Open Forum

Hello, all. I am working on getting our veteran interviews transcribed and wanted to know if anyone has any recommendations for safe, reliable, and affordable video-to-text transcription software. I am currently using Microsoft Word to transcribe the videos through the "Voice" function, and then going through the interview recording to correct any errors to the AI transcription. While Word does a fairly decent job, the service is limited to 300 minutes of uploaded audio a month, which are going to get eaten fast given that our museum has over a hundred hours worth of interviews that need to be transcribed. Are there any services that allow unlimited transcription work? Any suggestions would be greatly appreciated. 

Keith,

While it is not free, we use both Temi.com and Rev.com for audio/video to transcript and caption services. Temi is done through AI and cost .25/min, Rev is done by humans and cost .99/min. In either case you can then review and edit the text in their interface.

John ffrench
Director of Visual Resources
Yale University Art Gallery

Keith,

You may like what Descript does for you. You can transcribe and edit audio and video. You can edit the transcript like a Word document and changes will automatically be made to the audio or video. Go to https://www.descript.com/

------------------------------
Tania Marien

The Freelance Project | TALATERRA
Contributor, The Carbon Almanac
------------------------------

I like Otter.ai. They were the cheapest, good option, but they recently raised their pricing. I'm not sure how competitive their new pricing is as I'm grandfathered in, but I like them. You can try them for free. 

If you have the Adobe Creative Suite around somewhere, Premiere does automated transcription of everything you load into it, and it's pretty good. Advantages: you don't have to upload anything to the cloud, it's unlimited so effectively free if you already have Premiere. It is integrated with video editing, so everything you edit from the original is also transcribed as edited. You can improve the audio quality and re-transcribe if the original audio is weak, or if there are tracks you need to eliminate, and so on. Disadvantages: manual editing of the text is a bit clumsy compared to the cloud service, exporting is limited, and it's not shared automatically. You can export the transcripts, but only in text format, with no control over the formatting. Functional but maybe not ideal. I still use Otter a lot for bulk transcription of material I need to share with others prior to editing.

If the audio is bad and you are not getting a good transcript, try AI noise reduction. You can use very aggressive noise reduction to make bad audio more easily transcribed even if sounds awful to your ear. I like AudoStudio.com but there are many. 

Finally, re-reading can be a very effective problem solver. Essentially, you use any voice-to-text trained to your voice, listen to the recording and repeat to your voice-to-text, rather than typing. It's a bit tricky and slower, but can be a lifesaver for heavy accents or highly technical originals. It allows you to correct as you go, and even train the software. The brain is still smarter than AI.

------------------------------
Tod Hopkins
Hoptod LLC
443-472-5978
todhopkins@me.com
------------------------------

Have you tried uploading to Youtube? I know they offer a free transcription of videos. I'm not sure if you need to make the video public to get the transcription or not.... It could be worth a try.

I just saw yesterday that Audio Hijack, a well-respected MacOS audio capture software, has added a transcription component powered by OpenAI's model that doesn't have minute limits or anything. You'd have to convert the videos to an audio only format to run them through I think, but what I was reading was talking about it putting a major dent in Otter.AI's business model. While the software is $60 or so, if you’re doing bulk conversion it might be cheaper. I was planning to play around with it soon, but here's a link to a short write-up. If you're open to Mac, might be something worth checking out.

------------------------------
Drew Whatley
Museum Educator
Whatcom Museum
Bellingham WA
------------------------------

Hi Keith - sounds like there are a lot of good options already mentioned, but also wanted to add that if you find yourself needing translation as well (in addition to transcription and/or subtitles), I've had luck with Fiverr. The person we worked with there was quick, accurate, and affordable for Spanish/English translation and subtitles.

Rev.com is good. You might also try https://gotranscript.com/. I've found them to have a faster turnaround than Rev. (They'll quote 3 to 5 days and get the transcript back in less than 24 hours often). They're also a bit less expensive. 

Keith, 

We use both Adobe Premiere's internal transcription system and AWS Transcribe on a batch processing basis. AWS is a pay for the actual processing you undertake model. Both work well, though both too have challenges with particular speakers and dialectical differences in pronunciation. Both do require you to do post-processing clean up and transcription document formatting, so it may make sense to outsource the work. We have found this to be especially the case if the AIs seem to perform poorly with a particular speaker. 

Cheers,

Chris 

We have been using this site: https://speechnotes.co. It has been working nicely for us, and it's only $0.01 per minute and with a minimum purchase of 45 Minutes ($4.50) 

Keith,

    If you have a subscription to adobe you can download Premiere and it will do it within seconds to minutes depending on how long it is.  If you need help with your project I'd be more than willing to help out.  Premiere does a good job at creating the transcription but will need to be reviewed for verification.  It would save you a lot of time for sure. I'm a vet myself so I'd love to help just give me a call and we can talk.  Not sure what your deadline is for them.

Hi Keith,

We, like others on this thread, use Rev.com but the cost does add up. If you need a free solution, we have used Youtube. You upload the video, save it to Private, and in about 24 hours, YouTube will have created captions that you can download as a vtt or svt file. In the subtitles tab, select the 3 dots next to published. YT does a decent job. Never perfect but will save you some time. 

Hi, Keith,

We at Nubart are very satisfied with HappyScribe (https://www.happyscribe.com/pricing). We work on digital audio guides, so we use it quite often.

It's not completely free, but it's very affordable. They have a free plan to test it first. They make it very easy to edit mistakes.

Just to clarify, we don't have any business ties with HappyScribe – this recommendation comes straight from our positive experience with the service.

Wishing you the best of luck with your transcriptions!

You can try using https://youtubetotranscript.net/ for unlimited transcription. Upload your videos to YouTube as private or unlisted, then paste the video URL into the tool. It will automatically generate and display the transcript, which you can edit later. It's a free and scalable solution for transcribing long interviews without the limits of paid services.