Protected: Transcribing Audio Recordings – The Options for Law Enforcement
The expectation of the public changes and evolves, and this can often steer what the courts expect from police in terms of producing evidence. Here, we want to focus on one aspect of the criminal justice system that is changing and that is the expectation that recordings with audio content, regardless of from where they originate, will be transcribed.
Let’s face it dealing with audio recordings is problematic for police because the number of potential sources is substantial, and the volume of material can be enormous. Adding to the problem are the different types of recording device in use and the different formats that these devices use. Audio may be recorded on purposely designed digital audio recorders, body worn cameras, iPhone and Android phones, laptops, and tablets, to name a few. File types most people will be familiar with include the two most common mp3, and wav. If this didn’t create enough of a problem for law enforcement, how to securely store these files will involve significant expense and can require a degree of creativity to make everything work in harmony.
While many organisations face problems regarding transcription, it is fair to say that for law enforcement there are additional considerations. For law enforcement getting spoken words onto paper is further complicated by several stipulations:
1. The officer producing the original recording must be able to prove its integrity. Each time there is an opportunity for a recording to be doctored the officer must prove it wasn’t.
2. The transcription must be of evidential standard. It must be an accurate reflection of what was recorded. The officer will attest to this.
3. The transcription must reflect who, says what, at which time.
4. The transcription must be produced in a timely manner. Investigations should not be held up because an officer must wait for the transcription and the submission of an investigation file should not be delayed because evidence is not available due to administrative backlogs.
5. Officers are rarely dealing with just one investigation. They are likely to have multiple cases many of which will require transcription. Prioritizing these can be difficult.
6. Agencies are likely to have limited resources. Here, again, there will be competing needs regarding transcription, particularly concerning prioritizing, and paying for transcription.
Before looking at the pros and cons of each of the options for transcription, let us look at what is common to all these options.
1. At all times a law enforcement officer must be able to prove that the original recording presented to the court has not been doctored.
2. The officer must be able to state that what is written in the transcription is an accurate record of what was said on the recording. Regardless of how the transcription is done, the officer will always have to spend time checking it is accurate.
3. There will always be corrections to do, though the number will vary depending on the transcription method used.
4. There will always be time used to get the recording from where it is stored to where it is to be transcribed and back again. The smoother this is the less cost there will be – but there will always be costs: for example, uploading a recording to an internal or external transcription service, or accessing it on the department’s server. In some cases, this may still involve signing a physical tape out of storage.
5. Dealing with the quality of the audio recording is likely to add problems.
6. Dealing with human speech patterns can cause problems. These include accents, dialects, and the speed and tone of speech.
7. Identifying the different voices of those involved in the discussion/meeting/interview.
Currently a lot of transcription is done by people. However, software transcription is rapidly advancing and becoming a real option. Essentially there are four options for any agency regarding transcription.
1. Self-Transcription. Each officer transcribes the material they obtain for their investigations.
2. Dedicated Transcription Staff. The agency employs staff whose dedicated role is to transcribe material for the entire agency.
3. Outsourcing. The agency outsources transcription through a commercial company. Here transcription may be done by a person or by means of an electronic product.
4. Transcription Software. The agency purchases/licences a software application that allows them to do inhouse transcription.
Let’s summarise the pros and cons of each option.
Option 1 Self-Transcription
• The officer has control over when the transcription is done.
• The original recording and the transcription are held under the control of the agency.
• The officer has intimate knowledge of the circumstances relating to the transcription and is much more likely to be able to transcribe it accurately and avoid errors.
• Physical transcription takes time. As a guide we can use a figure of four hours continuous transcription for each hour of recorded material. Officers are rarely speedy typists, nor are they particularly adept at transcribing. They are likely to be slower. An officer’s time costs. As a ballpark figure the cost of an officer is US$1 per minute or US$60 per hour(1) . This cost increases exponentially as the officer increases rank. At best, each hour of transcription will cost US$240.
• A trained officer’s time is much better used doing law enforcement work. Using them to transcribe is wasted skills. More senior officers are unlikely to have the time available.
• Often the officer will have competing priorities. Tying them up in transcribing duties will create conflicts with other duties.
Option 2 Dedicated Transcription Staff
• Support staff are often less expensive than a sworn officer per hour. This option is cheaper than using an officer’s time. The cost here is likely to be US$40 per hour
• Over time staff are likely to gain speed and experience in transcription and are more likely to produce better results.
• The original recording and the transcription are held under the control of the agency.
• Employing staff brings with it many human resource issues, not least being the number of staff to employ. Identifying the number needed to ensure product is delivered in a timely manner may be problematic.
• Systems need to be well designed to ensure how and who decides the prioritisation of the agency’s limited resources. For example, just because a sergeant believes their transcription is more important does not mean that the detective’s work gets pushed down the pile.
• Physical transcription takes time. As stated, we can use a figure of four hours continuous transcription for each hour of recorded material. However, working in a pool is likely to take longer as the transcriber must also manage the mail in/mail out aspects for many officers.
• The officer will have to wait until their transcription is done. This may vary from one or two days to one or two weeks.
• The process may need to be repeated if critical errors have been made in the original transcription causing further rime delay.
Option 3 Outsourcing
• The agency is freed from using their officers/staff in transcription.
• There are few, if any, human resource issues for the agency.
• Transcription may be done in a timelier manner. However, see below re additional charges.
• Potentially very sensitive material is being shared outside the agency. Data protection and privacy concerns are real. It is difficult and costly to vet the company and their staff.
• Getting the original material to them and the product back may be logistically problematic. Given that some officers may have a large number of short recordings to deal with a poor process here can be extremely frustrating for all involved, with more time being spent in mail exchange than in actual transcription.
• Those outside law enforcement are more likely to make errors in the transcription because of unfamiliarity with terminology used.
• Cost is likely to begin at US$1.5 per minute of audio or US$90 per hour(2) of transcription even if they are using transcription software. This may cost more if is being carried out by a person.
• If transcription is costing less when a person is used than would be a normal wage for that location, care must be taken that non-native speakers are being used. This is likely to impact on the quality.
• The turnaround time is set by the outsourced company. Additional costs are likely to be incurred for more rapid transcriptions.
• Additional costs may be charged where an audio file contains more than two speakers. Typically, a law enforcement interview will involve, two officers, a suspect and a legal representative for the suspect.
• Structures are required to ensure the integrity of the product.
• Structures are needed to facilitate who in the agency, can send what, and when, to the out-sourced company. Expenditure must be carefully monitored.
Option 4 Transcription Software
• Speed. Officers can rapidly upload a recording and, depending on its length have it ready for checking, moments later.
• Considerable cost savings. Obviously, the amount saved depends on the price of the software purchased and the number of hours of transcription needed but this type of software will save money. For large agencies, savings may be hundreds of thousands of dollars.
• Control of the original recording and the transcribed product remains within the agency. This ensures integrity of evidence and addresses data protection issues.
• Officers can have access to the software when and where, they need it. This leads to better use of their time and speeding up of investigations.
• Enhance public confidence. The public want to see police officers on the ground and they expect investigations to be carried out in a timely manner.
• Officer wellbeing is improved. Few officers joined to spend their day at a keyboard. Frustration with out-of-date technology and poor bureaucratic processes are often cited as being major sources of stress for officers.
• Care must be taken that the software product selected meets the needs of the entire agency.
• Using transcription software not designed for law enforcement use, is likely to omit features that are needed/desired by the law enforcement user.
• There will be an initial and an ongoing investment. Care needs to be taken that the outlay does not outweigh the benefits described. The agency needs to have a good idea of how many hours of audio material they currently use.
• The accuracy of the transcription must be very high, or officers will have to waste time correcting it.
When it comes to choosing, the first question one needs to address is: “What is the total number of hours of recording done within the agency each year by all staff?” The answer to this question is likely to dictate what needs to be done. Next, each option needs to be explored with the previous figure used to explore how much each option is likely to cost. If outsourcing is chosen, then care must be taken regarding what each outsourcing company offers and how they compete with others. If the decision is made to purchase off-the-shelf transcription software, then what each company offers needs to be examined and balanced against the expenditure.
In brief, unless the law enforcement agency only does a few hours transcription per year, then purchasing the right software is the most efficient option and the best financial choice.