Scribie.com

Last month we completed four years of our company. We launched CallGraph Skype Recorder in April of 2008 with intention of offering services around it. The transcription service was one of those and it quickly became the most popular. The past four years we have invested all our time and effort in developing a human-powered transcription system with a single goal in mind: deliver the best quality transcript with the lowest amount of effort. In this series of posts we are going to write about this system in-depth.

So why build another human-powered transcription system? Why not just use a Automatic Speech Recognition system. Speech recognition has been in the limelight recently most notably because of Siri which uses Nuance’s Technology. In fact in Google Glass it’s a central component. Even Evernote recently added the support for it. However all of these systems employ keyword recognition; eg. commands that are spoken aloud. Our requirement was conversational speech recognition. The technology for that is still very immature. In fact we tried out CMU Shpinx and results were so poor that we ruled it out.

The big issue with human-powered systems is that it produces inconsistent results. Transcription is very labor intensive. And just like any labor intensive workflow, if you do not have processes in place, you will not be able to control the quality. The typical transcription process involves one person doing the typing work and maybe, another person proofreading it. On an average it  takes around four hours to type one hour of audio and around the same amount to edit it. This increases the cost of transcription. And even after that the transcript is bound to have mistakes, thereby affecting its quality.

So that was the starting point for us. Our system manages the transcription process end-to-end. It’s like a machine where you input the audio file and it outputs a high quality transcript in one day. This system is powered by our certified transcriptionists who do all the work. We have a well defined workflow and a robust process in place. We use some Machine Learning and Information Retrieval tools as well, but for the most part, it is all done by hand.

With this system we have completed more than 3000 hours of audio transcription till date and managed to survive four years in a highly competitive market. The best part is that we a high return rate of customers. For a startup, it might not be a stellar achievement like Instagram’s, but we believe that we have built something substantial; a scalable and reliable transcription service. The next post will cover the first part of our system, the transcriber certification process. Till then, if you are in need of a transcription service then you should try out our transcription service today. You will not be disappointed.

The next part of the series can be found here.

We have recently added two new options which can be specified when ordering Transcripts. A screenshot is below.

The first one is Flexible Delivery. If Flexible Delivery is enabled then the we will deliver the transcript anytime within a week as per our schedule. In the worst case the delivery will be done after 1 week. But it can also happen before that if we find a slot for it. We offer a 10% discount for Flexible Delivery orders.

The second option is Additional Proofreading. In our transcription process we do one mandatory round of proofreading. For Additional Proofreading orders we do a second round of proofreading as well. This is useful for files which are difficult to transcribe, i.e. files which poor audio quality where the speakers can’t be heard clearly or which have noisy environment, or files which have non-native speakers whose diction is hard to comprehend. We charge $0.50 per minute extra for this option.

In addition to these options we have also do more quality checking than before. We also pay more attention to blanks which are leftover after proofreading and try to correct them as much as possible.

We’d like to thank you for using our service and your valuable feedback.

So, you have an audio file which needs to be transcribed to text and have no idea how to go about it. Here’s how to do it.

The first thing you would need is ExpressScribe. It’s a free tool which enables quickly to start/stop/pause playback with hotkeys (or foot pedals) so that your hands are freed up for typing. To setup the hotkeys go to Control -> Hotkeys setup and enter your preferred keys. We suggest the following.

• F7 -> rewind
• F8 -> forward
• F9 -> play
• F10 -> stop

The useful thing is that these hotkeys are global. So you can control ExpressScribe even if it’s not the active application, which means you can type, play, stop, rewind, and forward without leaving your text editor. Very useful. Load up a test file and play around with the setup.

The second thing you would need is a good headset. Transcription is time consuming and it will start paining around your ears if the headset is not comfortable. Any headset which has a bit of padding around the earpiece is good enough.

The third thing you would need is a text editor. You can use Word, OpenOffice.Org or any other editor of your choice. At a minimum it should have word completion and spelling auto-correction. While typing you’ll find that you misspell lot of words and to correct them you’ll have to stop and go back constantly. Auto-correction will save you time. Similarly auto-completion will save you typing, at least for common words.

Once everything is set up, you’re good to go. Bring up your editor, play the file, pause and type whatever you understood. Play-pause-type. Rinse and repeat till the file is complete. One pro-tip is try not to rewind too much. Rewind as few times as possible. Better still, instead of rewinding just mark the inaudible portion with a blank or make a guess, and after you’re finished go back and review the file. You’ll finish the file faster this way.

On an average it takes around 4 to 6 hours to finish 1 hour of audio. It varies by your typing speed,  the audio file quality and/or the diction of the speaker. Difficult files take longer. Plus, you’ll notice sometimes you cannot catch a few words, no matter how many times you rewind and play it back. Ask for a second opinion if you can.

As you would have guessed by now, it’s a painstaking task. It takes a lot of time and effort. That is exactly why we have the Audio Transcription Service. We do all the work for you and deliver a high quality transcript after 1 business day. For $59.40 you can get an hour of audio transcribed. We have a rigorous process and have transcribed over thousands of hours of audio to date. Try it out and check out the results for yourself.

On the other hand, if you like transcribing then check out our Freelance Transcription Program. You can work as a homebased freelance transcriber and get paid on an hourly basis for the work done.

If you’re looking to sell our your own e-book then have a look at Jared’s post at Startups Open Sourced. He’s had great success with it and has written up a detailed how-to of the steps involved. One of his tips is, get your interviews recorded and hire a transcriber to transcribe everything which you can work off.

One of the things to remember while hiring a transcriptionist keep in mind the difference between audio hour and man hours, which Jared mentions in his post. An hour of audio can take anywhere from 4 to 6 hours to transcribe. If you’re paying by the hour then your cost will basically be multiplied by that factor. On the other hand if you’re paying by the audio hour, then the amount of effort spend does not matter. You’ll pay for the amount of audio transcribed and not time taken for it to transcribe.

On Scribie.com we charge always by the audio hour. You don’t have to worry about the amount of actual time taken. Even if the audio is hard to transcribe and takes a more effort, you wont have to pay extra. At the end of the day you’ll still get a high quality transcript of your audio file.

We have moved to a new domain; Scribie.com. All the accounts data from CallGraph.Biz has been migrated to Scribie.com. If you had an account on CallGraph.Biz then you can now login to your Scribie.com account and check.

https://scribie.com/account/login

With Scribie.com the focus is more on our audio transcription service. We have been working on our outsourcing platform for a long time and we are very confident about it. It’s the best audio transcription service you’ll find out there.

We have also discontinued the online backup and Skype Group Collaboration services. We are not accepting new accounts for those. If you had been using those services then your files will still be available and you can still keep uploading. But, we are not planning to add any new features or maintain it.

We still offer the CallGraph Skype Recorder for free and plan to continue working on it.

We now offer two types of transcript: Draft and Proofread. If you plan to do a comprehensive review of the transcript then you can choose the Draft Transcript Type while ordering. The Draft Transcript is not time coded and the speakers are not tracked in it. It contains just the text of the audio file with each speaker’s diction paragraphed neatly.The rate for the Draft Transcript is $0.60 per minute. Therefore for a 1 hour file it will cost $36.

The Proofread Transcript on the other hand requires minimal or no editing on your behalf. This type of transcript is time coded and speakers are tracked with initials. The blanks are marked with a time-stamp to make it easy to locate it in the audio file and correct it. We also do a complete review. This includes researching any terms that might occur in the audio and finding out the correct usage, correcting mistakes,  filling in blanks, etc. We also ensure that the quality is as high as possible. The rate for the Proofread Transcript is $0.75 per minute of recorded audio. Therefore for a 1 hour file it will cost $45.

This gives you a bit more flexibility while ordering the transcript. The most time consuming part while transcribing an audio is typing; thats the heavy lifting part. In the Draft Transcript we do the heavy lifting for you and you can then review it and modify it as per your need.

We now offer only Proofread transcripts at $0.99 per minute of audio.

We have been fine-tuning our transcription process to improve the accuracy of the transcripts. Our process is a bit more involved from the usual transcription process where a person works on the complete file. Instead we split up the file into smaller parts and several transcribers and reviewers work on each part individually. The parts are then collated into the final transcript.

Our process tries to remove, to the extent possible, tediousness from the transcription work. Since files are small (around 6 minutes) a transcriber has to spend no more than 45 minutes to 1 hour on each file. Studies have shown that the average human concentration span is around that mark; a person can maintain the same concentration level on one task for no more than an hour, after which a break is required. Therefore splitting up files has an noticeable impact on the accuracy.

But this also introduces the problem of inconsistency. Since different parts are worked on by different people, a specific term which appears in one part may not appear the same way in the other parts. Our first approach for solving this problem was to use a dictionary. After the final transcript was done we used to cross check the file against a vocabulary of commonly occurring terms and correct the mistakes. But this approach only solved the problem partially since it was hard to catch all the inconsistencies.

The best we accuracy we could get to was around 95% with this process. But we wanted to improve it. So now, we have added another step of review. This final review is done by a single transcriber who goes through the entire file and corrects all the inconsistencies. We have been doing this since the past few months and have found this to be much more effective than the dictionary approach. Consequently our accuracy has gone up to 98 to 100%.

Because of the additional review step our costs have also gone up; from 50 cents per minute to 75 cents per minute. This rate has been in effect for around 3 months and we are quite pleased with the results. We are still around 25 to 50% cheaper than other professional transcription services available. That is the result of our transcription process and the system we have developed to manage it.

The increased price also gives a little bit flexibility. Therefore we have also started offering discounts for bulk orders. If you order for more than 10 hours at a time then you can avail a 10% discount on the total price. We will also initiate a seasonal discount program sometime in the future.

So try out our improved transcription service and order a transcript today!

Deliveries: The delivery slabs are now 3 days, 1 day and 12 hours instead of 7 days, 3 days and 1 day respectively. The rates remain the same.

Video Files: You can upload video files and have them transcribed. We support wmv, avi, flv, mpg, mp4 and mov files. The files are converted to mp3′s and then you can order transcripts.

URL Upload: Files can be uploaded by URL’s. Eg. YouTube links or podcast links.

Partial Transcripts: After ordering the transcript you can view partial transcripts. The option is in the actions menu in the transcripts page.

Order Cancellation: After the transcript order has been placed you can cancel the order and we’ll refund the applicable amount to you. The applicable amount is calculated from the transcript progress.

Tools: We have also added a slew of tools to modify/clean up the transcripts after they have delivered. The timestamps, speaker initials, blanks and all meta information can be removed. A simple find a replace tools has also been added. These options are available from the actions menu in the transcripts page. The transcript text can also be copied directly to the clipboard from as opposed to download the files.

Thank you for using Call Graph! Comments/feedback are most welcome.

Our our transcription subsytem has undergone a complete revamp. As a result there are several changes to it. The major one’s are:

We have transitioned to a complete manual process for transcription from an mixed automated and manual process. This means

• No more screening of the files
• Flat 0.50 per minute rate for all kinds files
• Improvements in quality of transcripts

Transcript orders can be placed from the uploaded records page. Earlier it was from the transcript page which used to list all the screened recordings.

The limit of 60 seconds on transcripts orders has also been removed. Transcript orders can now be placed for files of any length.

Rates

Do you have a fast typing speed? Are you comfortable with American, UK and Australian accents of English? Are you interested in audio transcription?

Our Freelance Transcription Program is ideally suited for you. Our transcription process is partly automated and partly manual and we are looking for help in the manual part. All you need is a computer, internet connection and a good quality headset. We will tell you which tools to use and the way to do it. We’ll take a test and approve you if it matches our quality requirements. Thereafter you can start earning with Call Graph!

Please click here to check the details and apply.