Tech Notes for Making Waves Audio Producers:

RECORDING - FORUMS - INTERVIEW - DIGITIZING - EDITING - INTEGRATION - PRE-PROCESS - POST-PROCESS - BACK

The following instructions are intended to outline the simplest procedures for doing every step of the production process. There are of course always other ways of doing things and lots more to learn if you feel up for it. Here is a tutorial on the CONCEPTS behind audio production if you want to understand all the terms involved and what's really going on.

These instructions below assume that you are using a windows PC and not a Mac, although many of the steps should work on both.

You may need the following free software: Audacity (sound editor), CDEX(converter / CD extractor), WinAmp (player), FileZilla (FTP server / client) and UltraVNC (remote terminal software). Check occasionally for updates.

RECORDING

Record with the end product in mind. Before starting, spend a few minutes with Google familiarizing yourself with the topic and person being recorded, composing intros, insightful questions, and a one-sentence re-intro for longer pieces. Record everything likely to be needed including your transitions on location, so you have consistant background noise and don't have to dig out the mic again during editing.

Find a relatively quiet, not-windy location for recording. Foam mic covers must be used in wind, and you can put cheaper stand-alone recorders in a sock. It's a good idea to wear headphones while recording so you know what noises might cause problems. Any movement of your hands or other contact with the mic, wires, recorder, mic stand, the table it's on, etc will be audible in the recording. If you can both speak up and there is no background noise, just set up the mic or recorder on a table between you and who you are interviewing and leave it alone.

Otherwise, use a consistant hold on the mic and keep it 45 degrees to the side of the person, pointing the mic at them a few inches from their mouth. You DON'T want the wind from their speaking blowing directly at the mic, or the recording will have very loud and distorted "P" sounds. If your recorder supports it, set the mic sensitivity appropriately, depending on how loud the person talking is and how close you can get the mic to them. The volume adjuster often also changes how loud the recording will be. With extremely loud speakers, the mic should be a foot or more from their mouth.

Hopefully, all recorders that you are likely to borrow will have simplified instructions included with them.

RECORDING FORUMS

It's a good practice to get featured speakers to sign a simple waiver, simply stating that they authorize their talk to be recorded, edited, and re-distributed by Making Waves. However legally, any public speech, words said on a sidewalk, or other audio where speakers have no "reasonable expectation of privacy" is probably fair game without a waiver. Recording from the front row is OK, but get the mic within a couple of feet of the speaker's mouth is best, such as on the podium (briefly and politely coaching the speaker to avoid bumping the mic and podium is also wise). A close second is to wire directly into a "line-out" from the event's sound system or mixing board, if present. Most recorders have a "line-in," which is nearly always a "mini" plug (1/8th inch stereo). Connecting to the mixing board may be "mini," 1/4 inch, RCA, or XLR. So, a mini to mini wire, 1/4 inch stereo adapter, an RCA to mini wire, and a 1/4 inch to XLR adapter should cover all the likely possibilities.

Third best is to set the mic sensitivity to low and point it at the public address system, a few feet from it. Question and Answer periods may be useful but are technically quite difficult to record if there aren't mics set up specifically for the audience. More importantly, try to get a quick interview from the speaker afterwards and ask for reactions and comments from the audience as individual interviews.

INTERVIEW TECHNIQUE

Remind the person you're speaking with that the interview will be edited, so they can relax and take their time. Tell them that silent pauses are easier to edit out than filler sounds (uh, um, ya know, etc - we respect our audience too much to make them listen to a lot of meaningless words). It's OK to ask the person you're about to interview if there are specific ideas or issues they would like to bring up in the interview, so you can work that in gracefully if it seems relevant. Tell them to try to ignore the recorder and have a relaxed but focused and purposeful conversation. The hardest part about interviewing those who are not already media-savy is to put them at ease and keep them from straying off-topic.

DIGITIZING

Install audacity if it's not already on the computer you plan on using to edit. Save the project immediately to a descriptive (who, when, what stage in the process you're in) filename where you can find it. Keep your various files organized intuitively in folders. Ideally, edit in stereo (edit, preferences, audio I/O, Channels, 2). 44100 is the prefered frequency rate to use (...preferences, quality, default sample rate). 16-bit samples are probably plenty (edit, preferences, quality, default sample format...).

The best way to get audio off a recorder is digitally through a USB connection, but unfortunately special driver files may be required to set this up, if the recorder you're using even supports this. The next best is to connect the recorder's line-out to the computer's line-in.

{ Most computers have a line-in port, but if not you may be able to get acceptable results through your computer's sound card's mic port. Right click on the speaker icon in the bottom right of your screen, open volume controls, click options and make sure "advanced controls" is checked. Open the options menu again, and select properties. Click on the button to the left of "recording," make sure the check box next to microphone is checked, and click OK. Make sure "select" under microphone is checked. Click "advanced" under microphone. Make sure microphone boost is OFF. Close. Make sure microphone volume is about between 50 and 70 percent. Now it might actually work as a line-input, maybe. }

The final option is to use a stereo wire with 1/8 inch / 3.5mm male ends ("mini to mini") to connect the recorder's earphone jack to the computer's line-in. If you don't have such a wire, you may be able to borrow one from certain computer speaker systems, WSCA, or Steve. Start by setting the volume on the recorder at half and experiment. It may sound fine while being recorded into audacity and sound bad when you go to play and edit it. If it's too loud, parts of the audio may end up crunchy. If it's too soft, you'll probably hear an ambient hiss after it's amplified. As you record it into audacity, make sure louder parts aren't going off the edge of the waveform. As long as you can see the sound at all and it's not just a flat line, you can always select all and amplify... this is digital amplification, so it shouldn't introduce any noise.

I've found that for me, Audacity often locks up at random times when recording more than a few minutes in stereo, which is a problem when digitizing an interview through line-in above. Digitizing in mono solves this for the most part (edit, preferences, audio I/O, Channels, 1), but then you should convert to stereo (see below). Apparently this problem is just the result of driver issues with my "Vinyl" Realtek AC'97 integrated sound device. I've upgraded to a Sound Blaster 24-bit card and haven't seen that problem since.

MINIDISK recorders are very useful. We use a Sony MZ-RH10, which can transfer audio digitally and fairly quickly over a USB cable, but you have to use Sony's proprietary "SonicStage" software to do it. SonicStage transfers each track as a seperate file. This would be fine, except depending on the volume levels and their consistency during recording, and whether or not you used the mic input vs line-in from a mixing board, it could arbitrarilly split your audio into countless tracks.

In that situation, I just go from the headphone jack to the computer's line-in as described above to record the audio into Audacity, ending up with one long track. BTW - many minidisk recorders allow you to do basic editing without a computer, just using the buttons on the recorder. It may work fine, or may corrupt your disk and permanently lose what you recorded.

EDITING WITH AUDACITY

If the steps below are unclear, other instructions are available online for novice and advanced users. If you have an additional audacity question, check to see if it's already been answered on the audacity help forum archive and then post your question to the audacity forum.

Editing can seem daunting at first, but isn't too hard once you get into it. The help menu may be useful, and just experiment. Feel free to ask someone who does editing already for a tutorial session. If you're not sure what an icon does, just hover the pointer over it and a word will appear saying what it does. But in general, all tasks in audacity can be done by...
1. left-click-dragging over a section of audio.
2. Clicking "play" will play just that section, "cut" (scissors) will remove it, "paste" (clipboard) will place it wherever the mark point is.
3. By putting the pointer over an edge of the selected area and holding a left-click when the finger appears, you can drag that edge to a new spot.
4. When you have selected exactly the section you want to manipulate, click on "effect" and select the type of change you want. "Amplify" is the most useful effect, and 'negative amplification' makes the selected area quieter.

While editing, save the project often (File, Save project [if your project doesn't have a name yet, select "save project as..." instead]). For extra safety (if you have the hard drive space to spare), periodically export the project to a WAV file (File, export as WAV).
Make sure your segment to air has no swears, filler words, long pauses, jarring noises, or areas where someone didn't speak clearly enough to be easily understood. Make sure it does have a consistant volume throughout, a periodic description of what's going on for people just tuning in, and a sensible progression. Brief musical interludes may also be a plus.

The "Noise removal" effect can be very useful, especially when producing a phone interview with a cheap telephone patch. Consistant, loud hums are easiest to remove. The latest BETA version of Audacity removes noise best, but here are the instructions for 1.2.6:
Make sure you get a really good sample of the noise and nothing else. Select a few seconds of background noise, click "effect," noise removal, and "get noise profile." Now, select the entire project (Ctrl-A does this quickly), go to noise removal again. Move the slider all the way left towards "Less." Click "remove noise."
Depending on how good your noise profile is and how much "More" harshly you had to apply it, there may be distortion in the end result, but hopefully it will be better at least. If the noise was changed during the recording, you may have to get separate noise profiles in each area and apply it only to the nearby segment. Ideally, always arrange to get clean audio from the beginning. The professional phone patch at the WSCA studio works great.

The "Noise removal" effect in Audacity is useful, but I've found WavePad also has good noise removal tools.

The "Normalize" effect is helpful to make a section of audio a standard volume, with peaks of volume not hitting maximum, and quieter areas louder. But the only effect that makes quieter parts louder relative to the peaks of the audio selected is "Compressor." Manually adjusting the peaks and quiet parts with amplify or normalize usually yeilds better results, but is time-consuming.

INTEGRATION

While we used to use OGG compressed files because of their many advantages, I've decided to switch to MP3 because it's so standard. If you're having trouble encoding MP3s directly from Audacity, click here.

Because MP3 and OGG use different "lossy" compression approaches, every step of our audio production should now use MP3 compression. If absolutely necessary to re-encode between these formats, use at least 128 kbps. While voice interviews sound acceptable at 40-64 kbps if in mono, and the final show is now compressed to 96 kbps stereo, all audio by Making Waves segment producers should probably be exported to 192 kbps stereo MP3 to avoid degradation due to being compressed multiple times. Due to licensing reasons, MP3's exported straight from audacity will be lower quality than WAV files converted to the same bitrate MP3 with CDEX using the highest 'quality setting'. The CDEX approach will take longer, but gives you a better sounding, smaller MP3 in the end.

Export your project as MP3, using a descriptive filename and a folder you can find. If the file size is less than 10 MB (right click, properties) email the file to Steve as an attachment, assuming your email system allows this.

{ Note that files exported from audacity as OGG can be difficult for people to open with any other program. This bug is rarely a problem (and I think is fixed in Audacity 1.2.4), but could be an issue if you want to directly post your segment to radio4all. If this is an issue, export as a 16-bit WAV (edit, preferences, file formats, uncompressed export format, WAV - Microsoft 16-bit PCM, OK; File, export as WAV, choose a sensible folder and name, SAVE). Then convert it to OGG using CDEX (F4, generic, check normalize box; Encoder, select "OGG Vorbis..." from pull-down menu, check stereo, check "on the fly encoding," check "use quality setting," move the slider between 56 and 100 kbps; Filenames, note or change WAV -> MP3 and "recorded tracks" folders, OK; F11, browse to the right directory [click on 3 dots in the upper right], click once on the filename you want to convert, check normalize, check "delete original" if you wish, Convert). }

If the file is larger than 10 MB, the prefered method of getting it to the producer is through our FTP server. "Active Mode" needs to be supported by your FTP program. If you don't already have such an FTP program installed, get FileZilla. Install it and run it. Click "edit," settings, firewall settings, and make sure "passive mode" is NOT checked. Contact Steve for details on accessing our FTP server.

Other possible approaches to get audio to the producer / DJ include using AOL instant messenger (which is user friendly but has advertising) or Filetopia (a secure AD-free "peer-to-peer" file sharing and chat system).

A simple approach to get audio to us is to burn an audio or data CD onto a CD-R or CD-RW (read / write, which means other data can be overwritten on it later, but CD-RW isn't generally as reliable) disc and hand-delivered to Amy or Steve, or if time permits and you confirm we'll be in Portsmouth in the next few days, you can leave it in the MW mailbox at the station (we can also accept audio on minidisk).
The various CD burning programs all look different, but most of these programs start automatically when a blank CD is inserted in the drive, then asks what type of disk you want to burn (select Audio / Music CD, not "MP3 player disk"), then lets you select and arrange what you want on the CD. This can be altered as much as you want until you tell it to actually burn the CD.

PRODUCTION PRE-PROCESS

Download all segments needed from the MW crew, radio4all, Between the Lines, etc. All content should be converted to 44,100 hertz stereo before it is imported into the final audacity project for each half of the show. Also, files that are not WAV, MP3, or OGG may not be importable into audacity at all - most of these problem files will be WMA. Covert such files with CDEX to a temporary WAV file (F12, browse to source file [click on the 3 dots in the upper right of the window], select it, normalize checked if possible and delete original unchecked, click convert. When converted, rename the output file as appropriate and move it into this show's folder).

Play all content with winamp to ensure that it's in stereo and at 44.1 khz. Some audio that winamp won't play may still successfully import into audacity. Fix frequency first if needed by starting a new audacity project (file: new), import the audio file (project, import audio file...), set the project rate in the bottom left to 44100, and export as WAV to a descriptive new filename. If necessary, convert that file to stereo with CDEX, by converting the WAV to a stereo 128 kbps normalized MP3 file (F4, encoder: LAME..., "don't delete" unchecked, stereo, on the fly checked, 128 kbps, j-stereo, 44100 samplerate. Generic: clear tag-comment and encoded-by fields, normalized checked, click OK. Convert, convert WAV to compressed audio file, browse to source file, select it, normalize checked, delete original checked, click convert. When compressed, close CDEX, rename the output file as appropriate [it should be in the OUT folder on the desktop] and move it into this show's folder).

Converting to stereo can be done more quickly in Audacity. To make an entire project that is mono stereo, select all, copy, select project - new audio track, and paste into that new track. Click the triangle on the first track (see the left side of screen) and select "make stereo track." Export to WAV or MP3 as appropriate.

All content from audio CD's can also be extracted to stereo 128 kbps normalized MP3 files using CDEX as well (insert CD, select the track you want, F8...).

PRODUCTION POST-PROCESS

Once you've edited together each hour of the show in audacity, including the news, imported all other segments, and put in transitions, it's time to send it out to the world! Here's how...

Export each hour of the show from audacity as a WAV, one at a time. Don't try to do anything else on the computer while it's exporting. Play the output files with winamp to ensure they are the expected length, and listen to a few different spots for anything unexpected, like areas that are silent but shouldn't be (This hasn't been an issue since we selected the "safer" option in Audacity [edit, preferences, file formats...]). Using CDEX, convert them (WITH deleting the original WAV) to stereo 96 kbps MP3 files (F4, encoder: LAME..., "don't delete" unchecked, stereo checked, on the fly checked, j-stereo, 44100 samplerate, 96 kbps, quality 0 (very high quality) no VBR. Generic: tag-comment: NH Making Waves, encoded-by: < our email >, normalized checked, click OK. Convert, convert WAV to compressed audio file, browse to source file [click on the 3 dots in upper right of window], select filename, normalize checked, delete original checked, click convert. When compressed, close CDEX, rename the output file as appropriate (it's in OUT on the desktop... format is MW-XXX-A_date_XXmXXs), you can leave the MP3's in the OUT folder for the 'MW clean' script to clean up later. Copy these files to a flash drive and bring it to an internet computer if necessary, and upload to the FTP server. Do VNC tasks (see below) and post to radio4all with that show's map (Amy and Steve know how to do this).

We now only encode to one format (96 kbps MP3) for every purpose, including the DJ playing the files live, radio4all, web streaming, and our archives (the backup hard drive and MP3 CD's at WSCA). After each show, there is also a process to enable web streaming. An outline of this includes VNC'ing to WSCA, COPYING (don't just drag, that moves them) the show MP3's from the FTP server ("uploads") to the "for-stream" folder, and running the "make-and-move..." batch file in "for-stream." This should result in the MP3's and an approriate mw-XXX.M3U file ending up on the web server.

Each week after the show is done, the mwclean batch file (d:\mwc#.bat) should be run which (in addition to doing backup and archiving tasks), creates a seperate folder in the shared folder (of the production computer) for all files for the next show by making a copy of "MW-xx-skel" and renaming it appropriately. It also puts everything needed to post the current show to radio4all in the radio4all-temp folder. If you aren't sure how to answer any questions the script asks, just press "n".

Occasionally, audio CD's of the show will be burned by the DJ (generally Jo Ann) at the station just before the show. To burn an audio CD from the production computer, put a blank CD in the drive. Select Creator Classic when prompted. Click on the "Music" tab and make sure "audio / music CD" is checked. Update label. Drag the audio file(s) to the bottom screen. Select burn CD in the bottom right. Play each track off the CD using a boombox to ensure they are the right length and everything works.

COMPRESSION: A few concepts and misc notes
Note that every time a "lossy" audio compression format like MP3, OGG, AAC, and WMA are used, some of the audio data is thrown away. Repeatedly re-compressing audio will eventually make it sound distorted. The higher the bitrates you use at each compression step, the less is thrown away each time. Also, you should try to minimize the number of times your audio gets re-compressed. Using loss-less formats like FLAC and WAV when possible also reduces degradation.

--- Updated Nov-2007 ---


For any proposed updates to this document, questions or clarifications, e-mail us at: