Given that most prosumer HDV camcorders have evolved from existing prosumer camcorders, we can expect their internal audio components to retain the lackluster performance of their forbears. By lackluster, I mean that they fail to meet their own specifications. Published testing results have shown that the mic preamps in these camcorders do not have the frequency response or signal-to-noise ratio that "16-bit 48KHz" specifications would indicate.
Those are just numeric specifications of the amount of bits allocated to record audio. The quality of the audio that is recorded is a completely different story. From bad microphones and poor wireless systems to camcorder microphone preamps that hamper audio quality, the time has come for those striving for high-definition in their video to capture high definition audio as well. And it almost goes without saying, at this point, that you won't achieve that goal by relying on the onboard audio of the camera alone.
This article is written for videographers who have a working knowledge of audio but who are not audio engineers, do not have a dedicated digital audio workstation (DAW), and are not as fluent in audio editing as they are in editing video. I am one of these people. I like good audio, and I know it when I hear it. I can also recognize its absence. When I tried to record a music performance using my PD 150, I found the quality of the audio to be much poorer than most any dedicated audio recorder—even a tiny, consumer MiniDisc recorder.
Now as we move to HD acquisition, most of us are doing so via HDV. It uses 4:1 compression using MPEG-1 Layer-II—and this is just for two channels of audio. Canon touts its XL H1 as being able to shoot and record four channels of audio (previous XL models offered the same feature). This means a unique audio data stream that uses 8:1 compression. If you are going to compress the audio, why not use a more universal standard like MP3 (MPEG-2 Layer-III), where perceptual encoding and high bit rates can satisfy even the pickiest audiophiles in today's ubiquitous portable MP3 players?
Portable Audio Recorders
Portable audio recording devices are designed from the ground up to do one task, and do it well: record clean, high-quality audio. Recently, several manufacturers brought to market small, portable, solid-state digital audio recorders that offer high-quality audio in convenient packages. In fact, they can go well beyond 16-bit 48KHz, all the way to professional audio-level 24-bit 48KHz WAV files. Just as with oversampling on a CCD, 24-bit audio recording enables you to capture, edit, and deliver some of the most pristine audio you have ever recorded.
The one challenge here is that these recorders require the user to go back to the "film-style" double system for recording audio. That is, using a clap slate and syncing up the audio in post. But in reality, this is far easier than it ever was now that we have visible waveforms in our nonlinear editing software. Even though HDV tape will still be digitized in real time, we can download hours of HD audio in just a couple of minutes. It is worth the effort to step up both ends of your production when the consumers of your product will be enjoying it on digital screens and high-quality surround-sound systems that reveal the effort you put into the production.
To survey the current crop of HD-capable audio devices that are most likely to fit both the workflow and budget of event videographers, we gathered three recorders with comparable price points: the Edirol R1 ($429), the M-Audio Microtrack 24/96 ($400), and the Marantz PMD660 ($499). Each of these are two-channel recorders capable of recording uncompressed audio in WAV files, as well as compressed audio in MP3 format. True purists will opt for uncompressed recording, which requires adding a few very large compact flash cards. For our test, we opted for high bit-rate MP3 recording, which tests have shown to be aurally indistinguishable from uncompressed digital audio. This allowed us to use just one, much more affordable, 512MB card. We compared all three of these recorders by using them to record a live musical performance.
The Edirol/Roland R1 is the first unit we received. It is a compact flash-based audio recorder that uses two readily available AA batteries. You can use rechargeable AAs, but the R1 cannot recharge them. Using the included AC adapter bypasses the batteries. The R1 is about the size of a small Betacam cassette. It features both line- and mic-in, as well as built-in stereo microphones. It has an input-level adjustment dial that is always active. The screen can display an audio meter, but indicates the average of both channels. It has a limiter you can turn on or off.
The R1 can record WAV files in 24-bit and 16-bit at a sampling rate of 44.1KHz. It can also record 16-bit MP3 files at 64, 96, 128, 160, 192, 256, and 320Kbps. Compared to the other recorders we tested, the R1 offers good flexibility in terms of uncompressed/compressed, and lets the user determine the data rate to fit on the compact flash card. You can use cards as large as 4GB, but the largest single file you can record is 2GB. After that the recorder will stop. With a 24-bit WAV file, that's 125 minutes. With 320Kbps MP3, that's 831 minutes (13 hours). This should be enough time for any continuous sound recording.
The clear and easy-to-use buttons had us going through the menus very quickly without the manual. However, because the display has only two lines of text, we had to crack open the manual to find out what some of the abbreviations meant.
The manual touches on the many effects this little unit is capable of during both recording and playback. These include Easy EQ, For Speech, Voice Perform, Editable EQ, Noise Reducer, Hum Noise Cut, Reverb, Mastering, and more. The EQ alone has ten bands you can adjust ±12 dB. By reading through the descriptions and making a few tests, you can pick a setting and make adjustments pretty easily. Our only issue was with the settings for Internal and External Mic Recording. They offer a "Transformation" setting of 1 to 5, but the manual has no description of what the Transformation is. This is definitely a place where your ears are on their own. Test these settings thoroughly before using them, or follow the golden rule: record clean, process later.
The R1 surprised us with a few technical limitations that cropped up in testing. First, there's no stereo meter, and no calibration for the meter's 15 steps on the face of the unit. There's also no internal clock, so all your files have the same creation date: January 1, 2002. You can't pre-name a file and then have the recorder add subsequent numbers. The contrast setting, while variable, offers a very narrow angle to view the display. The unit submitted for testing also included no wind screens.
In using the recorder, we found it easy to get going and fairly simple to operate. The R1 has a basic "press Record to record" simplicity that lets you avoid the extra effects and such in their entirety if you so choose. This is good, as we found messing with the effects settings tricky at times because they reset themselves if you jump to a different effect and change a setting there. This also means you can't use multiple effects at the same time.
Some effects were astounding. The built-in microphones and the Hall reverb setting produced terrific audio—with a lot of reverb—with the unit just sitting on the table. The R1 also includes a few zany effects which we used effectively on one project. Contrast this to the Mastering effect, which proved to be too hissy no matter how we adjusted it.
Testing also revealed that the R1 requires more gain than other recorders. Even with the input level all the way up, we often found our audio levels to be average. Though it has an internal limiter, we did not find ourselves having to turn down the input levels because there was too much gain. Contrast this to a DAT machine, which never required us to turn up the same stereo microphone past halfway.
The ways the R1 display changes during recording are simple and clear. The top line always shows the name of the file. The second line can be toggled between elapsed time, remaining time, sampling settings, and the recording level. When you hit Pause, the second line always jumps to Rec-Pause and the elapsed time. This is handy.
The R1 interface seems to be a little sluggish sometimes, responding a beat or two behind button presses. This is especially evident when trying to jump around the tracks during playback. We also noted that the unit fades up the audio when it first kicks into record. The audio effects also ease their way in, though there is no delay in the audio heard. This approach can be quite good compared to all of a sudden being hit with some ear-splitting audio.
All in all, if used as a clean, basic audio recorder with an external microphone, the R1 will provide you with audio clips that deliver. We don't recommend it for the built-in effects or for the kind of work where the audio you want to record tends to be very quiet.
The Marantz PMD660 is a little thicker, wider, and longer than the R1. It uses this additional size to provide true balanced XLR connectors and space for four internal AA batteries. The XLRs are for microphone-level input only and can provide phantom power. Line-level I/O is only available through a 1/8" stereo jack. There is a headphone jack on the front and a full-size USB jack for data.
The compact flash card fits into a door on the front of the unit. The door is a bit tricky but is designed so that it can be screwed shut over the media. The eight segment meters are right above this door.
The meters tend to be a little rough, but the advantage is that they are stereo, while the R1 offers only mono metering. They are also LED and are easily visible from any angle. With green, yellow, and red for "over," you can even tell how the recording is going at a glance and from across the room.
The 660 simplifies recording options compared to the other recorders tested here, as well as the rest of the Marantz product line. You can record audio as MP3-44.1KHz, MP3-44.8KHz, PCM-44.1KHz, or PCM-48KHz. That's it. Other options include automatic level control (useful for reporting), a limiter, and an internal attenuator. The 660 also features two built-in microphones and adds a single internal speaker. It would be fairly impossible to critically assess the audio on the 1" speaker in the 660. As with the R1, there are also no wind screens for the 660's mics.
The LCD is clear and uses an even backlight, which makes it easy to read from almost any angle. Most every setting is visible all the time on the LCD display. The buttons on top allow easy navigation of the menus.
One convenient feature of the 660 is that there are user-definable presets. You can set one for ALC, mono, as MP3, for taking dictation of ideas. You can set a second preset, in stereo, no ALC but with a limiter, as WAV, for recording production audio. Then you can quickly toggle between Preset-1 and Preset-2.
In testing the 660, we found it very easy to get around. The ability to see and adjust the recording levels of different channels proved to be very handy. In one case, one performer was much louder than the other. We had placed a stereo microphone between them, and easily turned down the one channel.
It's worth noting that the Marantz mic inputs are more sensitive than the R1's, and we suffered some audio distortion from audio clipping at the inputs. We didn't have this problem with the same mic on the R1. We used the internal 20 dB pad, but when we cranked up the dial to compensate for the padded audio, we heard a bit of hiss. Apparently, the 660's limiter—which was turned on—doesn't prevent a hot mic from overloading the mic inputs.
Unlike the R1, which goes into Rec-Pause when you hit the Record button, the 660 goes directly into recording. The Marantz unit is also very responsive to adjustment of settings and navigating the buttons on the unit. Another very useful feature of the 660 is the ability to cut a track on the fly. When you hit the Record button while recording, the track number will change. The other two units required stopping and then restarting. This is probably not an issue for most recording situations, but for situations where the recording must not stop, but you still want separate tracks, the Marantz does it with ease.
The 660 also lets you set the date and time, and store a name. They say it is the name of the unit, but this eight-character name is what ends up as your file name. Leave a few digits at the end and the 660 will automatically increment the name of the file for you. This way, your files not only have the creation time of the event, but you can give them some name to help identify them later. Very professional.
M-Audio Microtrack 24/96
The M-Audio Microtrack is an astoundingly small and light recorder. You turn the thing over in your hand and find yourself looking at what appears to be a slightly thicker, rounded iPod Mini, but is actually a full-featured audio recorder. From the 1/8" stereo mic to the balanced TRS 1/4" L&R mic/line inputs, all the way to separate RCA lines-out and even coax digital-out, this little unit makes use of every inch of available space.
The right side of the unit is the simplest, with just the compact flash slot and a little rocker switch you use to navigate the menus. This is also the only unit to support MicroDrives, which is good for versatility, though they will use more power than compact flash cards and shorten total record time when powered by a battery. The bottom, in addition to the line- and digital-out, also has USB I/O, which serves to recharge the unit. Thankfully, this doesn't require a computer because M-Audio includes a minuscule AC to USB adapter and cable that you can treat like any other AC adapter to recharge or run the unit in the field.
The top of the unit, in addition to the mic/line inputs, also has a headphone output. The left side has several adjustable switches that are flush with the unit. A Menu button brings up the menu. The face of the unit features a Power button that must be held for several seconds to power up or shut down the unit. A decent, backlit LCD display provides most information. There are audio-level adjusters for left and right channels, a headphone adjuster, a Record button, and a Delete button.
In the menus, you can set the input-level adjustments to be linked or separate. Linking them means that using either the left or right adjuster will affect both channels. You can unlink them in the menu, and then the left or right adjuster adjusts only that channel. You can also turn linking on and off while recording. It's convenient and capable and offers operation similar to the Marantz with the two record-level adjusters linked by friction, but easily adjustable separately.
Watching the stereo meters while recording, however, will take some getting used to. Even when adjusting the recording levels all the way down to zero, we were still getting a good recording. M-Audio explained that audio adjusters work like input trim pots; they will not fade the audio completely to silence. Another feature is that M-Audio already has firmware updates available to update the capabilities of the recorder.
In addition to the on-screen display of audio levels, the unit also features green and red LEDs below and above the recording-level adjusters. The green LED lights up when a signal is present, while the red indicates peaking. You can easily adjust the record levels down from peaking without even having to take a critical look at the LCD display.
The Microtrack lacks built-in microphones but M-Audio includes a tiny "T" microphone system with two small elements directly attached to a 1/8" stereo plug. There are little fuzzy wind screens over the mics to provide some protection from breezes.
When using the navigator button/rocker to get around the menus we found it a bit touchy, sometimes jumping two items instead of one. Reading online forums unearthed a few posts noting the same issue. Careful, deliberate use is the best course of action here.
Lastly, space on the M-Audio is at a premium so some things have multiple functions. It is not intuitive to use the menu navigator button/rocker to pause recording or playback, but that's what it does. Also, there's no dedicated button to stop a recording; you just hit the record button a second time. While it may start to seem fairly intuitive after some time, things like this can befuddle a new user.
Knowing that most field production takes place away from convenient AC lines, and that prosumer camcorder batteries easily last on order of four to six or more hours, we tested the full-on recording times of the units on battery power, recording to MP3 with the backlighting forced on all the time.
Our testing used the built-in and included microphones with each unit. You will shorten your record times by using external microphones that require phantom power or by using MicroDrives, which consume more power than compact flash cards. Setting the screen backlighting to stay on for only a few seconds when a button is pressed will increase run time, but we found it much harder to monitor the status of the recorders without the backlight—especially the R1 and the Microtrack, which use the LCD for metering. We've also been told that different recording settings (like WAV versus MP3) can affect record time, so we had all three units record MP3.
We used the same 2200 mAh NiMH rechargeable batteries in both the R1 and the 660. We charged up all the batteries immediately before testing, and topped off the Microtrack's internal battery via a computer USB connection immediately before testing.
The R1 ran 3 hours and 20 minutes on two AA batteries before the low-battery indicator started appearing in the display. It ran for about ten more minutes before it shut down completely. Interestingly, because it has a "hard" power switch, the R1 comes right back on, but does not resume recording. The first time we saw this it caused a bit of confusion about why it stopped recording on its own. Now we understand where to look for the low-battery warning and know to check the recorder every five minutes or so. A separate, bright LED would be a far better low-battery indication because it would grab the attention it deserves. A battery graph that stays on screen the entire time, like the other units have, would also be better.
The Marantz 660 ran about 3 hours and 20 minutes on four AA batteries. Despite the fact that we had set the machine to know we were using NiMH batteries, we saw no difference in the battery-level indication. In fact, before the battery icon changed at all, the 660 shut itself down. Despite having the hard power switch still turned on, the 660 stayed off. We turned the switch off, waited a bit, and then back on. The unit stayed on for over a minute, with the battery icon displaying a full charge, before the unit turned itself off again.
The only other internal battery setting on the Marantz is for alkaline. Given how capacities of NiMH batteries have increased from 1400 to 2500 mAh in the few years I've been using them, I can understand how the calibration of the recorder may not take into account higher-capacity batteries, which may reach the end of their life before the recorder notices any change in power supplied. If there is no way to calibrate the two or adjust the setting in the recorder for newer capacities, a simple voltage reading would be more useful and would let the user determine when to change the batteries.
The M-Audio Microtrack has an internal lithium ion battery that ran for more than four hours. This was quite impressive considering the how small and lightweight the recorder is. When considering the backlit display and various LEDs on the unit, it is clear that M-Audio did their homework on power usage and optimization.
Moreover, the battery indicator is clearly calibrated to the internal battery. The indicator works its way across the battery icon diagonally and has more than ten steps. The Microtrack only died when the indicator was on the last few dots of power. Even though four hours of run time is comforting, we can't help but wish there were a way to change out the battery in the field—perhaps like a cell phone—in case USB power were not available.
So in terms of run time, the Microtrack wins hands down. The R1 is a close second with good warning before a battery change and it only uses two AAs that can easily be replaced in the field. The 660 worried us with no warning before shutdown and because it required double the number of AAs to get nearly the same run time as the R1. This is not really a big issue in terms of cost, but we would have expected a little more run time than the R1 and more accurate power metering.
With all three units tested here, the included or built-in microphones are not high-end elements. Also, the built-in/ attached microphones are very susceptible to noise from handling the recorders. Even resting on a table or floor they pick up surface noise very easily. Using an external microphone separated from the recorder by a cable is essential. It is the only way you can touch and adjust settings on the recorders without recording handling noise.
Each of these units also has certain strengths. They all produced very good audio. While there are certainly more capable, and more expensive, portable recorders, these three seem to be vying for the same place in our camera bag.
The Marantz offers ease of use, good run time, fast response, and balanced XLRs. When the mic preamps get overloaded, the 660 has an internal pad you can add to salvage your audio. It restricts flexibility in recording settings—probably to encourage producers with bigger budgets or higher-end requirements to step up to Marantz' bigger flash recorders. The 660 takes more power to give the same or less run time and it's also considerably bigger than the other two recorders tested.
The R1 offers a small form factor, many recording settings, as well as numerous features and capabilities including optical out. It gets a decent record time on just two AA batteries. It lacks balanced inputs and input pads, and you can't monitor or adjust individual channels. The display is the weakest of the bunch.
The Microtrack is a truly diminutive recorder with many recording settings, coax digital out, and RCA line out. It offers amazing run time from an internal battery. It lacks the ability to change batteries; in the field this can really become an issue. It also lacks an internal limiter or pad so if something is too loud, it's distorted and there's nothing you can do about it. Lastly, it seems that M-Audio still has a few tweaks left to implement in their firmware to make the record-level adjustment work properly.
Each recorder takes several seconds to start up and get ready—no different from most any camcorder. The Marantz is the only one that can cut a track on the fly. The others have to stop recording, save, and then can start recording again. For some people, this will be the key factor in their purchase. The R1 is second fastest, and the M-Audio slowest.
When it comes to which one worked best for me, personally, I have to go with the R1. While not the snazziest or fastest, it works. It functions simply, and it offers great flexibility and run time with batteries I can change in the field. I can deal with the file date and naming issues. Now, add to this that Roland has just announced the R-09, a smaller, lighter and simplified version of the R1. Unfortunately, they also removed the limiter, which is critical for the work I do. But if you liked the R1, and don't need a limiter, this new recorder may make your choice even easier.