This is a rather different type of article from the normal. It is part of an occasional series in which I recommend a tool or service I’ve found particularly useful and which I think others might too.
We know that video is increasingly popular across the internet these days. Social media really champions video content above all else, and the second largest search engine in the world is YouTube.
The problem is though that increasingly people are watching video with their volume muted!
It makes sense – scrolling through whichever platform takes your fancy, with all the videos auto-playing, it does get very noisy. Especially if you are out and about. Or perhaps browsing where you shouldn’t really be 😉
So a very important part of creating video is to ensure they have subtitles or captions, so that viewers can still get the benefit of what you are saying even with the sound off. Most social media sites and many other laces (including YouTube, of course) offer the facility to upload your own subtitles (they use .srt files which is an industry standard for this).
Which just leaves the question of how on earth do you create those files in the first place?
There are three main ways this can be done –
- Do it yourself manually.
- Outsource to someone to do manually.
- Automated transcription.
I don’t know about you, but I hate #1. I’ve attempted it many times before, but it is pretty time consuming.
Apparently a good typist takes around an hour to transcribe 15 minutes, on a good day!.
Plus once you’ve typed it all up, you still have to convert it to the right format to create subtitles. Which gets even more time consuming.
#2 is a very popular option, and there are a lot of companies out there who offer full transcription services. You give them your audio, they give you a full transcript with timestamps so you can use it for your subtitles etc. The downside is that it can get quite costly – one well known company charged $1 per minute which doesn’t sound much, but for me that gets pricy – 4 x 25-30 minute radio show episodes per month is already $100 – $120 in transcription fees, before I take into account all my other videos.
#3 is a far lower cost option, but finding an automated service with good quality and high accuracy has always been the stumbling block.
However (you knew that was coming!) I have found something which really does work very well for me. It has good accuracy, a fast turn around, is easy to use, and is very cost effective. So I thought I’d share it with you here too.
It’s a service called Sonix, which offers fast, accurate, automated transcription.
I’ve been testing it out, and it really does do what it says!
I’ll give you a quick overview of the process.
You start by uploading your file (video or audio, it accepts the usual sorts of file). You can do this either directly from the website, or it integrates with DropBox, GoogleDrive and YouTube, which is pretty handy.
Once the file is uploaded, you confirm the language being spoken, and hit go.
After the transcription is complete (I recently transcribed a 30 minute track, which it did in under 15 minutes), you are offered the chance to check and edit if you wish.
This is where Sonix really shines. As you can see here, the page shows your text nicely transcribed, split into different speakers, and with time stamps.
It also has an “accuracy thermometer”, where words which it feels may have been less accurately translated are highlighted, making it very easy for you to go in and fix them if needed. You can play each paragraph, OR you can highlight a word and it plays from that word, making it even quicker to catch what was said and to ensure the transcription is accurate.
If you come across words with which it is having difficulty (uncommon or specialised words, for example), you can correct them and add them to its dictionary, which it uses for all future transcription – certainly a lot easier than creating special dictionaries as some other automated services require!
Other editing options include striking out or deleting some words if they are not needed, and the option to highlight a section and download the audio for that section is particularly useful for generating short clips.
Using the transcriptions
Once you are happy with it all, you are ready to download and use the transcript. Again here there is a lot of flexibility.
For normal text output, you can save it as a Microsoft Word doc file, as a plain text file, or as a PDF. Here, you can specify whether or not to show the speaker names and the time stamps for each paragraph.
For subtitling etc, you have the options of the normal .SRT and .VTT files, or you can export for Adobe Premier or Apple’s Final Cut Pro X. With the SRT for example, you can also specify how many maximum characters may appear on the screen at one time, and whether to use one or two lines. Great flexibility right there!
You also have the options to export the original audio in various ways.
There is another neat option – you can either embed it on your site, or just give a direct link to it on the Sonix servers. When you use either of these, then it provides the visitor with an audio player to play the sound file and at the same time as displaying the transcription, it highlights to words as they are spoken, so you follow along! This can be a pretty neat way of consuming the content and ensuring you do not get lost.
I’ve used Sonix to transcribe a few radio show episodes already, and I’m really pleased with the results. It is way better than the previous automated system I was using, and it does get an accurate transcription pretty quickly.
The fact that you can download the result in multiple different formats is a definite plus.
The only real downsides I found were
The only English option is US English. Slightly frustrating not to have UK English, but in fairness it does not seem to have affected the transcription much, so that might not be an issue after all;
The FCPX output, whilst easy to load into FCPX, did not seem very useable there, and there’s no real documentation covering it, so it’s rather less useful – unless I’m missing something, of course, which is entirely possible! I’ll carry on playing with it, and if I make any breakthroughs I’ll update this accordingly.
For those who need them, Sonix has some other bonus features tucked up its sleeves.
- It has Zapier support, so you can integrate it with thousands of other online apps – you could set it up so that you upload a video to YouTube, Sonix automatically transcribes it and emails you the transcription.
- It has a Chrome extension for those using that particular browser, making it easier to get your audio or video files in to Sonix.
- For developers it has an API, so you can access it directly from within your own software.
OK, How Much?
Obviously, pricing is an important consideration here, and so far I’ve fond Sonix to be very reasonably priced for what it does. The standard no-contract pricing is $10 per hour of material. Bear in mind the more popular manual transcription service out there cost $60 for an hour, so it’s already a lower cost. For example it works out around $20 per month for my radio show transcriptions instead of $120 per month!
There are other options also available for higher volume, which brings the cost down to $5 per hour with a $22 monthly fee for that level of service. Obviously you need to work out which one is more cost-effective. If you plan to transcribe more than 4 hours 25 minutes per month, the latter option is definitely cheaper.
Should you use Sonix?
Only you can answer that, of course.
What I can say is that it is far quicker than doing it all by hand yourself. It is certainly a lot cheaper than the manual services around.
It offers a lot of power and flexibility, in an easy-to-use way, so it works with you rather than against you.
All in all, I like it. So much so, that it is now my go-to transcription service.
Pop over to their site and have a look for yourself. There is a free 30 minute trial which gives you a flavour for how it works, too.
Disclosure – this page contains one or more affiliate links, which means if you buy after clicking on them, I will get rewarded. I hasten to add that is NOT why I am recommending this! I genuinely do find it very useful, but hey, they offer an affiliate link so I’d be daft not to use it. You are, of course, welcome to work round that and use a normal link instead of you prefer!
Award-winning coach, international speaker, multi-time best-selling author, hypnotist, occasional magician, and writer of this post, Keith spent his first 40 years suffering from several phobias including being terrified of speaking with strangers. After one incident too many, he started studying and training in NLP & hypnosis to conquer his own issues, found he was rather good at it, and changed careers (aided by redundancy at just the right moment after 20 years in IT). He helps people transform their deepest fears into their greatest strengths, and having helped over 5,000 people across 5 continents, he is the UK’s #1 Fear Strategist.