What do I do when I get a poor quality audio file to transcribe?

[forwardvision]

Help! I need some advice please. I've picked up a new client, who has asked me to transcribe some interviews. I have 4 hours of audio to transcribe. There is a lot of background noise and whilst the interviewer's questions are clear enough, the interviewees' answers are not. Like I said I have 4 hours of audio, and with all the rewinding, listening, re-listening, re-listening again & then again, it's taken me almost an hour to just 5 minutes of audio. I'm starting to get a little frustrated.

I've done audio transcription before. But that was in an office setting, typing meeting minutes and letter dictations.

I have emailed the client, outlining the challenges I'm having but I'm still awaiting a response. (I only have email contact details for this client).

So firstly, is this sort of challenge usual with transcription work. And secondly, how do I get around this issue?

Anyone got any advise or suggestions please?

[Tess]

When I was doing transcription (note: WHEN...I don't offer this service any longer ) I had a fee for 'good quality audio' and a fee for 'poor quality audio' with the fee for poor quality being higher, naturally. Having this set in place before you accept the project is so important, otherwise you're a bit stuck with the rate you agreed to initially. (If you don't have it set up this time around, now you know and can implement it in your fee schedule for next time!)

Have you tried slowing the audio down a bit? That sometimes helps with clarity issues as well.

[forwardvision]

Have you tried slowing the audio down a bit? That sometimes helps with clarity issues as well.

Hi Tess,

Firstly, thanks for your advice and for reponding so quickly after I just posted this question.

I'm not using any specialist software or equipment for this task. Just playing the audio through QuickTime & listening with my earphones.

I'm not sure if it's possible to slow an audio down. Do you know how?

[Tess]

With QuickTime I have no idea.
I've always used Audacity - it's free and has a speed control feature, among other things which are super fabulous when you have clients who need sound files from radio shows or interviews cleaned up and prepped for loading to a website. Also great for transcription (in my experience) - I never used a foot pedal or those specialist tools either...

Hope that helps

[MoragB]

Tess has already given excellent suggestions; I have a rate for clear audio and another for poor quality audio - it can talk almost twice as long to transcribe a bad piece of audio.

I use Express Scribe and it's got an audio cleanup tool on there which can elimiate background noises.

Good luck & please let us know how you get on!

[forwardvision]

Thanks to both Tess & Morag for your advice. I've actually got a copy of WavePad on my laptop. I must admit, I'd not thought about using that. Actually, I'd completely forgotten about it.

Anyway, I'll have a play around with that. I reckon this could take a good chunk of the weekend. But it's all experience!

I've had a response from my client. She has advised me that if any parts are just to difficult to catch then just type "INAUDIBLE" and the time index. So that's helped a bit.

Just one more question if I may, and this is probably gonna make me sound like a complete novice. But here goes. How important is it to get an exact word for word transcription? I mean, I'm no short-hand typist. I have done some audio transcription in a work setting. But that was a little different to this.

[Tess]

I'd say pretty darn important....unless the client has specified she is using it for something enirely casual and specified that it's not important. I've done transcription where the client wanted the background/ambient noises of an event recorded as footnotes to the main content - talk about exacting work! It's good to get a feel for what the client is looking for at the get go but in most cases (99%) I'd say it's important to get every word and syllable and leave it up to the client to edit the content out that they don't want.

Just do your best It sounds like you got saddled with a really crummy file...and isn't this THE best learning experience? LOL - next time you'll have a whole new approach to a request for transcription.

Good luck with this one! And I hope you get a chunk of weekend for yourself too

[Prime Admin Solutions]

I don't transcribe the "ums". etc. and if I get bad quality audio I tell the client immediately that it is going to take much longer to transcribe and let the decide if they want to go forward. I had that situation last month. I called the guy and told him that it was poor quality and he said he really needed the transcirption so I did it. It is mentally exhausting to transcribe poor quality audio. I had a 96 minute transcription last month that was professionally recorded and it took about 4.5 hours to transcribe compared to a a really poor quality audio recorded on a digital recorder on a table in what sounded like a coffee shop. That took over 7 hours to do. I think those I won't even bother with in the future.

[Tess]

LOL...I agree with Julia and should have been more clear...I wouldn't include the 'ums' either

[MoragB]

lol....Tess, I've got a client who wants each and every 'uhm', 'erm', etc!!!

What I do now is offer 3 types of transcription and ask the client to state on the submission form which they would like - saves any misunderstandings later on.

Here's how I define the three different types:

Verbatim: Includes every 'urm', 'ahh', [pause], [cough], background noise etc, etc;
Semi-Verbatim: Omits the 'urms' and 'ahhs' but includes pauses, change of thought, and false starts; and
Intelligent Verbatim: This essentially produces a 'clean' piece of transcript with no 'urms' etc, pauses, changes of thought.

I ask the client to stipulate whether they want background noises and interruptions to be include in the transcript.

Also, if it's the audio contains technical words or terms / names it's useful to have a section on your submission form for the client to enter those.

Hope this helps a little!