Quick Review of Temi – Inexpensive, Automated Video Transcription

Videos are a powerful way to share your message, especially now that many people have Smartphones. But if you’re not also providing text transcriptions, you’re missing out on the SEO benefits and an opportunity to help the hard of hearing.

Watch this quick review of Temi, an automated video transcription service made by the same team behind Rev.com (audio transcription). It’s very affordable for even occassional bloggers and will help to maximise your content across different media.

I’ve included the transcription of this video below, provided by Temi.com (total cost: $1!)

This should help you decide if the quality of machine transcription is suitable for your needs.

Video content is a powerful way to engage your audience and share what you have to say, but if you’re not providing transcriptions at the same time, you’re missing out on a lot of value for SEO and, of course, helping those people who might have accessibility needs.

So today I want to share with you a quick and inexpensive solution to automatically provide transcriptions for the video content that you create. This frees you up to do the things that you enjoy, like make more videos! Let’s have a look at Temi.

So, let’s test this out and upload our own video file to Temi.com to see just how good and accurate and reliable the transcription files are.

So first of all, we have all these checkboxes that by default, these are on no anyway. Is the background noise free? Yes. Are the speakers close to the mic? Yes. Is conversation clear without overtalk yes. Do the speakers have clear accents? I think I’ve a pretty clear accent. So Click on. Yes, we can not show this again, but it’s useful I think at the beginning to set the expectations on what is a good upload file and what is a bad one. Let’s click on proceed. So here’s a video that I recorded before I’m going to upload this. It’s about 16, 17 meg or something like that. So it might take a while.

Now it’s uploading. We’ve got a nice upload bar there. I like that. And the great thing about these transcriptions are with it being only ten cents a minute, it’s cheap enough to do an automatic transcription. This is machine led. It’s not manual, there’s no person behind this listening and typing in the transcription, but it’s cheap enough to be able to do it for all your videos and put those transcriptions on youtube to improve the usability for people who might want to listen. Um, and also to presumably improve things like SEO performance or where it’s positioned in the search results. If those factor in things like, you know, does it have a transcription?

We put our email in here. This should take, I don’t know, five minutes or something. So okay, so here we’ve received the email from Temi telling me that my video has been transcribed. This is the free test, so this is the very first time that you upload a video. You will be able to do it for free and after that it will charge you ten cents. But I thought we’d use our free trial to test out Temi here and see what the quality is like.

It tells us that our video is just over five minutes long. The cost was free in this case. Um, and we can view the transcript and it took two minutes. It’s really fast I think. So let’s view the transcript. The first time you do this it will ask you to set a password. So I’ll do this now and we’ll just jump over to checking out the transcript.

So we’ve set the password and we’ve got ourselves an with Temi for any future transcription we want to do. But for now let’s have a look at the online revision editor. They’ve done the transcription and here we can check on the quality of it. We can see what they’ve got, right, what they got wrong, we can make some changes. Um, and I’ll, I’ll lead you through this as I explore it here now. Over to the right hand side, we seem to have the video itself and we can play this. We can change the speed, we can change the volume. We can jump back five seconds and we can write some notes. And over to the left, this is the main editor. Just close that for a second.

They correctly identified me as speaker one, although some point here they’ve also put speak two. So, I don’t know if I changed my voice. I don’t think I did.

I don’t know what these words are in red, so that was going to be very interesting to see. Maybe that’s words that they’ve had to guess for. We can do a search and replace, so if we find something that’s really bad we can change that all across the board. We can highlight something which might be useful when we download this as a word document or PDF or anything like that. We can strike through something as well to show that for whatever reason, something he’s taking out. Read along here, I think when we play this, I think what happens is the cursor follows the text within the editor, and we can turn that on or off. I think that’s really good. We have some shortcuts here. So for example, play pause tab, that’s going to come in and really useful as well.

Um, what else can we do? Strike through and highlights so we’re not playing around with the mouse too much. I like that. And then finally we can reset it. So I guess any changes that we’ve made here, we will be able to reset that. I can’t see us doing that because that would reset a whole lot of work. Quality is on. So let’s see, the audio quality is high. So Ah, here we go. Anything marked in kind of a brown color is at 19 percent. Low confidence phrases and these are phrases that Temi had a hard time with will be colored Brown. Fair enough.

So let’s see how we did. Let’s put the video back to the beginning and press play.

So today we’re going to install Thrive Architect on a brand new wordpress installation. Thrive Architect is offered by these guys, Thrive Themes, who I’ve used their excellent plugin.

So here our very first, unknown. It got thrive correct here. And it got thrive correct here. But the third time I said thrive, it’s thought I said throwing. So I think is a good idea at this point to take that and find and replace. Search for throwing and change it to thrive. Replace all. Now I wish it told me how many of those it had replaced and maybe it did. I wasn’t paying attention here. So yeah, let’s take thrive in small letters and change it to capital t and replace all. Great. And come out of that. so I can already see how this is going to work. Um, as I click in here, it actually changes the video, so I’m going to click there.

It’s a visual editor for wordpress.

Great. So this is really accurate so far. I’m wondering how it chooses where to put the paragraphs, I think it would normally break them up by speaker, but I mean, so for example, here we’ve got this little sentence on it’s own. I just don’t know why, but that’s fine. I mean we can probably go here and yeah, I can just hit delete on that, which is great. Uh, there are no speaker twos so I’m going to put this on his own line. And Speaker two, like, that’s great. So he thinks that there’s a new speaker there, but I don’t know why.

So I’m going to run through this now and make a few changes and speed up the video and then we’ll see what the exports are like.

Okay. Now let’s have a listen to this because this is an interesting issue that I’ve picked up with Temi. It’s not huge because because this is automated, it’s obviously going to make some guesses as to the sentence structure and cadence, but just listen to this and read at the same time

[Audio too quiet to transcribe “Here we have a vanilla website, a wordpress installation”]

So this was meant to be “here we have a vanilla website, a wordpress installation” and it’s put the word “here” as the final word of the previous sentence with a full stop. So these are the kind of errors you’re going to run into quite often I imagine when you’re using Temi because it is machine lead and is not manual. If you’re okay with that, and I am for the sake of making sure that there is a decent, half decent transcription available on Youtube, so if you’re okay with that, then this service might be perfect for you. Otherwise you’re going to need to spend a little bit more time making sure the revisions are in place to get it as good as it can be.

Right? So I’ve done a little bit of messing around here. I’m not going to do too much because it is a test and let’s see what the export options are like. You can download it and we can choose Microsoft Word, PDF or plain text. We can include the speaker names, have timestamps or we can just highlight certain sections and export those, but we’re going to download the whole thing. Microsoft word with the speaker names and the timestamps. That sounds useful.

Here’s the download transcription. Enable editing at the top and we can click, we can control-click on any of these timestamps to jump back into the editor that we’ve just been using at the correct place, which might be useful. We can also share, so it looks like here we can either email somebody or we can get a shareable link. So let’s try this, copy that. And we’re going to go into, um, an incognito window so we’re not logged in. Great. So anybody will be able to access those, but that means it also be able to download it, I think. Yes. So anybody who has access to that link can download this.