[OBSOLETE -- New Video in Description] Transcriptive Tutorial - Requesting New Transcripts
**UPDATE** NEW Version of this video at: • [OBSOLETE -- see description for new ...
Check out our updated Tutorial page here: https://digitalanarchy.com/transcribe...
Transcriptive - Requesting Transcripts
In this tutorial we go over how to request a transcript for an active Sequence and how to use the "Batch" transcription method to request transcripts for multiple files at once.
Learn more about Digital Anarchy & Transcriptive here:
Try the Transcriptive FREE demo:
http://ow.ly/A8Ey30gygtX
Truncated Transcriptive transcript:
Hi, and welcome to another Digital Anarchy tutorial. I'm Tor Olson, software Q.A. and tech support, over here at Digital Anarchy and in this tutorial we're going to be going over Transcriptive. And more specifically, how to call for her transcriptions once you've logged into your speech engine.
First, you need an active sequence, and luckily, I have one down here of goofy little video that Jim and I recorded about the mascot's of Digital Anarchy.
And when you are calling for transcription for a single sequence which we're about do you want to make sure you do it with your active sequence. Make sure it's your active sequence by looking in the list of sequences, and your active sequence is the one that is underlined.
So I know I want to call for transcription for this sequence.
So I would go ahead to click "transcribe".
So from here I'm opening the transcribe dialogue for a single sequence. It'll show you the sequence name. It'll show you how long that sequence is.
So, in other words, how much are you going to be paying for. Here we can see how many credits we'll be using since we're charging this to Speechmatics.
You can select either speech engine by using the "Transcribe With" dropdown. So in this case we have Speechmatics. The information with the engine we're using is shown here in terms of: the credits we got, the minutes we can use, how the server on Speechmatics end is doing, and how heavy a load is on right now.
So since there is zero minutes, we expect that Speechmatics will come back in about a fifth of the time that it takes for our sequence to play back in real time. And that speed estimate goes into both Speechmatics and Watson. In that amount of time for a transfer to get back to you is what we expect for both speech engines.Speechmatics is a little faster, but only incrementally so.
If I were to choose Watson that information will change a bit. It'll show you the total minutes that I have logged for the current month. So this renews every month of course. So you get a thousand free minutes at the beginning of each month.
For now we've only used about 44 minutes here. So our balance due is going to be zero once this goes above 1000, you'll start to see the ticks on your account. So just below where we can like her speech engine.
You have the languages that each speech service supports. Each has a unique list of different languages that it supports, with Speechmatics having a little bit longer of a list. You also have speaker diarization.
So, if you say, have a sit down interview and you know that you only have two speakers. You can make sure that you limit the speaker identification to two speakers. And when you ask for your transcript, you can later take the speaker identification tweak it if you want.
You can change the name so they can be really useful tool for searching your transcripts maybe looking for one particular speaker, or the other. Transcriptive transcripts by default and come back with every sentence ending in a period.
So by checking the toggle for identify questions will look for sentences that start with: what, why, or certain sentence patterns and try to associate question marks with those questions.
We have an option to align current text, but this is for a different tutorial that will go over later. So once we've chosen the speech service that we want to use, in this case I will use Speechmatics, I will go ahead and click continue, and that'll encode all the audio that we have down here, and turn into a raw audio file. Send it off to Speechmatics. Speechmatics will then create a transcript, and return it back to us.
Alright, so it took about 25 seconds, or so, we have a diarized transcript. Where each individual word has its specific timecodes so we can click, and you'll notice that the play head will skip to that portion of the video.
Now, there are different ways that you can transcribe footage. The method that we use just by hitting the transcribe button up here, is to transcribe a single active of sequence. But what we can also do is batch transcribe multiple clips for audio at the same time...
Смотрите видео [OBSOLETE -- New Video in Description] Transcriptive Tutorial - Requesting New Transcripts онлайн без регистрации, длительностью часов минут секунд в хорошем качестве. Это видео добавил пользователь Digital Anarchy 08 Сентябрь 2017, не забудьте поделиться им ссылкой с друзьями и знакомыми, на нашем сайте его посмотрели 1,165 раз и оно понравилось 3 людям.