps, each "step" is 1/16 of the representable range. A "STEP" in 16 bits is 1/65536 of the representable range. The STEPS are fixed size; there's no "or" there.
 
killfilm said:
I do wonder why is it that a few theatricals were recorded at 16bit audio into camera!. This info defies what my instructor told me (i know we're not perfect),
But isnt 16bit roughly adds up to 96db? is this enough to represent dialogue in theatrical films?

This makes me wonder if double system is even necessary anymore
Click to expand...

16 bit was the standard for production of great soundtracks for many years. 16 bit is still the standard for distribution. It can be very good. That said, 24-bit gives you better quality and more flexibility - most specifically, the ability to leave a lot more headroom and decrease the risk of getting unusably-distorted audio. So given a choice, I'd say always use 24 bits. But 16 bits can work for most things.

Recording direct to camera, it's unlikely that the sample size will be the most important factor in your audio quality. Double system can give you much better analog and conversion quality. As I said, even a crappy little MicroTrack fed by a MixPre sounds a lot better than a DVX or HVX fed by the MixPre - and these cameras have pretty good audio sections!

For all the theoretical/technical stuff we're having fun discussing, none of us are saying that 24 bit recording doesn't provide better sound; ands none of us are saying that the analog stages aren't, in general, more important. Just having 96dB of dynamic range representable by your camera's data format doesn't say anything about the actual audio quality you'll get - it's a theoretical limit. It's just like you can buy a blank can of film which COULD hold "Citizen Kane", but that doesn't mean that's the result you'd get by using it.
 
But isnt 16bit roughly adds up to 96db? is this enough to represent dialogue in theatrical films?
Click to expand...

Certainly it is. This whole discussion isn't really about if 16 bit is good enough after all that is what is shown in the theatre if your looking at Dolby Digital. The question is whether there are any benefits to 24 bit recording and if so what are they. At least that's what it has evolved in to.



This makes me wonder if double system is even necessary anymore
Click to expand...


For who? If your shooting 35mm it is. RED yes. A camcorder with good audio section... yes and no. If you know what your doing maybe not. Oleg is a firm believer in single system. My experience is that the DP or camera operator can really screw you up single system so I always run a backup that I have control over. But that isn't a quality issue. In post 99 times out of a hundred I use the OMF and don't go back to the master tapes unless there are problems. Then it's nice to have a bit more meat to work with. Titanic was recorded on DAT 48/16. THe only thing 24 bit was the music that came in on Sonic Solutions DAW (Pro Tools didn't do 24 bit at the time).
 
ok nice info thank you,

1. Why not just use the tools optimized for the jobs, a sound recorder for sound recording, the camera for videotaping...let the baker bake kind of idea

2. How Are most feature films today recorded today, on the nagra, dat? 16 or 24bits?

3.Is it not better to have a 24bit master in case your production is picked up,
say (one day) your feature is picked up for theatrical, tomlinson holman states that 24bit would be better for theatrical productions?
 
Don't put too much credit in anyone source, including me. TH has had some foul ups also, not that I'm disagreeing with him but seek a lot of info and then use common sense. If your production is picked up your track could be on a wire recorder, your picked up. If they are going to redo the sound thee in for some serious $$$ whatever your tracks are on. That would not be a reason to use 24.

1 - agree 100%

most features are recorded on some kind of HD recorder. Nagra missed the curve and came in late with a (very nice) two track recorder at too high a price. They were sort of competitive for 9-18 months then the price dropped out of the recorder market with the FR-2, SD recorders, Marantz, Tascam, Zoom, M-Audio etc. The Nagra is still nice but at 7K with out TC for two tracks....
I believe they are coming out with something to compete with the Deva but I think they missed the market. In the US Deva is probably #1 with direct to computer (Metacorder, Boom Recorder) coming in second. Cantar is pretty big in Europe.
DAT is dead. Nobody is making DAT recorders anymore and expect tape to start getting hard to find. Nagra analog is probably extinct for production sound though it is still used for FX recording for some things.

I suspect (though don't know for a fact) that 48/24 is the standard rate for production work in film features. For video it's probably whatever the post is going to be, generally that would be 48/16. I don't know any working FX recordists working at less than 48/24 and a bunch of the guys are recording at 96/24, though the logic for doing so may be a bit dodgy.
 
killfilm said:
ok nice info thank you,

1. Why not just use the tools optimized for the jobs, a sound recorder for sound recording, the camera for videotaping...let the baker bake kind of idea

2. How Are most feature films today recorded today, on the nagra, dat? 16 or 24bits?

3.Is it not better to have a 24bit master in case your production is picked up,
say (one day) your feature is picked up for theatrical, tomlinson holman states that 24bit would be better for theatrical productions?
Click to expand...

money is what make the production not to record separate .
DAT is probably 10-15 % , but not because 16 bit issue , no spare parts , less facilities who keep these ,
16-24 probably depends on post production- what they ask or not ask .

there is a chance of 0.1% that someone would pick you in the feature so if it cost you more - keep the money :-)
 
Didn't we have the 16 v 24 bit just last summer?

Killfilm,

What matters if your audio will be good enough for festivals and distribution is not the bit depth or type of recorder used. Good audio is good audio. The number and types of mic used to get good audio would be a more pertinent question but still more or less moot to what YOU need for your production.

That would be a minimum of one person on your set that understands how to pull good audio out of a location and scene. This does not mean you need a pro, although a pro would be you best choice.

From your Critique thread you appear to be on the right track.

Good audio can not be bought like a product; it is a process that happens. Good equipment and good hands nuture the process along.

To that end, you need a budget for audio and then a knowledgable overview of everything you will need to fit into that budget. Seperate recorder? Good idea, but do you have the more pressing equipment covered already? Making equipment decisions a la carte will either drive you over budget or leave you shorthanded on equipment.

If you already have a 24 bit recorder I would recommend 24bit mode. While in practice you won't utilize the full dynamic range of even a good ~75db 16 bit system, you will lose less of this practical range when you "sacrifice" the top 12 or 20 to headroom of a 24 bit recording. Secondly, as already mentioned, you set yourself up for cleaner post production processing. Let's face it, small budget indys, by their very nature, tend to need more post work then average. With gigabytes of storage costing between pennies and a few bucks, depending on your storage medium, the cost is negligable.

Cheers
 
good points everyone, SO, why is there a lot of features that have recorded into camera with major success such as the DVX (iraq in fragments), no seperate recorder, no 24bit, and the audio is nothing short of awesome. Some people advised never to use the DVX/HVX preamps, but are proven wrong again.
 
1) "Content rules"? I haven't seen "Iraq in Fragments", but I think viewers are more forgiving (even exepectant of) compromised sound in war zone documentaries...

2) How did you view it? Sound that's "awesome" in even a pretty good home theater setup can fall apart in a real theater.

3) But, look. Most of our favorite movies were made with equipment that's far worse than what's cheaply available now. Many of them are mono low-frequency-range analog. Most treasured music recordings were made on equipment with what would now be called ridiculously low-spec equipment. Listen carefully to what everyon'es saying here: first comes having enough people to properly handle audio; then comes technique, skill, artistry in the audio process, after that comes equipment specs.

Given a choice, all else being equal, most people would of course choose the better equipment. That doesn't mean that it's absolutely necessary to get a good result. Certainly, new equipment purchases tend towards the higher-spec equipment. But there's no rush by seasoned pros to ditch the tried and true 16-bit equipment they have.... it's certainly a slow change as old equipment breaks down and needs to be replaced; or driven mainly by other considerations (workflow, reliability) than quality differences. We're reaching (or have reached) a point where the majority of usage seems to be shifting to 24-bit, which makes it seem the minimum standard. For many people, this alone will make it necessary; people without audio experience will occasionally ask you "what's the audio format" and make judgements based on that, just as they do about film vs. video, HD vs HDV vs SD, etc. But people have and certainly some will continue to produce great results with "crappy" equipment like DAT, camera audio, and analog cassette decks.
 
killfilm said:
Some people advised never to use the DVX/HVX preamps, but are proven wrong again.
Click to expand...

People who say "NEVER do" anything are ALWAYS wrong. ;-)


"Try to avoid the HVX/DVX preamps" is good advice. But everyone's used 'em in a pinch, and they can provide usable results, especially if you know what your doing, get your levels right, and don't need a lot of leeway to "fix it in post". But a decent external preamp or recorder will do better, all else being equal.

You seem to be drawn to extremes (perhaps this teacher you refer to has inspired that!), but there's no such thing as best, never, always, necessary, etc. Especially in real-world production, you're making 20 compromises a second. We strive for the best, but do the best we can in the situation. In the end, ingenuity and flexibility will trump equipment. You can be sure that someone right now is shooting a film on a cell phone that will blow you away!
 
Last edited:
To expand on the above if I had the choice of doing my own production recording on what ever is the latest and greatest or having Oleg do it with a cassette deck (assuming the same mics pres etc) I would go with Oleg. I know he has more experience doing production than I do and I'm pretty sure he's a perfectionist. His experience is worth a LOT. I wouldn't make the same deal in post and maybe not in FX but I'm sure he's a better production mixer than I am.
 
You must log in or register to reply here.
Back
Top