Home > JoeRants > Audio Stuff, pt 3

Audio Stuff, pt 3

October 24, 2012

All right, in the last couple of blogs in this audio series (part 1 is here, and part 2 is here – go check ’em out!), we talked about getting your microphone and your recorder to work together, because microphones audio levels are all over the map.

Now I want to switch gears a bit and talk about what I look for in an audio recorder. And in the process, we’ll discuss what some audio features mean.

First things first, my audio recorder needs to use XLR inputs. Any audio recorder that doesn’t use XLR’s probably isn’t very good, because they’re automatically assuming that you want to convert your input from your mic (and almost any good mic will be an XLR output) to something else (usually an 1/8″ mini headphone jack). In general, converting is something to avoid if you can, because you’re invariably going to lose something in the conversion.

Without getting too technical, there’s a reason that pro audio equipment uses XLR connectors—because they help avoid interference in the signal, so the audio has less a chance of getting corrupted as it travels across the cable. XLR connectors are also strong – they’re generally made of medal and make a really strong connection that is hard to accidentally interrupt. There are high quality XLR to 1/8″ mini converters, but again, you don’t want to have to convert if you can avoid it.

Second, you want a recorder that can record uncompressed WAV or AIF files. There are a lot of mp3 recorders out there, but mp3s are compressed audio. The way mp3’s work is they throw out information in the audio signal that is considered disposable – the stuff that you’re less likely to notice if it’s missing. But when recording production audio, you’ll often need to filter and process that audio in post-production. And when you do that work, you’re often pulling out parts of the signal that might be that same, less noticeable stuff  – the exact parts of the sound that mp3s compress heavily. Recording in mp3s really limits your sound designer’s ability to bring out the best in your production tracks.

Third, I look for a recorder that can record 24bit audio. When working with uncompressed audio, there are two things that affect the technical quality of the recording: the audio bit depth and sample frequency. I say technical quality because many things can affect the aesthetic quality – like recording the dialogue next to an air conditioning unit or passing train, or pointing the mic away from the actor’s mouth, or…many, many other things.  We’ll talk about that in a future posting

You might ask – hey, what are bit depth and sampling frequency? Well, here’s the answer.

Sampling frequency is simply how many recording per second are made. It’s direct video counterpart is the frame rate. There’s a sweet spot of frame rate that hits at about 24 fps. Less than that, you tend to notice that there are individual frames playing in sequence – it doesn’t look continuous. More than that, well, different people will tell you lots of different things (google “Peter Jackson frame rate”). But 24 fps is generally considered “good’ for a frame rate.

Most audio recorders can record at a sampling frequency of 48k, as well as other rates. The “k” in 48k stands for 1,000, so 48k = 48,000 recordings (or samples) per second. That’s a lot of recordings. It’s also generally referred to as 48kHz, with Hz meaning “hertz” which means “per second.”

Some recorders go higher than 48k (usually 96k). There are instances where 96K is good for recording, though I tend to find it unnecessary for most of the work I do. 48K is a really good sampling rate – it’s more than twice the frequency of the highest point of human hearing, and there’s a bunch of math that says that any digital recording needs to be twice the frequency of the highest pitch you want to record. So a 48K recorder should be able to record pretty much anything the human ear can hear. Going to a higher frequency will get you more info (which can be good when you’re using certain filters in post), but often, it’s overkill.

Bit depth is the other part of the quality equation. CDs play audio at 16bit, which is pretty darn good. Most video cameras record at 48kHz and 16bit. But I look for recorders that can do 24bit. There are also ones that can do 32bit and 48bit, but they tend to be out of my price range.

If sampling frequency is like video frame rate, then bit depth is like resolution. How much detail is in each of those 48,000 recordings? My standard metaphor is that 16bit is like standard def TV (except the of 16bit quality is better) and 24bit is like high def. There’s a lot more info in 24bit. The SD/HD metaphor falls apart when you then decide, well, I should be playing everything in 24bit, if it’s so much better. Most people won’t notice the difference between hearing 16bit and 24bit audio. There are people who can, but they’re audiophiles. We don’t talk to those guys, the conversations get a bit… difficult.

So why do I want to record in 24bit if I’m only going to play back in 16bit? Two reasons – the first is practical, the second is a rule of thumb.

First, I want 24bit audio to get that extra detail in my recording so when I raise the volume, and add a bunch of filters to bring out different qualities in the sound, I’m working from a more detailed original audio, so there’s info to play with. While upping the sampling frequency doesn’t get you that much more usable detail, upping the bit depth does – by a lot. I can really manipulate the sound without losing too much quality (if any!) at 24bit or higher.

Secondly, 24bit audio tends to mean you’re going to get a better recorder. That’s not a hard and fast rule – every single 24bit recorder isn’t awesome, but you can filter out a lot of lousy ones by only paying attention to the 24bit ones. Same thing with any recorder that doesn’t have XLR inputs.

Lately, I’ve been using Marantz recorders. I mentioned the PMD661 in my last post, and I really like it. It’s compact, records 24bit, and has very little noise – most recorders tend to add some “noise” into the audio as they record – it’s hard for them not to introduce some, but the PMD661 is very quiet.

My favorite feature that is new to digital audio recorders (I come from back in the days when we used to record on reel-to-reel Nagra recorders) is referred to on the Marantz as “DLmono.”  DLmono records a single mic on two channels. On the 1st channel it records the audio exactly at the level you set it. On the 2nd channel, it records the same microphone signal, but lowers it, just a bit. This gives you some safety that if the audio gets too loud and overmodulates, you’ve got a lower-level backup on the second track that likely won’t overmod. You can ride your recording levels a bit higher because you’ve got a safety recording on the other track.

All right, that’s enough for tonight – Next time, we’ll get into what recording levels you want, and a few other toys that you need to invest in beyond a recorder.

  1. No comments yet.
  1. November 12, 2012 at 8:41 pm
Comments are closed.
%d bloggers like this: