Basic Audio Settings in Editing
 

Should I just stick with 16 bit 44.1 khz audio settings in post if I'm using audio from an HVX? Is that the standard for DVD? Would changing it to 24 bit 48 khz not be a good idea since that isn't what my source audio is, right? I apologize for these stupid newbie questions.

It really depends. If you're a newbie and your projects aren't very complex, 8 tracks or less, then yes, 16 bit is more than enough. However, as your projects grow in complexity and breadth 24 bit audio can provide distinct advantages over 16 bit. Primarily in the dynamic range that larger words and frequencies allow when processing filters and effects. It's kinda like the difference between sd and hd. You can't push sd as far as hd because it breaks down sooner. Good Luck, Charles.

Is the HVX really 44.1 kHz? I had assumed it was 48 kHz.

In any case, I would set the project to the same rate as your most important source. That should avoid resampling while mixing. If you have a 24-bit audio card, set it for 24-bits. Otherwise set it for 16.

This might depend on your NLE. I use Vegas. As far as I understand, it takes the input files at their native rates and bit depths, and mixes and outputs to your monitors at the project rate and bit depth.

When you do your final render, the NLE should take your input files at their native rates, internally convert them to the final sample rate at high resolution - probably 32-bit floating point. Finally, it dithers/rounds/truncates to 16-bits, or whatever your render setting specify.

If it's like Vegas, the project setting have no impact on the final render. The project settings just affect how things are processed during editing/mixing/monitoring.

The HVX is 48k, I think there is just a typo.
Will, 16bit audio is just fine for virtually anything. Getting into 24bit can cause computer problems if you're not up to snuff and you've got lots of tracks.
Indeed, you can push 20bit or 24bit harder, but in the same breath, it should be mentioned that it will be several years before the number of 24bit productions remotely approaches the number of 16bit (or less) productions. Audio is (currently) always delivered at 16bit regardless of the format, unless it's a specialized format. Dithering 24bit to 16bit isn't as easy as some folks believe it to be.
FWIW, I've produced well over 1000 projects that include 400+ albums with a wide variety of artists in studios ranging from Power Station to Hit Factory, Motown to Village, Caribou to my basement. All were 16 bit until roughly 3 years ago, when *some* projects were done in 24, and even today, only a handful are done 24bit, most are 20 bit. 20bit, 96k audio is great for just about anything. 24bit/192 gets you into a whole different realm.
Be careful out there...

Spot...

How are ya? I have a question for you concerning 16/24 bit. In my experience,especially with the recording of music where the dynamics of the music is very high, the audio signal exceeds the bandwidth capacity of 16 bit recording. There are two solutions, that I'm, aware of, 1-record at 24 bit or 2- apply compression/limiting to the live recorded signal, before it goes onto tape. Which of these methods do you use/prefer? I've been using a Presonus limiter, but, lately it seems I can avoid the problem by recording at 24 bit and dithering to 16 bit in post.

Howdy Bill,
I've mentioned in other threads, I use a compressor between cam and mic in many instances, depending on what I'm shooting. 24bit or 20bit, I don't worry about it much. Then again, even without anything between anything, I'm always shooting for peaks and not for averages.
If I'm shooting in a controlled environment, I'll usually double and capture 20bit or 24 bit, depending on the device I choose prior to the shoot. If I'm in the field, which is most of my shooting, I'll shoot cam only, worrying about peaks.
There is a huge variety of battery powered compressors, some of them quite nice, for field use.
There are now very nice 24bit recorders for field use too. But sometimes, you're just a 2-3 man band, or you have a lot of cams to work with and the master can't be fed everything.

Spot...

Man I hear you! Pretty tuff to handle camera, mixer, audio recorder, compressor.... all at the same time. Busier than a one armed paper-hanger, I beleive is the expression. Problem is, it's hard to find an audio guy who's easy to work with...;o)

I'm loving my Sound Devices 702 recorder. Makes the job that much simpler....even tho' it's still not simple enuff. The JVC HD100 has an audio quality limiter in the audio circuit that works pretty well, altho' it's set to the -20dB broadcast requirement.

Thanx for your feedback, I think we're on the same page.

We're on the same page.
After our experience with a "union guy" in the Grand Canyon, I'd have given my left body part to have had you with us up there. In the dictionary under "terrible" you'd have found his picture, I think.
He comes from the school of broadcast, and claims that -20 is the level at which he's "always recorded."
I guess "always" means "I read this on a website or ina book, and it seems the right thing to do."
There were at least 3-4 times I listened (and watched) during playback, and wanted to toss him over the Rim. In a doc setting with heli time costing 2K an hour, it's sorta hard to go back and get re-takes, especially when 10% of the dialog is either inside a heli, or taking place as the heli is taking off.
I'll never, ever use a cold union call again. Ever. This isn't the first time we've been sent a moron, but it surely was the last. 10 days locked in the wilderness with a guy who whips out his card every time you get tweaked at his bad work, and no reasonable way to send him packing...Unforgettable.

I'm mainly a composer (my son does the video & editing), and I always use 24-bits for capturing tracks from samples. My computer is modest by today's standards, and it's never been a problem with oodles of tracks. By working at 24-bits up front, I don't have to pre-normalize my levels before capture. As long as the individual tracks don't clip, it's fine. I simply find 24-bits more forgiving at the front end.

During the mix the internal operations are likely at 32-bit float. SAW studio is a bit different - it uses 32-bit (or is it 64-bit?) integer. In any case, your DAW or NLE is almost certainly processing internally at more than 16-bits, even if you use 16-bit captures.

When ever you mix, you virtually always have more than a 16-bit result. When you multiply a 16 bit number (your signal) by another 16-bit number (your fader position) you get a 32 bit result. Adding two 16-bit numbers yields a 17-bit result. Incorporate compression, EQ and convolution reverb over lots of tracks and the result without any rounding/truncation/dithering can be huge.

If you are rendering stems or intermediate mixes that will be further processed, mixed or mastered, I recommend always saving 24-bits or more. As Douglas mentioned, dithering to 16-bits isn't transparent, so do it as little as possible.

Only at the final stage would you want to dither down to 16-bits for the final project. It's a good idea to save a 24-bit or better master, in case you will create different versions for TV, radio or the web. That way you only do a single dither down at the end.

I know some golden ear (deep wallet) types who dislike all digital dithering. They will use a high quality 24-bit D/A and record into a high quality 16-bit A/D - effectively dithering in the analog domain. I don't have the budget for boutique converters, and our product is for the web, so I go the digital route.

To summarize, there are three phases to consider with bit-depth:
1) Capture, where 24-bits is more forgiving of low levels (but optimized levels are still best)
2) Intermediate processing which WILL be over 16-bits for anything but a pass through, and
3) The final render, which has a bit depth defined by your product media, generally requiring a reduction in bit depth from the wide processed form to the final, often 16-bit, form.

Very nice, Jon!!

Hi,

If I'm going to be mixing, EQing, or otherwise messing with the audio later, I'll record 44.1 or 48 kHz 24-bit.

If not and it's not an "audiophile" sort of thing, I'll use 16-bit.

Regards,

Ty Ford

So the HVX records at 48khz? Anyone know if it's 16bit or 24bit? And you're saying if I'm going to be doing mixing and complex audio manipulations, I should stick with 24bit assuming my computer can handle it?

Yes, the HVX is 48k/16bit. virtually all of the low cost (and most of the higher end camcorders as well) camcorders are 48k/16bit, as either PCM or MPEG audio files. This is the digital video spec. Unless you move to very high end camcorders, you'll only be able to record higher sample rates on an external recording device.

Okay so it sounds like we have conflicting opinions about 16 bit vs 24 bit.

Even if all your audio is recorded at just 16 bit, editing your project in 24 bit in a program like Vegas could provide for advantages? Is this a true statement or not? Sorry I'm still a little confused.

Say I'm working on a feature film for mass distribution, is 16 bit acceptable? Spot it sounds like you definitely favor 16 bit.

Will,I don't think there's really a conflict. Everybody seems to agree that 16 bits are sufficient for the delivery of the final sound. For acquisition and processing/mixing, you have a choice of using more bits for your intermediate artifacts, or ensure with other means that 16 bits gives you enough dynamic range. Using a compressor (like Douglas Spotted Eagle suggested) is one way of doing the latter.

It looks like the advanced versions of Dolby Digital and DTS can actually use 24 bit sound, so depending on what the requirements are for such a project, I guess it's not impossible that 24 bits (or 96 kHz, for that matter) might be called for even for the final delivery. How many people will actually be able to hear a difference? I bet it's a very small number.

- Martin