Title: Making, Managing and Analysing Recordings of Interviews and Focus Groups
1Making, Managing and Analysing Recordings of
Interviews and Focus Groups
- Duncan Branley
- Goldsmiths College, University of London20
January 2005duncan_at_gold.ac.uk
2Contents
- What do you want to do with your recordings?
- Recording and working with sound.
- Equipment for interviews and focus groups.
- Transferring your recordings onto computer.
- Improving your recordings.
- Transcribing.
- Moving transcriptions to NVivo.
- Automatically coding for turn-takers in NVivo.
3What do you want to do with your recordings?
- In-depth, qualitative exploration - longer term.
- Overview for commissioned research short
timeframe. - Archiving formats and metadata.
- Publishing on CD or the web
- Quality broadcast v research interviews.
4What is sound?
- Energy felt as vibrations.
- Converted to electrical signal.
- Heard or recorded.
- Frequency Pitch
- Hertz (Hz) and kiloHertz (kHz) waves/second.
- Amplitude volume
- Human ear hears
- Range 16Hz - 20kHz
- Range of normal speech 250Hz 8kHz
- Recorders frequency response
(Images source www.howstuffworks.com)
5Analogue Digital Recording
- Analogue exact replica - magnetic
- Digital takes snapshots or samples
- Sampling rate/second
- Sampling precision or resolution
- Each block is one sample
- Higher values match sound wave better
- Channels number of signals
(Images source www.howstuffworks.com)
6Digital Recording and File Size
- How many samples in 1 second?
- CD 44,100 upper hearing range 20,000 Hz
- Sampling precision
- 8 bit (256 possible values) or 16 bit (65,536)
- 8 bits 1 byte, 16 bits 2 bytes
- Channels monaural (mono) or stereo
- 1 minute of stereo sound at CD quality
- 44,100 samples/channel/second x 2 bytes/sample x
60 seconds x 2 channels 10,584,000 bytes
10,336 KB 10.1 MB - mp3 and RealAudio compression 1/10 1/14
7Recording and Working with Sound
- Sound converted
- From analogue to digital to record (ADC)
- From digital to analogue to play (DAC)
- Compressing digital audio files saves space
- Uncompressed WAV (PC), AIFF (Mac)
- Compressed
- Space double highest frequency of source
- Speech 250 Hz - 8,000 Hz gt16,000 or 22,050
- Telephone 400 Hz - 3,400 Hz gt 8,000
- See Ready Reckoner
8Analogue and Digital Audio Compared
- Sound quality
- Not necessarily better depends on sampling rate
- Main problem with analogue recordings hiss
- Copies
- Each analogue copy degrades, digital doesnt.
- Unless compress already compressed file.
- Confidentiality blanking identifiers
- Storage
- Media degradation v file format obsolescence
9Recording Equipment (1)
- Practice beforehand only one go at interview
- If disaster, write details of research encounter
asap - Consumer products v professional
- Easily replaceable supplies
- Ubiquity may reduce alienation
- ALL equipment should have
- Recording indicator
- Ability to pause and resume easily
- Mike-in connection
- Line-in may not produce loud enough signal
10Recording Equipment (2)
- Cassette recorders
- use at least Chrome CrO2 tape.
- turning tape may be a distraction, but longer
lower quality - Dictaphones
- Convenient but not too good quality often mono
- Minidisc
- Longer recording time 74-80 mins LP
- Better quality recording less mechanical noise
- Microphone
- External usually produces better results
- Best if a microphone-in connection if just
line-in may not produce loud enough signal. - Lead rather than direct plug-in lessens machine
noise.
11Preparing Before you Arrive
- Check batteries or other power supply and carry
spares - Have spare media unwrapped and labelled
- Have a checklist of how to use the recorder
- Carry the instruction book
- Cassette recorders and dictaphones
- Wind to beginning of recordable section
12Preparing the Recording Environment
- Sound from speaker moves in all directions
- Hard surfaces reflect delay gt echo
- Soft furnishings, smaller room and in corner
- Background noise
- Ambient noise in room
- Outside intrusions
- You and your participants moving, eating
- Microphone
- Position as close to interviewees as possible
best 30 cm 1 m - If focus group, place in middle to pick up all
sounds stereo useful - Place on stand naturalises, but beware
vibrations - Check the recording level although you may not
be able to - This will amplify the signal from the microphone
13Recording Telephone Interviews
- Same advice but some extra issues
- Permission to record ethical and legal
- Thinner recorded sound frequencies 400 -
3,400 Hz only - Could record from handset/speaker, but not good
quality. - Rather RJ-11 connector
- As easy as connecting a computer to a modem
- Position between telephone and handset balanced
signal - Telephone and line gt your voice louder
- Also may not work with advanced systems
- Then connect recorder
- Might want a Y splitter too to use two
recorders
14During Interviews and Focus Groups
- Concentrate on the interview!
- Monitor the recorder from time to time to ensure
that its still working.
15After Interviews and Focus Groups
- Dont stop recording until interview is really
finished - You may get some really interesting material
towards the end - When finished prevent over-writing by disabling
media - Minidiscs slide the tab
- Cassettes click the tabs out
- Label the recordings so that you know whats on
them later - May want to label using a system to preserve
anonymity - Work it out in advance
16Transferring your Recordings onto Computer
- Connections whats making and whats recording
the sound? - Source Line-out or Headphones/Speakers
- PC soundcard Line-in or Microphone
- Set up sources using Windows Play Control
- Set volume level on sound source.
- Start recording on sound recording software
- Audacity free to download and use.
- May have own with eg soundcard or CD drive
- Then start playing the source.
- When play finished stop recording.
17Connections
- (Images source http//geraldtomyn.tripod.com/diga
n.htm) - Miniplug cable 1/8 inch or 3.5 mm stereo (two
rings) jack at each end - Carries Analogue only.
- Source Line-out or Headphones/Speakers
- PC soundcard Line-in (arrow into curves) or
Microphone - Connect speakers to PC soundcard line-out to
monitor sound.
18Windows Volume and Recording Controls
- If youre not getting any sound this is the first
thing to check - click on the volume button on the Windows toolbar
- Recording you can only have one source selected
at a time. - The volume on both sides affects the loudness of
the soundfile - If it is too loud, it will distort too quiet,
you wont hear it. Test it.
19Recording with Audacity
- Start Audacity creates a new project
- File Preferences
- Audio I/O make sure its the same as the Volume
controls probably your soundcards name. - Quality Sample rate 22,050 Hz format 16 bit
dont need CD quality - File format WAV (Microsoft 16 bit)
- File Save As give it a name
- Click the circular red record button
- Start playing the source
- When finished click the yellow square stop button
- File Save
- To use in other programs you need to export the
file - File Export as either WAV (full quality) or mp3
(slightly lower, but smaller file size).
20Improving your Recordings in Audacity
- Have created a WAV (PC) or AIFF (Mac) file.
- Keep as original work with a copy.
- Could compress mp3 fine for speech.
- Can work with parts of file (drag to select)
- Make quiet parts louder.
- Effect Amplify accept default suggestion
- Blank identifying information (possibly after
transcription) - Click on silence button (wave with flat line in
middle). - Remove hiss find hiss with no words be
careful - Effect Noise Removal 2 steps
- Save brief extracts as separate files
- File Export extract as
- Remember to save your copy file.
21Transcribing
- Do you need complete transcriptions?
- Summaries in Word or NVivo (with labelled proxy
document) partial transcription. - Useful first step in field and for
familiarisation - Transcribe cassettes in full foot-pedal control
or auto-rewind - Use Transcriber to tie your transcription to the
sound files clause by clause cp sub-titles. - Can export to NVivo or a webpage.
- Transcription not neutral part of your analysis.
22Transcriber
- This enables you to create your transcriptions
while following your sound files on screen. - Start Transcriber and open an audio file WAV or
mp3 - Click on the play button and youll hear it.
- When theres a pause, press Enter and youll
create a segment break - When finished select Signal Playback Mode
Loop on Segment this enables you to replay each
segment while youre transcribing it. Press Tab
to stop and start it and the up and down arrows
to move between segments. - When finished set back to continuous playback.
- To create a new speaker Segmentation Create
turn or Crtl-T - You can double click to enter the full
identifier. - When finished save and can export to html or
NVivo format.
23Moving Transcriptions to NVivo
- From Transcriber
- Must ensure have NVivoTxt.tcl installed
inC\Program Files\DGA LDC\Transcriber\converte
r - File Export Export to NVivo Plain Text
format. - This will create a plain text file with the same
name as the soundfile you are working on except
with a .txt ending. - In NVivo Create a new document by importing this
file using the 'Read marked-up title and
description paragraphs' option.
24Automatically Coding in NVivo
- Document formatted with Heading styles.
- NVivo can see structure.
- NVivo can use this to code for each turn-taker.
- Can use for structured, semi-structured and
unstructured interviews.
25More Information
- Duncan Branley (2004) Making and Managing Audio
Recordings in Clive Seale (ed) Researching
Society and Culture (2nd edition) Sage, London
pp 207-228. - If you have further questions, email
- duncan_at_gold.ac.uk