FAQ: Recording & Editing
- How do I get my sound-recordings from my recorder into my computer?
- Can’t I just copy my minidisc data files?
- Why can’t I output digital audio from my consumer minidisc?
- Can I record onto my iPod or other portable digital audio player?
- Can I record Interviews Direct to CDR
- What cable(s) do I need to connect my minidisc to my computer?
- What sound-recording software should I use?
- Can I use my laptop as my audio workstation?
- Can I use my laptop as a field recorder?
- How should I archive my audio?
- Should I record at 16 or 24 bit?
- Why isn’t, you know, everything working properly?
How do I get my sound-recordings from my recorder into my computer?
Plug the line-out (analog) of your recorder into the analog line-in of your computer (most computers have built-in microphone inputs, some need an added soundcard or external interface). Go to your computer’s sound control panel and make sure that audio input is set to the right device (probably “external mic.”) Launch your computer’s sound-recording software, make a track, click record on your sound software, then play on your source machine (your minidisc or DAT or whatever.) Some computer’s sound inputs set to recieve mic level signals and will distort if you feed the louder line-out or headphone-out signals from your player. Better to connect to a or line-level input if possible. In either case, adjust your playback levels and watch the input meters on your record software, sometimes the connection will work fine even if it’s not electronically ideally-matched, . It’s best to feed as loud of a level as possible before distorting, rather than fixing it up later. To that end, you may want to use a mixer to regulate the levels, or in a pinch, use the headphone-out of your playback device, adjusting the volume to show strong levels on the input meters of your computer’s recording program. That’s one way, but there’s a better way…
If your computer has a USB port, use an external audio interface to capture audio (US$35 to US$600). You may have no choice if you have one of a few models of Macs, which were made with USB ports but no Audio-in. Using a good external interface can save one or two additional analog/digital conversions of the sound, and so, will sound better, especially if you can keep your signal digital. This external inrface can be simple, with only input and output connectors, or mgith have several connectors, volume knobs, meters, and perhaps digital in and outputs. This allows a real-time transfer of your sound, just like playing it into the computer’s audio input. It’s not a data copy, but it will get audio in and out of your computer without installing a card, and usually with better quality than an off-the-shelf computer sound input. There are more and more every day, but the most promising are listed here:
Firewire, aka IEEE1394, is the most common method of capturing video into a computer. Firewire transfer rates are many times faster than standard USB, although USB2 is fairly equivalent with the original Firewire rates. Recently, some computers have adopted a faster firewire protocol called “Firewire 800.” Firewire interfaces allow simultaneous multi-channel audio recording and playback., for moving more channels of audio simultaneously than USB allows. They usually offer more input and output choices, and usually include at least two built-in mic preamps, making them a versatile tool for a small studio as well as for remote recording. This may be more power than you need for basic audio production, USB does fairly well with 2 tracks of 16-bit audio, but Firewire offers a more robust and reliable connection. The Firewire interfaces we know of are listed here
An installed PCI card can give greater bandwidth than a USB device, allowing more simultaneous channels of input and output, and there are many to choose from, from basic $200 models to high-resolution multichannel systems for tens of thousands of dollars. The popular Digidesign Audiomedia card has been discontinued, replaced by the M-Box USB interface but there are many others with similar capabilities from M-Audio, Echo, Aardvark, Digigram, Digital Audio Labs, Hammerfall and many others. There are many software/interface systems, fom the project-studio oriented MOTU 2408 to Steinberg’s Wavelab and Nuendo, to pro-studio, high-overhead systems that allow simulation of large multitrack tape machines and mixers. But one should try to keep an eye on how much horsepower is actually required to create your radio program, and perhaps spend the extra bucks on a better microphone. The ease of installation and portability of USB and Firewire interfaces have made PCI card systems less attractive except for in the most demading multi-channel environments.
Do I really need a separate hard drive for audio?
While it is sometimes possible to record audio to the same physical drive as the computer’s system and recording application, you’ll get much better performance by recording to a separate drive. Audio eats up lots of space, but we’ve finally reached a time when it’s rarely a crisis anymore. If you get an 80 gig hard drive you will have some trouble filling that puppy up for a while. 160 gig, 250 gig, and even larger drives are easily found, and not all that expensive, If you have a tower computer, you could install an internal EIDE drive for really cheap. If you have an iMac or a laptop, or just want portability, you can get a firewire or USB 2 drive.If you want to record to it or play back sessions directly, you will need to get drives that meet certain specs. Check with your editing software for specific requirements, but for instance, ProTools is very picky about specs on Firewire drives. (Check the digidesign site for specifics) The basic rule is that if the drive has a rotational speed of 7200 rpm, fast data access and if it’s a firewire drive, an Oxford 911 or 922 bridge chip.
How big a drive do I need?
Uncompressed 16-bit audio takes 10 megs per stereo minute, so that’s 600 megs per stereo hour. How much room you’ll use depends on how you work: do you dump-in every minute that you recorded and edit from there, or do you only put sections you need into the computer? Do you do lots of bouncing or other processes that write new files? Are you using a destructive editor and want to save many different versions of the soundfiles? Non-destructive editing generally writes small session files that refer to one version of the soundfiles, so many versions don’t take up too much more room. Do you use 24 bit files? Do your projects have many tracks?
In any case, 80 gigs is a lot of space: over 12 hours of 16-bit stereo. And drives are getting cheaper and cheaper, a big drive will make your life easier.
The Sony HiMD minidisc recorders, released in mid-2004, promise to allow file transfers over USB from the minidisc to the computer, but we haven’t had a chance to try this yet to see how it works. It’s unclear how easy this will be, and seems likely that it can be done only via Sony’s own Windows-only software. Most minidisc recorders do NOT allow movement of data from the minidisc to the computer vis USB. Even the HHB Portadisc, which will transfer sound to a computer via USB, is transfering digital audio, in real time, not files. Be aware that the small consumer portable minidisc recorders that advertise a USB connection are designed to load digital audio onto the minidisc from the computer via USB, and not the other way around.
Yeah, where IS the digital output on those minidisc recorders? Many of the little consumer minidisc recorders come with a digital cable, but as you discovered, there actually is no digital output, only an input.
The answer is easy: these were made as consumer machines, and in order to keep size and price down, they built these devices with only the features the average user would want. They imagined these being used as a portable listening device, and that’s about it. They supplied a digital INPUT so you can load CD tracks onto minidiscs, and some models even have a USB connection, but that similarly is configured to load music from you computer ON to the minidisc, not the other way around.
In mid 2004, Sony introduced new discs and recorders that use a format they call “HiMD” that has larger capacity, therefore longer record times, as well as the ability to record in an uncompressed format, rather than the lossy ATRAC format that all other minidiscs use. But the even more intriguing potential is that these recorders allow fast transfer of your audio data from the recorder to a computer over USB. Unfortunately it’s not as simple as one would hope, the process is mediated by proprietary Sony software (Windows only at this time) that users seem to find needlessly complicated, and there are some irritating roadblocks set-up supposedly to prevent pirating copyrighted material, but that can get in the way of our using the device as a field recorder. For instance, audio recorded through the optical digital inputs can’t be output as a standard .wav file for later editing. The technology is still young, and perhaps it will be smoothed out in the future, but at this time (mid 2004) it’s not ideal.
The only minidisc recorder that has both digital out and USB out is the (discontinued) $1500 HHB portadisc. This is an excellent machine, but that’s quite an investment, if you can find one.
So, you are stuck outputting your sound through the analog outs. But you are still ahead of the game compared to a fully-analog machine, like a cassette, the minidisc audio is still cleaner and clearer because it was recorded digitally.
Even if you find a recorder with a digital output, there are very few computers that have built-in digital audio inputs, the Mac G5 towers are the only ones that come to mind, but there are many external USB or firewire interfaces that can accept digital audio.
In theory, this should work pretty well, a good-quality MP3 is high enough quality for many purposes. But there are a couple of potential problems.
The first is that the recording features of the small consumer MP3 players that offer this option are not ideal, they rarely have good analog-to digital converters, or good level control, or good metering.
Also, almost none of them have microphone inputs, so you’ll need some sort of preamp to get a microphone signal up to line-level. Or if there is a mic input, the sound quality of the built-in preamp is rarely very good.
I love my iPod, but the adapters that allow using built-in or external microphones leave a lot to be desired. It’s a consequence of the iPod’s 8khz sample rate, which allows the recording of intelligible audio that’s fine for voice memos or dictation, but not even as high quality as an old analog cassette.
The other bigger problem is the potential for “transcoding” degeneration, a rather nasty audio distortion that can occur if a file is coded multiple times via “lossy” methods such as MP3. It’s usually not a big deal if a file is converted to MP3, or MP2 once or twice, but there is a line at which the file can break-up, drop-out, or exhibit weird metallic-sounding artifacts if it’s coded too many times. So it’s a little scary to have the original field recording as an MP3, which you might want to convert to different file for editing or integration into a larger production, which might get compressed again for delivery or output, and could get squeezed more times down the line over the course of uplinking or storage, if that applies….
More and more folks are using field recorders that record in MP2, (which usually applies less severe compression than MP3, but still is a lossy scheme,) and not reporting problems. And even minidiscs use a form of audio compression, and could be subject to the same issues in some circumstances.
I have personally experienced a version of this problem when coding a radio show to high quality (44.1 khz, 256kbps joint-stereo) MP2 for delivery to a satellite uplink over the internet. The uplink needed to convert it back to .wav for transport, but then the satellite modulators re-coded it to MP2. A station took that signal down and recorded it in to their hard-drive storage, again as an MP2. On the air, it sounded terrible! The first compression sounded fine, I’m not sure about the second, but the third one created some very obvious artifacts, wavering metallic halos around sustained tones.
It’s tempting to take advantage of the increased record times offered by using a compressed format, but unless you are confident that your audio will not be compressed more at a later point in its life, doing so risks creating unpleasant-sounding audio.
There are a couple of portable real-time CDR recorders, and some folks are using them to record interviews in the field. Here are the down-sides: you really can’t move around, these units are portable, but they must be set on a horizontal surface and kept still while recording. That’s not a problem for most interviews, but you can’t follow someone as they move as you could with most field recorders. Also, you can only get 70-80 minutes max, and when you reach that point, you’ll need to “finalize” the disc, which might take a few minutes, so it’s not as simple as flipping a cassette or even slapping in a new tape or minidisc (although minidiscs need to write a Table of Contents too.) so if your interview goes over the length of one CDR, it will be a bit inconvenient.
And I’ve read a few complaints about the quality of the mic preamps, that they’re a bit noisy, but many portable recorders suffer from this until you spend big money.
The benefits are that the CD blanks are cheap and easy to find, and the recordings are easy to play back, copy, and transfer to a computer for editing.
So it’s a mixed bag, some good some bad….
This depends on your specific recorder and computer interface. Many computer soundcards and built-in sound inputs use the same type of connector as the consumer minidisc recorders: a 1/8″ stereo “mini” connector. In this circumstance you need a cable with stereo mini connectors on each end (there are three conductors: the tip, ring and sleeve.) If your minidisc player has a “line-out” use that, if not, use the “headphone out”. The line-out will be at a fixed level, while the headphone out will have a variable output level, controllable from the front panel. This can sometimes be an advantage, allowing you to optimize the input to your computer. If your computer’s sound input will only accept microphone levels, you will get excessive noise and distortion when you connect a line-level input. There are not a lot of devices built to knock-down signals to mic level, but an attenuating cable can help, or a headphone-extension cable with a volume control, or most of the the Mackie VLZ-series mixers can output microphone level from their main XLR outputs with the push of a button. But rather than dropping your levels down, only to amp them back up again inside an electromagnetically chaotic computer, or using a noisy headphone output, try using one of the line-level interfaces described above instead.
If your soundcard has different connectors, or if you are using a mixer in-between the minidisc and the computer to regulate your levels, you will need a cable with a stereo mini on one end, that splits to two of whatever connector matches your soundcard or mixer, probably RCA plugs or 1/4″ plugs. Make sure you are not getting a mono splitter or “Y-cable” that sends the same signal to both outputs, you want a cable that sends left and right signals to the seperate RCA or 1/4″ outputs. These cables are commonly used to connect a “walkman” style CD player to a home stereo.
If you are lucky enough to have a pro deck with a digital output, invest in an interface that accepts digital signals and use that.
Your personal computer can also become a multitrack audio recorder/editor/effects-processor/mixer. All you need is the right application. The programs most commonly used by independent radio producers are Digidesign’s ProTools (usually the Free, or LE versions) and Syntrillium’s CoolEdit, now called Adobe Audition. There’s also a free multi-platform application called Audacity that is pretty powerful. Descriptions of these and other popular programs are available here: Which program to use depends on your needs (are you just editing dialog to be mixed elsewhere, or producing a complex, layered piece?) and the type of computer you use.
Check the editing program’s minimum requirements, but most any new laptop will run a basic audio editing program. You may have a problem running lots and lots of simultaneous tracks with lots of real-time plug-ins, but basic stereo editing will be fine, and you should be able to get multiple tracks with a few plug-in effects even on a basic machine.
Laptops don’t tend to have really fast drives, or the fastest processors, so they won’t be quite as capable as some desktop systems, but all computers have sped up so much in the last few years that laptops still usually have ENOUGH power to do the job. As with all systems, RAM is key, put in as much as possible, I wouldn’t try running an audio program under Windows XP or Mac OSX ojn less than 512 megs. Many of the manufacturers use 256megs as a minimum.
The built-in audio inputs and outputs are OK for basic work, but an external interface will give you more control and better sound quality. USB(1) is fine for stereo, you’ll need USB2 or Firewire for multitrack. Remember, we’re just talking about simultaneous inputs or outputs, one can internally mix many tracks to a stereo output using USB. There are many external audio interfaces that provide good mic preamps, digital inputs, metering and headphone monitoring, and it’s worth that extra expense and bulk to improve the quality and control of your audio input.
It’s always good to have a second drive to record audio onto, and on a laptop it has to be external. Data transfer needs to be fast, so you need a speedy external drive: a firewire or USB2 connection, a 7200rpm drive, and the Oxford 911 or 922 bridge is preferred.
In theory, there’s nothing wrong with using a laptop as a field recorder, but in practice it’s not as convenient as it would seem. Laptops take up a lot of space, the screens can be hard to read in bright light, the battery life is not as good as a typical audio recorder, and being computers, they are more susceptible to crashing or freezing-up. It can work, but it’s way more convenient to use a smaller device devoted to audio recording.
In a perfect world, archivists tend to recommend making copies to multiple media: make an analog 15 ips reel-to-reel tape, AND make an Audio CDR AND make a .wav (or .aiff) data CDR. But to be realistic, analog copies are slow to make, expensive and cumbersome to store, and many of us have experienced the fragility of analog tapes.
There’s some doubt about how long CDRs or other dye-based optical media like DVDR will last. It’s certainly affected by exposure to light and other environmental conditions, but even in ideal storage there’s some reason to believe that the dyes might fade, and the data may become hard to recover.
A reasonable approach to this is to make multiple copies, then date and file them in such a way that one can systematically copy them to a new form every few years, 5 years seems to be the current paranoid consensus. If one copies the CDR to a new CDR, or whatever storage format will be the most practical in 5 years, one might be able to avoid the problem of the original copy having faded.
As for the file format on the CDR: data CDs have more error-checking and less interpolation than CD-Audio. So burn a data back-up of your .wav files, or .aiff files, or .SD2 files or whatever format you are using, rather than making audio CDs. These are file-format issues, not a question of the kind of CDR.
There’s no physical difference between “data” and “music” CD blanks, except for possible variations in dye formulations. CDR’s labeled “music” are formulated for single-speed burning, and so may be better matched for stand-alone burners or even computer-based burners that operate at lower speeds.
The process of burning an audio CD is kind of weird, and often introduces errors that would not be created if copying data. The over simplified explanation is that audio CDs just keep going and if they missed some data, there’s be an error. In most cases error-correction on the CD-players will cover it up, but that’s not reassuring when archiving important audio. Copying data allows some pausing and a “wait, I didn’t get that, send that again” kind of handshaking, that tends to give more accurate copies.
CDRs are so inexpensive these days and take up so much less space than analog tape, so it’s practical to make audio CDs that can be played in any CD player, then also burning a CD with that audio as a .wav file or .aiff file. Many people are burning larger projects to a DVDR disc, as data.
Drives are getting cheap enough that it’s almost practical to get a few gigantic hard drives and save the .wav files and .aiffs there also. Drives can die, CDs can be damaged, so redundancy is key. And one has to plan to make copies in a systematic way to preclude the possibility of media degradation.
I have some 15-year-old DATs that won’t play back properly (although others will.) I have CDRs that did play a few years ago and no longer do.
But one has to be careful about media, we still have a few stacks of 5-1/2 floppies that we keep wondering if there’s anything important on there… a missing transcript maybe? But none of our computers have those drives any more. None of my Macs have floppies at all, and this could happen with CDRs in a few years, we might scoff at the idea of putting data on such a fragile and puny medium.
The advantage of analog tape is that we’re still playing back tapes from over 40 years ago, it’s not always pretty, but we can usually get information off of them…. we don’t know if that will be true for CDR or even IDE or SCSI hard drives, they could end up as some weird format that nobody can read. I expect ADAT tapes will get like that pretty soon, and we’ll eventually be searching for someone with an operational ADAT in the attic somewhere.
Here’s a good site with info about the care and feeding of CDR and DVDR: http://www.itl.nist.gov/div895/carefordisc/
There are endless arguments about what dye formulation, what burn speed, etc gives the best result, and the answer is usually to find a good match of CD brand, dye color, burner, and burn speed. Data can probably be reliably burned at higher speeds, it doesn’t have the same linear data stream issues as audio does.
There are pricey CDRs that claim to be “archival quality,” tested to last 100 years or something, but will they REALLY be significantly better? Who knows? I would try to get at least good quality CDRs, not the cheapest store-brand bulk discs.
24 bit really does sound better than 16-bit, but there are a couple of down-sides. The soundfiles are significantly larger, which even with today’s big drives, always ends up making a crunch for space, and can often make it more difficult to archive, spilling to multiple CDRs or whatever. And even the bounce-to-disc is now harder for the computer, and depending on the number of tracks and kinds of plug-ins, you could choke the processor, even a nice new speedy one.
And the thing is, most of your sources are still 16 bit at best, the interview you did on DAT, or the music beds. Loading those into a 24-bit session is now going to convert all of those files to bigger ones, but to no real gain in fidelity, they’ll be 16 bits with 8 bits of zeros at the end. Additionally, you’ll usually want your production at 16-bit eventually to put it on CD, so you’ll need to dither your final stereo output, which is no big deal, but a good dither can take some processing cycles, which could end up pushing you past your chip’s abilities if you’re using a few power-hungry plugs, like the L1, etc. (Of course there are work-arounds, for instance, bounce the session as undithered 24-bit, save that file as a high res archive, then make a new dithered file from that bounced file, which needs no other plug-ins…)
And if the end result will be radio or internet, you won’t even hear the difference.
Now, after talking you out of it, I do sometimes work at 24-bit even with 16-bit sources if the processing will be a significant element. Fades, reverbs, and other elements that rely on fine detail can really benefit from the better resolution down near the noise floor, with smoother, cleaner, more realistic fades, or reverb tails or whatever. So I do music stuff at 24-bit, even if I’m working with live stereo recordings from DAT, or multitracks from ADAT, just so any reverb, other processing, or fade-outs sound better.
Similarly, if your productions rely on a fair amount of Audiosuite processing: selecting a range of audio and applying an effect to that region in non-real-time (an old-school approach that I’m ironically finding myself using more and more these days), rather than just applying RTAS plug-ins on a track, having the project at 24 bit can reduce the negative impact of multiple processing. 16-bit sound can get pretty gritty-sounding if one does a gain-adjust, then a limiting pass, then an EQ pass, then bounces it out… It’s not so bad at 24-bit.
It’s always hard to predict the future uses of things, most pundits thought we would be working with better and better quality audio, not more and more lossy, compressed, crappy sources. But there may be a market in the future for high-res audio, so there could be some point to keeping everything at 24-bit, but it’s not all that practical. There’s a big push on for surround-sound stuff, and it CAN be really cool, but I’m still not sure that it will ultimately be a common, popular format. Same for hi-res.
If you get a field recorder that records at 24-bit (like the Sounddevices HD recorder, or even that Core-Sound thing that runs on an iPaq PDA) you might want to start doing everything at 24-bit, but if it’s only your narration tracks that actually go in at 24, I’m not sure you’ll hear an improvement, and you will take a performance hit.
So in short, I wouldn’t bother, I’d stay at 16bit, unless there’s something really delicate or super hi-fi that you want to try to preserve.
There are a myriad of reasons why things might not be working in concert with one another, before you buy something, check the company’s website and make sure that it is appropriate for what you want to do. Editing programs will give indications of minimum computer requirements, interfaces will usually indicate which programs they will connect with, etc. There is always a risk of a conflict with some hardware or software in your particular set-up, but here are some general rules that will reduce these problems:
1) Buy as much RAM as you can afford, and your computer can hold — RAM is cheap and makes the most cost-effective improvement in sound-processing. (It also can determine how many tracks and effects your software can handle.) It is NOT overkill to load up your machine with 500 megs or more.
2) Set your sound software’s memory allocation (the amount of RAM the program uses) to high: 100M or more. If you’re in Mac OS9, try to run as lean a system as possible, only running needed programs and system extensions, leaving more processing power open for your audio work. Newer operating systems handle this automatically, but even in that instance, try to restart your computer before doing heavy audio work, and run as few other programs simultaneously as possible.
3) The faster the clock speed (in MHz: Mega Hertz) of your computer’s CPU (Central Processing Unit), the faster it will process sound. That translates to more tracks and more plug-in processing.
4) The larger your computer’s harddrive (in M: Megabytes), the more sound it can store. You’re always better off recording audio onto a second drive, different from the one where your computer’s system and the recording applications reside. If you have a tower computer, it’s easy to install an additional drive. If you have a laptop or iMac, or just want the flexibility of portability, use an external firewire drive. Make sure it has a rotational speed of at least 7200 rpm, low data access times, and uses the Oxford 911 or 922 bridge chips. External USB drives are rarely fast enough to use reliably. USB2 and firewire 800 are faster than the original specs and might be good choices. Check with your audio software’s web site for specific compatibility.
5) Use these minimum sound sampling rates for radio production: 44.1KHz, 16bit, uncompressed (the rate used for Audio CDs). It’s true that FM broadcast does not take advantage of the full 20 khz bandwidth of this resolution, but recording and editing at lower bit-depths or sample rates will often result in murky, noisy sound. Better to get some additional hard-drive space if you’re running low. Recording at higher sampling rates and bit depths will give higher-quality sound, but will also require more disc space for the same length of audio, and tax your computer’s processor more. if the end -use of your production is FM broadcast or internet streaming, this may be overkill.