Recording Phone Calls
Intro from Jay Allison
For years, people have been asking Transom for advice on recording over the phone. Slowly, we’ve been acquiring phone interfaces and software tools for Jeff Towne to review and incorporate in his primer, and it is finally done. Jeff covers high and low tech options–analog phone couplers and hybrids, digital hybrids, cell phone taps, computer-based telephony like Skype, taping little mics to telephones, and instructions on configuring “mix minus”. Various hardware and software options are reviewed, with audio samples. There are links to other resources. This is a really good rundown. You can stop asking now. Drop by to tell us about your own tips and ask questions.
from Jeff Towne
The first thing to consider about recording phone calls it to make sure that what you’re doing is legal. Laws vary widely from place to place, with dramatic differences even between states in the U.S. The safest policy is to always be upfront about the fact that you’re recording, and to get clear consent from the person being recorded.
The second thing to consider is that from a purely sonic standpoint, phone audio is almost always a compromise. It’s often a necessity, and can even be used to dramatic effect, adding a sense of distance and mystery, but in terms of pure sonics, recording from the phone always involves some loss of quality.
All telephony involves some kind of bandwidth-reduction. Standard telephones filter out low, bassy tones below 400 hertz. Likewise filtered are the high treble tones above 3,400 hertz, a narrower bandwith than even a cheap microphone and recorder would capture. That filtering was very carefully designed to increase the efficiency of phone lines, while maintaining intelligibility. So even though a voice on the phone sounds thin and tinny, it’s usually pretty clear and comprehensible. Cell phones, satellite phones and internet-based communications generally employ various forms of data-compression and bandwidth-reduction to increase the efficiency of their transmissions as well. These all have some negative impact on sound quality, so for the cleanest audio, no method of capturing a phone call will match recording both ends of a conversation at full-bandwidth with good recorders and microphones. There’s still a place for what’s known in public radio as the “Double-Ender” or the increasingly anachronistic term “Tape Sync” in which an engineer is sent to record an interviewee in-person while the interview itself is being conducted on the telephone.
The increasing speed of computers, internet connections, and wireless data transfer have made moving audio from place to place easier, and created some alternatives to standard telephony, but collecting interviews still often employs old-fashioned technology, sometimes called POTS lines (for Plain Old Telephone Service.) Even when using a modern technology like internet-based VOIP, the final leg of the communication might involve an analog telephone.
When it’s not practical to have the equipment or personnel to record the interviewee in person, there are several techniques to record phone calls that can yield acceptable results. The frustrating thing about phone recording is that there are always many variables beyond your control. The quality of phone connections varies widely, and can be affected by everything from the kind of transmission utilized by your particular phone service provider, and that of the person you’re calling, to the physical condition of the wires connecting each phone to the nearest switch.
Phones themselves vary widely in quality, and increasing numbers of people don’t even have wired landlines anymore. Ideally, your interviewee will be on a wired landline, not a cell phone, and not a cordless handset for a conventional landline. Headset phones are tricky, they can sound very good if the mouthpiece is positioned correctly, but make sure that the microphone part is positioned above, below, or to the side of the mouth, to avoid P-pops and breath distortion. If you have the opportunity, it’s worth suggesting this before the interview, so that your interviewee can arrange to use that kind of equipment, if possible.
On the recording end, if using a phone, rather than a computer-based Voice Over Internet Protocol (VOIP) service like Skype, you may need to have a certain kind of telephone to use some phone-recording devices. Some phone taps are designed to be plugged between the telephone base and the handset, so if you plan to use one of these, make sure your phone has a (detachable) cable on the handset. A cordless handset will not work with many conventional phone taps. Other devices plug between the wall jack and the phone, but most of those are designed for use with single-line phone jacks and cables. If your wall jack is a multi-line, you may need to use an adapter to split that line into two single-line cables. Some commercial spaces use digital phone lines, which require specialized equipment, and possibly special configuration by someone familiar with that particular system.
In many cases, getting sound off the phone simply requires inserting a box somewhere in the chain, and plugging a cable between that box and your audio recorder. But the quality of the recording, and the amount of control one has over the levels of the sources will vary widely based on the equipment being used. As so often happens, there’s a rough correlation between price and quality: digital hybrids that cost several hundred dollars, generally do give better quality and more control than taps costing much less. But depending on the situation, sometimes the relatively inexpensive boxes can be sufficient.
Henry Howard has a good overview of telephone recording devices here:
Analog Phone Couplers and Hybrids
The most basic device is the coupler. This type of interface is often sufficient for simply recording a remote caller, or for sending audio down a phone line. But a coupler offers little control over the relative levels of incoming and outgoing audio, which could create a problem if the intent is to record both sides of a phone conversation. The sound of your local voice will be MUCH louder than that of the remote caller, and both voices will be mixed together on one audio output.
This is not always a big problem, if the person on the local end of the conversation is careful not to speak over the caller, levels can be balanced in post-production. Perhaps the local voice will be edited out altogether, and if so, you may want to try muting the mouthpiece on the local phone handset while the caller is talking, or turning down the mic feed if the interface is being fed from a mixer to reduce extraneous sounds mixing in with the remote caller’s voice.
Here’s a recording made using the JK Audio That-2 coupler (approx $200.) This device connects between the telephone base and the handset, and mixes both sides of the conversation to a single output. As you can see on the green waveform representation above, and hear in the soundfile, the local voice is significantly louder than the remote voice, but the quality is not bad, and the discrepancy can be mitigated in post-production.
A little more control is gained by using an analog hybrid. These devices do a better job balancing the volume of the two sides of the call. Here’s the same conversation as above recorded on the JK Audio Autohybrid (approx $180.) As you can see in the blue waveform display, and hear in the soundfile, the two sides of the call are closer in loudness, although still not equal. This device has no gain controls for the respective sides of the call, and both are still mixed to one audio output.
For even more control, a digital hybrid can more effectively separate the two sides of the call, allowing for balancing the levels at the interface, and for sending each side of the call to separate mixer channels and/or recorded tracks. This allows individualized processing of the tracks to reduce noise or optimize EQ for each voice. The JK Audio Broadcast Host (approx $450) is very flexible, allowing a mic to be plugged directly into the device, or for audio to be fed from a mixer or other line-level signal, or both, each with a volume control. Audio is output on a stereo line level mini jack, which is easily split out to two mono cables, with only local sound on one, only remote sound on the other.
As you can see, the tracks are nicely discrete, but in getting a good loud signal from the remote caller, we’ve also picked up a good amount of hissy, rumbly noise along with the voice.
|That’s where having each voice on individual tracks comes in especially handy. Applying EQ to the remote caller’s track can clean up much of the noise. Many EQ plug-ins have a “phone” preset intended for use as a special effect, simulating phone calls by knocking out frequencies that aren’t transmitted by a phone line. That same EQ setting can be used to reduce extraneous noise: if it’s not within the range that sounds like “telephone sound” it’s probably noise.|
But even after EQing, there’s still some hiss, and some weird artifacts from the less-than perfect digital suppression of the local voice bleeding onto the caller’s channel. In fact there’s a low-level of very weird, ghostly voice bleeding over to the caller’s track, which can serve to undermine the sound of the local voice. Here it is amplified:
So it’s best to reduce both the remaining line noise and the ghostly voice bleed by ducking the volume when the caller is not speaking. An expander plug-in can do this, but it’s difficult to find a setting that will suppress the noise without creating chattering artifacts or clipped-off words. Sometimes tedious manual gain tweaking is a better choice.
Alternately, broadband noise reduction programs are getting better and cheaper. Here’s the caller audio tweaked with Izotope Audio Rx’s broadband de-noising algorithm. (Izotope de-noising software suite approx $350.)
The amount of noise on the line will vary significantly depending on the particular phone lines and telephone equipment.
The Telos One (approx $600-800) operates in a very similar way to the Broadcast Host, but the configuration of the outputs makes it best suited for use in a “mix-minus” configuration. Mix-minus requires a mixing board, and that mixer must have at least one aux send on the input channels. In complicated set-ups involving multiple simultaneous phone calls and/or multiple local sound sources, more aux sends may be required, but for the basic one-on-one interview, a small mixer and a single aux send will suffice.
In this scenario, plug the local subject’s mic in channel one. Plug the output of the phone interface into channel two. If this is being recorded and will be edited and tweaked later, pan the channels, one all the way to the left, the other all the way to the right. That way you will end up with each voice on its own channel. If you’re transmitting the session live, or just want to have it mixed live, leave each channel panned to the center. Next, connect the mixer’s Aux 1 output to the “line-in” or “to phone” jack on the phone interface. Then turn-up the aux 1-send knob on channel one to send a signal from the microphone to the phone interface, and therefore down the line to your remote caller. Be very careful to NOT turn up the aux 1 knob on channel two, which would create a feedback loop, resulting in screeching noises, or echoes, or both.
Wiring-up this mix-minus configuration offers several advantages. First, the local microphone can be recorded directly, not through the phone interface, or through the phone.
Second, the local host doesn’t need to hold the phone or wear a phone headset. He can hear himself and the caller through headphones connected to the mixer. The caller can hear the local voice through the aux feed to the interface, rather than through the telephone mouthpiece. Most interfaces can seize the line, allowing the receiver to remain on the hook. (It may have to be off the hook to dial-out, but after dialing, the line can be seized and the receiver hung-up. If the line is being called from outside, it can pick-up by seizing the line, and hang up by dropping it, usually controlled by prominent buttons on the front of the interface.
Third, multiple sources can be sent to the telephone interface by simply turning up the aux sends on each channel on the mixer that has relevant audio. This way, there could be multiple local subjects, each on a separate microphone, and the remote caller could hear all of them. Other audio can be sent down the phone line as well, whether an audio clip played from a sound source connected to the mixer, or the audio from additional phone interfaces for multiple remote participants.
Additional phone interfaces will require additional aux sends, one for each phone line, so make sure your mixer has enough simultaneous aux sends if you’re doing something fancy like this.
There are several tiny, inexpensive mixers available that will allow a basic mix-minus set-up with one phone interface, just be careful to check for an aux-send knob (sometimes labeled an effects send.)
The Behringer Eurorack UB802, Samson MDR 624, Peavy PV6, and Yamaha MG102c, are all very small mixers that have aux sends, and can be had for less than $100. In the $100-200 range, small mixers from Mackie, Tapco and Yamaha have been reliable performers in many production studios over the years. (Note that the smallest, least-expensive Mackie mixer, the 402-VLZ3, does NOT have any aux sends. So it would not be helpful for setting up a mix-minus rig for phone recording. )
Cell Phone Taps
It’s becoming increasingly common for people to rely completely on cell phones and computers for communications, and not even have a landline. Or you may be traveling and need to acquire sound while away from a conventional studio setting. Thankfully there are ways to record phone calls in these circumstances.
It’s never ideal to record from a cell phone. The audio has been subjected to data compression for increased efficiency of transmission, and while it may be sufficiently intelligible, voices often sound muffled and gurgley, and those artifacts may get worse as the soundfiles get compressed again later in the delivery process. That said, cell phone signals avoid many of the problems of landline phones, usually free from grounding hums and buzzes, and many cell phones do not provide a “sidetone” of your own voice to your earpiece, meaning that it’s easier to get a clean recording of the remote caller. In fact, that lack of sidetone sometimes makes it impossible to record both sides of the conversation easily. If both sides of the conversation are required, a small mixer or a version of the mix-minus may be required.
A few companies make cell phone taps, we tried a couple from JK Audio. The Celltap(approx $80) is super-simple: just plug a (provided) cable from your phone’s 2.5mm earphone jack to the Celltap. Then you MUST plug an earphone/mic into the Celltap. Plugging a cable into the earphone jack of a cell phone disables the internal mic and earpiece, and so you can’t hear the caller, nor can he hear you without an earphone/mic plugged into the Celltap. My phone does not provide sidetone to the earpiece and so my test recordings contained only the remote caller. As you can hear, it’s relatively clean, but it’s also gurgley, sounding like a low-quality MP3 file.
Similarly the Daptor 2(approx $160) can easily be interfaced with a cell phone, but requires a mix-minus configuration to facilitate communication with the other end of the call. This box doesn’t even provide a jack for an earphone/mic, and only a line-level in, so it’s best as a utility box to send audio from a remote location back to a studio, or connected to a mixer.
There are Bluetooth interfaces as well that will pair with a cell phone without cables, much like an earpiece will, but the additional stage of wireless communication strikes me as adding even more sonic compromise, so they may be fine for convenience, but perhaps not for critical audio.
There’s another problem with the cell phone taps: mobile phones often create electromagnetic interference on recording equipment. Some phones and networks are more problematic than others, but an active phone, even with the ringer turned off, can create bursts of noise that are picked up by microphones or cables. The usual solution is to turn off any nearby cell phones, as in power them completely down, but you obviously can’t do that if you’re using one as part of your recording chain. So, proceed with caution.
Keeping calls in the computer realm is no less of a mixed bag. The world of telephony is changing radically, largely due to VOIP (Voice Over Internet Protocol.) Computer processors and internet connections are now generally fast enough to allow smooth communication using packets of data, rather than a continuous electrical connection. Skype, as well as other real-time conferencing software packages, offer some exciting opportunities for affordable and easy recording of remote conversations. Of course there are challenges too.
On the positive side, a service like Skype offers very inexpensive communication, free in some circumstances. The software itself is free, and when connecting to another user who is also using Skype, that call is free. But the more practical application is that Skype can call standard phones. Rates to both domestic and foreign numbers are very low, compared to conventional long distance charges, and can be purchased in unlimited monthly packages or in pay-as-you-go dollar amounts.
On the technical side, VOIP call quality isn’t subject to some of the phone-line problems of a conventional call (although remember, if you’re connecting to a conventional phone, that “last mile” is still carried by regular phone lines, and bad wiring, or a bad phone, will still negatively impact sound quality.) But at least one generally avoids the hums and buzzes associated with the grounding differences inherent in the phone system and the electrical system.
Unfortunately, audio quality can vary widely depending on the speed that data moves along the internet, and that speed is rarely predictable. There’s usually plenty of bandwidth along the main trunks of the internet, but the upload and download speeds any user tends to get at any given computer can be unpredictable. If you’re connecting to the internet via WIFI, the throughput of your connection can dramatically affect the sound quality of your call, so make sure you’re getting a strong WIFI signal, or better yet, use a hard-wired ethernet connection when possible.
Beyond the strength of the connection to your modem or router, one needs a good fast connection to the internet. And perhaps even more important than the theoretical throughput of the connection is the absence of interference from networking tangles, and most crucially, firewalls. If your local network is slowing the throughput of your packets, you’ll get drop-outs or swirly, gurgley audio, as the data can’t keep up.
But if you can get a fast, clean connection to the internet, you can get surprisingly good-quality audio transmitted in real-time. As always, you’re only as good as your weakest link, and so if you’re ultimately connecting to a standard telephone, it’s still only going to sound as good as a telephone. But if you’re connecting directly from Skype to Skype, and the remote caller uses good quality equipment to connect to his computer, the resulting audio can be much better than the sound of a phone call.
There are several tricks for optimizing your connection for Skype, and they’re summed up extremely well by the folks at The Conversation Network here:
I highly recommend playing the video that describes all the steps to optimize your computer’s configuration – it’s daunting – I needed to watch it a few times, and skipped back for another look more than once, but it’s worth it. It’s very helpful to follow along: have Skype active and go to the preferences and control panels when they do, the video is easier to follow that way.
There are far too many variables in computer and networking configurations to address them all here, but some of the best tips from this tutorial are:
- Shut down all unnecessary programs while recording from Skype, especially anything that could be accessing network resources.
- Make sure you have a sufficiently fast internet connection:
At least 100kbps up AND down.
- Make sure your version of Skype is up-to-date. There have been some major revisions to the codec, so make sure you’re running version 2.6 or later for Mac, version 3.2 or later for Windows.
- Open Skype’s Technical Call Info window (Preferences>.Advanced>>Display Technical Call Info) and make a test call. Check for Packet Loss, number of Relays, and whether you’re connecting using the SVOPC codec. If you’ve got a high percentage of packet loss (above 10%) or if you have a large number of relays (preferably none) you’ll get dropouts, and you may need to dig around your network settings, or more likely adjust your router’s firewall settings. If you’re not connecting using the SVOPC codec, make sure you have a recent version of the software, and that your connection is fast enough.
Sadly, there may be some circumstances under which you just won’t be able to get sufficient throughput to get clean audio over Skype, and sometimes those conditions are beyond your control. Network congestion happens, and you never know what’s happening on the other end of the line, so if relying on this technique, you may need to accept that occasionally you just won’t get a solid connection, and you may need to try again later. But if you follow the tutorial linked above, you’ll increase your chances of getting clean audio.
That tutorial also includes some good advice about recording techniques to optimize your sound quality. When possible, they recommend connecting Skype-to-Skype, and that both parties use a USB headset mic. If it has a boom that comes around in front of one’s mouth, you’re advised to position that boom up about nose-height, to avoid P-pops and breath sounds.
Skype also has some tips for adjusting your sound settings here:
If the distant person you’re recording also uses Skype, and also has a clean, fast connection to the internet, and is speaking into a good-quality microphone connected to his computer through a good quality interface, rather than a phone, the remote voice can sound quite good. Meeting all those conditions is a lot to hope for, but more and more people are using Skype to communicate with distant friends and relatives, and random people might have some of this software and hardware. Even a mid-level USB headset mic can sound much better than a telephone when connected this way. Even better would be a good quality USB studio microphone, and even better than that would be a high quality mic connected to a USB interface.
You do NOT want to have the remote party speaking into a built-in mic on his computer, monitoring on the computer’s speakers. That works fine for informal communication, but actually sounds oddly distant, echoey and clipped when one pays close attention to it. A small step-up is an analog headset mic plugged into the computer’s mic and headphone jacks. USB phones reduce the echo and clipping of built-in mics and speakers, but still sound a bit like telephones. Vastly preferable is a device that amplifies and digitizes a good mic outside the computer, like a USB headset mic, or any kind of external USB interface. In some of the samples below, the remote caller actually sounds cleaner and more present than the local host doing the recording, particularly when the remote caller was using a good microphone and interface.
Here are some tests run using Skype over the internet. Both local and remote parties are using MacBookPro laptops, connecting to their local networks via WIFI. We were careful to make sure that each computer had a strong WIFI connection. During tests, less than ideal WIFI signal strength created dropouts and garbled audio. Both computers were on fast cable-model connections. Skype’s Technical Data window was showing near-ideal conditions.
Local host (first voice) on USB phone – remote guest on standard wired telephone landline.
Local host (first voice) on USB phone – remote guest on laptop built-in mic and speakers.
Local host (first voice) on professional microphone and USB interface – remote guest on USB telephone.
Local host (second voice) on USB phone – remote guest (first voice) on professional microphone and USB interface.
If both parties are using recent Macintosh computers, iChat (which comes pre-loaded) offers another alternative. Like Skype, audio or video chats can be conducted, and external audio devices can be used to improve the quality of the audio. There’s even a built-in recorder function, but we found it didn’t work very well, resulting in unbalanced levels between the two sides of the conversation. It also couldn’t split the local and remote voices to separate tracks, narrowing the options for fixing imbalances in the mix. Audio Hijack can capture the audio from an iChat in the same way it does a Skype session, and the result might even be a little bit cleaner and smoother than Skype, although the absolute audio levels are noticeably lower.
Local host (first voice) on USB phone – remote guest on professional microphone and USB interface.
Using a good mic and USB interface does yield results that rival a face-to-face recording, but there are still some compromises: there’s a slight gurgle and smeary quality to the sound under even the best circumstances. But using this good equipment allows for significantly better sound quality than any phone call.
There are several good-quality microphones with built-in USB interfaces available, and any good mic can be plugged into a USB interface, the simplest of which is something like the Mic Port Pro(approx $150). Just plug the mic into the XLR end, plug the USB cable into your computer, and then pick that as your audio input source. The Mic Port Pro has a headphone jack and gain knobs for the mic level and headphone volume. It can also provide phantom power to condenser mics. There are many other USB audio interfaces with microphone inputs that would work equally well.
Of course this is unrealistic in a standard interview scenario, you can’t expect the average person to have the equipment or knowledge of how to configure it. In those circumstances, you may just need to use Skype-Out to connect to a standard telephone, or if your interview subject has Skype and a good internet connection, you may get lucky and be able to ask them to use a USB headset or some input source better than a telephone.
If you have an ongoing need to record someone from a remote location, it might be worth getting better input devices for that remote party, like a good USB microphone, or a mic and an interface. But it’s important to remember that the audio is still compressed for efficient transmission over the internet, so this real-time audio is never going to sound quite as good as a full-bandwidth recording made at the source. If you don’t need the remote audio in real-time, it may be worth having the remote party simply record his end of the conversation and then send you the soundfile via FTP. But if you need the audio quicker than that, or don’t feel like you can count on getting the audio sent to you, you might want to record the Skype conversation. It’s never a bad idea to do both things: have the interviewee record himself and send the file, and record the phone conversation also, as a backup. The locally-recorded file almost always sounds better, even if it suffers from P-Pops or bad levels.
Getting your computer, and Skype, to use the better microphone as a source is a matter of selecting the proper device in your computer’s sound control panel, and in Skype’s Preferences>>Audio. Windows machines may need drivers for some devices. The audio inputs and outputs may need to be set in both Skype and at the system level.
Once you have your inputs and outputs configured, the good news is that unlike with a standard phone line, you don’t need any extra equipment to record from Skype. There are inexpensive utilities that will allow you to record any audio that passes through your computer directly to your hard drive, so you won’t need to buy any external boxes, figure out how to wire them up, or worry about hums and buzzes from unmatched electrical grounding!
For the Mac, Audio Hijack Pro ($32) is highly recommended. It’s easy to configure, and under Input>>Advanced you can choose “Megamix Mode” which allows you to record incoming audio to one track and outgoing audio to another, for greater flexibility in balancing the sound of two sources, or applying processing to only one of the voices. Under Recording, choose “For Burning to CD (AIFF)” as the file format, to maintain the maximum quality.
The Megamix Mode creates two audio tracks, one with the local voice, one with the remote voice, which makes for much more flexibility at the edit and mix stages. As with the digital hybrid used with standard phone lines, the two separate tracks can be treated differently with EQ and volume automation, or the local voice can simply be thrown away, then the session can be bounced or exported as a mono file.
There are many similar programs for Windows. Total Recorder is an inexpensive ($18), easy and flexible program that allows capturing both ends of internet telephony, and even includes basic editing functions.
A quick search around the internet will reveal plenty of other software that can record audio running through your computer. If all else fails, one can connect your computer’s sound outputs to an external audio recorder, even another computer running a recording program. This is significantly less convenient than simply capturing the audio within the same computer, but it does lighten the processing load on the computer making the internet call, and so might be the only solution in a pinch.
Recording Skype calls or iChats, or other computer-based teleconferencing sessions has the distinct advantage of being relatively low-cost, and relatively free from many of the noise issues of standard phone lines. When the interviewee uses a good microphone and converter, the audio can even sound much better than a phone call. But even the best input quality is subjected to file compression and the timing inconsistencies of packet data transmission, resulting in some negative audio artifacts. Even at best, there’s a slight gurgling and swishing quality to the sound, at worst, drop-outs and garbling can render a session unusable. The biggest problem is that data throughput can be slowed at random, uncontrollable times, and audio quality will suffer.
There are other low, or medium, tech solutions to recording from the phone. Olympus makes a small microphone embedded in an earplug, which simply makes it easy to record from any phone’s earpiece (approx $20). That microphone can be connected to any recorder that takes a mini plug, not just Olympus recorders. The main advantage is that it holds the mic element in a good position to pick up incoming sound, but one must be careful not to rub the phone earpiece against it.
A similar result can be gained, in a slightly less elegant fashion, by gaffer-taping a small lavalier mic to a phone receiver. Even better are dual lavaliers, like the (sadly) discontinued Radio Shack model pictured here. People have even reported success pointing a microphone at a speakerphone, which could surely work in an emergency.
Increasing computer speed and internet bandwidth will likely offer-up even more options in the future. But for right now, sometimes the lowly phone, not much changed from its debut over 150 years ago, offers the best way to communicate with someone far away. With reliable, fast computer connections, one can get near-studio quality audio in real time over the internet. Using some of the techniques described above, we should be able to capture that sound and spread it even further.
Thanks to JK Audio for lending us the Daptor-Two and THAT-2 interfaces for evaluation. And thanks to Viki Merrick, Chuck van Zyl and Herb Wolfson for making small talk on the other end of phone lines and Skype sessions as I tweaked my record-settings.