Speaker 1: Hello, and welcome to another tutorial for Digital Anarchy's Transcriptive. My name is Evan and I provide software QA and tech support here at Digital Anarchy. In our last tutorial, Tor Olson showed you how to set up your Speechmatics and Watson accounts. Today, I'm going to show you the two different ways one can request a transcript from these two speech services using Transcriptive. So let's jump into it. The first method is by using this big blue transcribe button in the Transcriptive panel. First, be sure to have your desired sequence active. Here I've made a sequence using the clip Jason Take 3. I know that the sequence is active due to the highlighted name and the underline beneath it. So, I'll go ahead and press the transcribe button. Go ahead and jump in and make this bigger. And now we have our transcript request dialogue. At the top, we have our sequence name and sequence length. We can see Jason Take 3 and that our sequence length is a minute and 30 seconds. Speechmatics charges seven cents a minute or six credits a minute. Further down, we have the option to choose which speech engine we would like to use. For Speechmatics, we can see our total available credits in our account, the status of Speechmatics' servers. Here, it's showing good or that the servers aren't under a heavy load. And because of this, there will be zero delay in executing our request. If we change our speech engine to Watson, the information changes slightly. You'll see our total minutes used for the month. Remember, Watson gives you 1,000 free minutes a month. And because we haven't used up our 1,000 minutes, our balance is zero. After 1,000 minutes, Watson charges two cents a minute. Next, we have the option to choose which language we are transcribing. Note that Speechmatics has roughly double the options as Watson. Next, we have speaker detection. If you know that you shot an interview with only two people, you can max out how many speakers the speech engine is attempting to identify. This comes in handy in the future with transcript formatting and searching through your transcript. Below that, we have identify questions. As a default, Transcriptive returns everything as a sentence or everything ending in a period. So by choosing identify questions, the speech engine will attempt to identify certain words such as who, what, where, why and how, as well as specific sentence patterns that would denote a question and use a question mark rather than a period. Next, we have align current text. This is for syncing existing text transcripts and we'll be going over that in a future tutorial. Now that we've gone through this dialog box, we can go ahead and press continue. Premiere will now encode the audio from your sequence into a lossless FLAC audio file and send it to Speechmatics, where Speechmatics will then create a transcript and send it back to us. OK, the 90 second clip took roughly 20 seconds or so to return, and now we have a transcript where each individual word has its own unique timecode, so we can click on any word and our playhead will jump to the corresponding time in our sequence. At this point, I'd like to give a warning. When you've requested a transcript for a single sequence, the transcript is only linked to that specific sequence. It is not linked to your individual clip in that sequence. As for now, I can go through my transcript, make changes, and the transcript will save, but it saves to the sequence, not the clip. If I take this clip to any other sequence, the transcript will not follow. In order for the transcript to link to your clip, you'll need to export to Speech Analysis. This burns the transcript into your clip's metadata. So let's say, for example, that I've accidentally requested a new transcript and discarded the current transcript without having exported to Speech Analysis. If I go to Load Transcript, nothing will happen. That's because I did not have my Speech Analysis metadata. When you request a new transcript and discard what's currently in the Transcriptive Panel, you're erasing the link between what's currently in the Transcriptive Panel and your sequence. If you find yourself in this situation, don't worry. You can still import the transcript from Speechmatics by selecting Import Transcript and choosing Speechmatics. Or if you used Watson, you can find the corresponding JSON file in your Transcriptive folder on your computer. And unfortunately, any changes you may have made will be gone. So as a precaution, if you're requesting a transcript via a single active sequence, always export your transcripts to your clip's metadata by selecting Export and then choosing Speech Analysis and finally Export. So we've successfully requested a transcript for a single sequence. But what if you have multiple clips to transcribe? Well, we can go over to the Transcriptive drop-down menu. Here you can see our batch options. It's all pretty straightforward from here. If you click on Batch Files, your File Explorer will open or finder if you're on a Mac. From here, you can choose as many files as you'd like to be transcribed. I'll go ahead and select these two clips from Alexia. From here, Media Encoder will open up and batch encode your file's audio and send them off to either SpeechMax or Watson depending on what you choose. We are now prompted by the same Transcript Request dialog box where you can see that we show the total accumulated time of the selected clips to be transcribed. So let's go ahead and send this off. This time we'll go ahead and use Watson. This being around 12 minutes in total length, it'll probably take a few minutes. Notice up here in the status bar, we can see how far along Watson is in transcribing our files. And while we're waiting for these files to be transcribed, we can go over the other methods for batch transcribing. If we go back to the Transcriptive drop-down menu, below Batch File, you'll see Batch Folder. This is exactly as it implies. If you click here, File Explorer or Finder will open up and you'd simply choose the folder containing the files you wish to transcribe. I'll go ahead and cancel out. The last method of batch transcribing would be to choose Batch Project. Unlike the previous batch methods, this option pulls files from within your current project. To do this, you must create a bin in your project and name it Transcriptive with a capital T. You'll then move the clips you want transcribed into the bin. Go to the Transcriptive drop-down menu and select Batch Project. Media Encoder will then pull the clips from that bin, encode them, and you'll be prompted again by the Transcript Request dialog box to send them to your choice of either Speechmatics or Watson. Alright, so you'll now see that our status bar indicates Batch Successful. That took roughly a little over, almost 10 minutes. I will say that Speechmatics is faster than Watson, but only by a little. So let's go ahead and pull those recently transcribed clips into our project. Now what's great about batch transcribing is that your file's speech analysis metadata gets automatically burned into your clip. You can see here our two clips both have their individual metadata. If you want to populate your Transcriptive panel with this transcript, simply make a new sequence with one of these clips. There we go, we have our new sequence. Now let's go over to our Transcriptive panel, and I'm going to go to the Transcriptive drop-down menu and select Load Transcript. Here you'll see that the transcript is now automatically pulled from our speech analysis into the Transcriptive panel. And once again, we have our unique timecode connected to each word that will allow us to move through our clip. So now that we have our clip in our sequence with our metadata, we can make a few edits and we can now have our Transcriptive panel reflect that of our edit. Let's go ahead and we're going to request a new transcript, discard what's currently in the Transcriptive panel, and then we're going to Load Transcript, and this will pull from our active sequence to reflect the edits we made. You can see here now at the three minute mark is the end of the clip. Now we can do this with multiple clips as well. If we pull in the other Alexia clip and make a few edits, and then go up to the Transcriptive panel and request a new transcript, discard what is currently in the panel, and load a new transcript, the Transcriptive panel will now reflect the edits we have made. So if we scroll down here to the bottom, click Organizations, there we are in the middle of the second clip, same with that, and here with Intimidating. And there you have it. The two ways one can request a transcript using Transcriptive, the single sequence method as well as the batch method. So if you want to try Transcriptive out for yourself, you can go to digitalanarchy.com where you can download a demo. While you're there, you can also check out some of our other special effects plugins. We have Beauty Box, Clicker Free, and Samurai Sharpen. And yeah, that's it for me. Again, my name's Evan, and I'm software QA and tech support here at Digital Anarchy. And I'll see you in the next tutorial.