Chapter 11

Using Recorded Speech

In This Chapter

Setting up NaturallySpeaking for portable recorders

Dictating text and commands on portable recorders

Transcribing voice files

Correcting transcriptions

Thank Thomas Edison. He’s the guy who figured out that you don’t have to be standing there talking to your transcriptionist. Nope. Instead, you can record stuff in the privacy of your own office and hand it off to your transcriptionist later. Of course, in Edison’s day, people had to yell into a big horn and their words were recorded on a wax cylinder. But the idea was a good one: Record now, transcribe later.

The idea was so good that now you can use any number of wonderful gadgets to dictate into while you’re on the go. This chapter deals solely with a voice recorder.

If you want to use NaturallySpeaking with mobile devices such as the iPhone, iPad, iPod touch or Android, see Chapter 14.

In this chapter, I explain what you need to know about setting up your equipment, speaking into a recorder, and having NaturallySpeaking transcribe that recording.

Why Record?

Perhaps the most attractive benefit of recording first and transcribing later is the same one that Thomas Edison probably had in mind. I suspect that he had no desire to master the manual skill of typing, preferring to let his assistant Watson do that job. Likewise, you can simply dictate your text and then let your assistant (if you’re lucky enough to have one) handle the transcription. That way, you never have to master the intricacies of NaturallySpeaking itself. You still have to master the intricacies of using a recording device, however.

One of the nice things about recording on a portable recorder is that you can sit there talking into a little box in your hand (a recorder). It doesn’t strain your eyes or cramp your fingers.

The second advantage of recording first and transcribing later is that, surprisingly, it’s often more accurate! Because your recorder provides a digital audio file, NaturallySpeaking’s transcription doesn’t have to keep up with your rate of speech. It can take its time and read your speech from the file at its own rate. As a result, it will be more accurate.

The disadvantage of recording first and then transcribing is that you don’t get to correct NaturallySpeaking on the fly. As a result, you may find that you have to make the same correction repeatedly throughout your document. Subsequent documents will, however, benefit from your corrections.

One fantasy to shoot down right now is the one in which you transcribe meetings using your recorder and NaturallySpeaking. One problem is that NaturallySpeaking has to be trained to each speaker’s voice. What’s more, the acoustic environment for meetings is invariably far too poor to get a decent recording from even one person. Besides, who really wants everything he said in a meeting to appear in a transcription?!

Setting Up to Use a Portable Recorder

Hopefully, if you already own a portable recorder — it will be good enough to work with NaturallySpeaking. Nuance lists on its website (http://support.nuance.com/compatibility/default.asp) the categories of recorders that the company has tested with its products, as shown in Figure 11-1. In general, you need a good-quality recorder that outputs digital audio files. For serious remote dictation work, you want a recorder that enables you to store multiple separate recordings. Or do you feel lucky? You can try using your existing recorder and see how well it does.

You can improve the audio quality of some recorders by plugging a separate microphone into them. Look for a microphone jack on your recorder. Of course, a separate microphone often makes the recorder significantly less convenient and portable. “Stub” microphones (microphones on a short stalk) that do not make the recorder too unwieldy exist for this purpose. Check with a good audio equipment supplier.

You have to do a few things before you can make remotely recorded dictation work, including the following:

• You must be able to make a physical connection between your recorder and your PC. You can’t just let the recorder play into the microphone.
• You must install any additional software required by your portable recorder.
• You must train NaturallySpeaking to recognize your voice the way it sounds after being altered by the processes of recording and transferring to your PC.

Figuring out your connection

To get started, you need a physical connection between the recorder and your PC. If your recorder isn’t digital, you’ll have to use the Line-In connection discussed here.

Digital data transfer connection

Digital transfer most commonly takes place through a data cable, running from a connector on a digital recorder to a connector (usually a USB) on your PC. From this connection, you copy the data (your voice recording) to your PC’s hard drive. Check your recorder’s manual for details on how to make this connection.

Other possible ways to transfer digital data include a memory card that you remove from the recorder and place in a slot in your PC (or in a device connected to your PC). Check your PC’s manual or your recorder’s manual for instructions on copying the data from this memory card to your PC’s hard drive.

Use the Line-In connection (see the next section), if you have a digital recorder but it doesn’t provide digital data transfer. Some other reasons for using the Line-In connection with a digital recorder are as follows: You may not have the digital cable you need; your PC may not have a connector available; or your PC may not be equipped to read the recorder’s digital storage medium.

Line-In connection

If your recorder still uses tape (seriously?) or if the recorder is digital but you can’t transfer the data for any of the reasons given in the preceding section, use the Line-In connection on your PC. This is also called an analog connection (as opposed to digital).

A Line-In connection requires a cable from the audio output jack (a round hole) of the recorder to the round Line-In jack on your PC. You can use this sort of connection with any recorder that has a Line-Out jack or a headphone jack (sometimes marked “ear” or “audio out”). If you use a stereo recorder for a Line-In connection, you need a special cable or adapter that creates a monaural (single-channel) output.

To be able to create a User Profile from another audio source (your recorder), you must have Administrator privileges. If you are the licensed owner of the software, you likely also are the administrator.

Adding a dictation source to your current profile

Most people think their voice sounds pretty terrible after it has been passed through a recorder. So does NaturallySpeaking. In fact, as far as NaturallySpeaking is concerned, your voice is so different that it needs to train with the recorder to recognize it.

Just like the first time you used NaturallySpeaking, you used the New User Wizard to set up NaturallySpeaking. Additional training of NaturallySpeaking to understand your recorded voice is just like training it for direct dictation, with one difference: You read the training material into your recorder, transfer the dictation to your PC, and then have NaturallySpeaking transcribe it.

The process is as follows:

1. Choose Profile⇒Manage Dictation Sources, as shown in Figure 11-2.

   A window pops up and you see your current source and an Add New Dictation Source button.

2. Click the Add New Dictation Source button.

   

3. Pick your source, as shown in Figure 11-3.

   

4. Choose Digital Audio Recorder from the list and click OK.

   A screen pops up that says, “Recorder Training has not yet been successfully completed for this User Profile and dictation source.” Here’s your opportunity to do so.

5. Click OK.

   As shown in Figure 11-4, you see a screen saying it will guide you through the five steps to get your recorder ready.

   

6. Click Next.

   You will now configure your recorder. If you haven’t done so already, as noted earlier, click the link shown in Figure 11-5 to check for Dragon supported recorders.

   

7. Click Next.

   The wizard displays, as shown in Figure 11-6, a selection of fine, edifying reading material to choose from in order to train your recorder. You are presented with several options to read.

8. Make your selection and click the View the Selection button (or click the Print the Selection button if you prefer).

   A window appears with your chosen text.

   

9. Read the copy provided. When you are done, click the Back button to return to the previous screen (refer to Figure 11-5), then click Next.

   A screen, shown in Figure 11-7, tells you that you need to locate the file you just recorded.

   

10. Click the Browse to Locate the File button to locate the file on your PC that you just recorded.

    Find the file and choose it so that it appears in the window.

11. Click Next.

    You see a screen that shows the name of the file you have selected for training, as shown in Figure 11-8. If it is correct, you can proceed to training.

    

12. Click Start Training.

    The training process will take several minutes, so be patient. It is recommended that you don’t touch the computer keyboard or cursor until the training is complete. This gives you the opportunity to take some victory laps to refresh yourself.

    When the training is complete, you see a congratulatory screen, as shown in Figure 11-9.

    The next time you open your list of User Profiles, you will see that a new profile has been added with your name and your audio source: Digital recorder.

When you dictate using a recorder, NaturallySpeaking remains in Normal mode. If you are going to do something that requires dictation of a specific type, switch to one of the other Recognition modes: Numbers, Spell, or Command. (See Chapter 3 for more information on Recognition modes.)

Getting better sound quality from portable recorders

For lots of reasons, you are more likely to have sound quality problems when you use a mobile recorder than when you dictate into your PC. Here are six tips to avoid problems:

• Avoid noisy environments. Moving cars, traffic, machinery, wind, surf, weddings, car washes, rock concerts, or active airport runways can negatively impact sound quality and recognition.
• Don’t move your fingers around on your recorder while recording. This causes noise.
• Don’t speak so directly into the microphone that it records puffs of breath when you speak. Keep the microphone of your recorder off to one side of your mouth. Keep the microphone a constant distance from your mouth.

  See if you can fit two fingers in between the microphone and your mouth.

• Don’t set your microphone’s sensitivity higher than necessary. If your recorder has different microphone sensitivity settings (or a microphone volume control), use the lowest sensitivity that still gives you a strong, clear recording. (Too high a sensitivity picks up background noise and sometimes distorts your voice.)
• Choose the recorder’s highest quality setting. Make sure your recorder is set for the highest quality of recording, if it offers different quality levels. Check your recorder’s instructions. Highest quality usually comes at the expense of maximum recording time, so choose the setting that gives the shortest recording time if quality level isn’t an option.
• Train in the environment in which you’ll be dictating. That way, Dragon knows what to expect. If you train in a quiet room and dictate in a coffee shop, the accuracy will be dismal.

Recording Your Dictation

When you record text for NaturallySpeaking to transcribe, speak that text just as if you were dictating into NaturallySpeaking directly. Chapter 4 tells you how to do it.

Certain aspects of recording, however, make that process a little different from dictating directly to NaturallySpeaking. Using commands, for instance, is tricky because you can’t see the transcription in progress. In addition, dictating into a portable recorder introduces some new issues that affect sound quality. The best thing to do is to limit your voice commands to dictation commands.

Because you can’t see the result of NaturallySpeaking’s transcription as you dictate, using certain commands in recorded speech is risky. NaturallySpeaking might, for instance, edit or delete the wrong text in response to a command. You wouldn’t know that until you see your text on the screen.

Because of that risk, NaturallySpeaking ignores most editing commands it encounters while transcribing your recording. NaturallySpeaking does, however, accept dictation commands in your recorded text — the ones that control capitals and spaces. The safest procedure is to use commands that apply only to your next spoken word, such as “Cap <word>.” Even though NaturallySpeaking allows you to use the dictation commands that turn something “on,” such as “Caps On,” NaturallySpeaking may occasionally miss the concluding “Caps Off” or other “Off” command. You may end up making more work for yourself (or whoever does the final cleanup) by using those on/off commands.

Following are some of the commands that, in addition to punctuation, I think work most reliably in recorded speech:

• “All Caps <word>”
• “Cap <word>”
• “New Line”
• “New Paragraph”
• “No Caps <word>”
• “No Space <word>”
• “Spacebar”
• “Tab Key”

You can also use “Scratch That” (which deletes back to the last time you paused) if you make a mistake. Use it only if you’re sure when you last paused, or you’ll delete more or less than you intended! You can repeat the “Scratch That” command to back up through multiple pauses if your memory for pauses is very good.

To avoid having to remember your pauses, a better command for amending recorded dictation is “Resume With <word>.” This command enables you to back up to a specific word within the last 100 characters and then dictate new text beginning from that point. (Of course, it only works if NaturallySpeaking got your word right in the first place!)

See Chapter 4 for more about “Scratch That” and “Resume With.” Both commands are allowed when you transcribe from a recording. They are called the “restricted command set.”

Transferring Files from a Digital Recorder

When you use your recorder, you need instructions from the manufacturer for transferring audio files to your PC. It may have its own program for handling file transfers that you need to install on your PC. Check your recorder manual for instructions.

Where on your PC’s hard drive should you put the digital audio files from your recorder? You can put them anywhere, but the NaturallySpeaking transcription feature looks first in the Program folder in the NatSpeak folder on your hard (C:) drive, where NaturallySpeaking is normally installed. For convenience, put them in that Program folder.

If you have a digital recorder but it doesn’t offer digital output or you do not have the necessary cable or software to make a digital transfer, you may be able to make an analog connection instead. See “Figuring out your connection,” earlier in this chapter. Take your recorder to an electronics store and ask for a cable to connect its audio output jack to a PC’s audio line-in jack.

Transcribing Your Recording

Watching NaturallySpeaking transcribe a recording is somewhat magical. You sit there and your words (or something like them) appear on the screen.

How does it work? NaturallySpeaking transcribes recorded speech from a sound file (a file with a .wav extension or a WMA, MP3, DSS, or DS2 format), created by a digital recorder, which you have stored on your PC’s hard drive. (Oh, you haven’t? See “Transferring Files from a Digital Recorder,” earlier in this chapter.)

To transcribe a recording from a portable recorder, NaturallySpeaking must be set up with a special additional source User Profile specifically trained to handle recorded speech from that recorder. (See the earlier section, “Setting Up to Use a Portable Recorder.”) You don’t need to choose a special user to transcribe files from the NaturallySpeaking Sound Recorder (assuming the files were created on your PC); use the same user that you use for dictating directly to NaturallySpeaking.

Launch NaturallySpeaking if you haven’t already, and take the following steps to transcribe:

1. Click Profile⇒Manage Dictation Source and select the Digital Recorder source you previously created.

   The source you created for transcribing from a recorder loads.

   If the microphone in the DragonBar is red, your headset dictation source is loaded. If the microphone is gray with an X, you already have your recorder source open.

   Make sure to change to the recorder source before moving to the next step to avoid negatively impacting your dictation source data, because you sound different when using the recorder.

   If your portable recorder uses the analog (Line-In) connection, NaturallySpeaking will expect a certain volume from your recorder. When you transcribe text, either make sure the volume is set to the same level you used for training or run the microphone check again at this point. To run the training, choose Audio⇒Read Text to Improve Accuracy and choose the Adjust Volume Only selection.

2. In NaturallySpeaking, choose Tools⇒Transcribe.

   A Personalize How You Transcribe screen opens, as shown in Figure 11-10. Choose the defaults unless you have a specific window you wish to work from.

   

3. Click Next.

   A Transcribe a Recorded Audio File screen opens and asks you to select where Dragon NaturallySpeaking can find the audio file you want to transcribe.

4. Type in the file name or click the Browse to Locate File button to select the file you want to transcribe from your hard drive, as shown in Figure 11-11.
5. Click the Transcribe button.

   The DragonPad opens and begins transcribing.

Correcting Your Transcription

It’s kind of exciting to watch as your words print magically on the screen. The hitch comes when you spot errors. Just like regular dictation, you need to proofread and correct your errors.

To proofread, I recommend that you transcribe into the DragonPad or your word processor and use the capability to play back your own voice if you are unsure about what you actually said. Here’s a method you can use:

1. Begin reading your text. If you spot an error, you can say, “Select <text>” and then, “Correct That.”

   The Correction menu pops up just as it does for your regular dictation. Choose the correct number of the correct version or say, “Spell That” and correct it that way.

2. Use Playback when you spot an error by selecting the incorrect passage and right-clicking. Choose Play That Back from the list of options, as shown in Figure 11-12.

   You can alternate methods until you have corrected the entire transcription.

Correct Dragon’s mistakes, not yours. If you hear yourself say something incorrect and NaturallySpeaking transcribes it, just go over it and correct it. Dragon didn’t make a mistake, so it doesn’t need to be corrected. (Yes, this is an operator error, not a software error! And for the record, it happens to the best of us!)

Are you lucky enough to have a staff member or other willing participant who will transcribe your recording for you? Make sure that individual transfers your User Profile to his PC before he begins. If he doesn’t have your User Profile on his computer, the audio won’t be recognized properly. (Of course, your faithful human transcriber also has to own Dragon NaturallySpeaking software in the Premium or higher edition to hear your voice when playing it back.)