Speaker 1: Hello everyone and welcome to Una Cosa Menos, the video podcast in which we talk about tools and productivity for your business. Today we are going to talk about transcriptions for podcasting. Well, as you already know, if you have been following me for a long time or if you have consumed my content such as the Business and WordPress podcasts, I love audio. I've done radio, I've done speeches, I do podcasts, of course, and since I like technology, I've always been interested in transcriptions. There is a technology that allows you to transform audio into text. For years now, there have been algorithms, tools, whatever you want to call it, to do this task, to convert any type of voice message or recorded audio with the voice, I mean, a spoken message, which transforms it into written text. And now, with the arrival, in quotation marks, of the hype there is with artificial intelligence, tools like Zetas are coming out. So, well, I'm going to tell you a little bit about the different strategies that we can use to do this task. And in fact, I've been collecting different tools, services, etc. for a long time, because there are actually a lot. And I'm going to tell you about the different types or methods that exist to be able to do this. Well, the most artisanal way to do it is by talking directly to the computer, dictating it, we could say, right? For example, Google Docs itself has a section to write with the voice. Well, just like our mobile keyboard, this has existed for a long time, so it seems strange to say it, right? But hey, we've had this for a long time in tools like Google Docs or Dictanote. And I even remember that IBM had an engine called Watson, in which you could talk and the text recognized you. I think it also admitted an MP3, but what it did was play it as in the background and it was transcribing it in real time, okay? It wasn't like today's engines that process it at full speed. Here, what we can do is install a virtual sound card and tell it that the reproduction, for example, of VLC, I'm going to say, the multimedia player with the file that we want to transcribe, and that the dictated tool, its input device, is that virtual card. Then, directly, the audio is going to be transmitted from one application to another and the transcription will be done. But of course, we have to install an additional software. We have to wait for that file to be reproduced in real time. Also, for example, Google Docs, if you change tabs, it stops working because it detects that you're not talking to that tab. In the case of Watson from IBM, you need to choose the language every time. Well, it's a bit of a mess, all this kind of tools, which are also totally obsolete today. Another way that can also be used for a long time is with YouTube, because YouTube has been transcribing automatically for years, of all its videos, the automatic YouTube subtitles. And what we can do is either generate a video with our audio, which there are also ways to do it very quickly. I'm also thinking of doing an episode about this. And if not, then a video, record a video as such. If we already have a video or if we can consider the option of generating that content in video, as I am doing now, which could perfectly be an audio podcast, but I also want it to have a video so that you can see me and sometimes I put accessory content so that the explanation is better understood. Well, YouTube is going to give you a transcription for free and automatically. Of course, you have to wait for it to be transcribed, etc., etc. It can be a bit of a cumbersome process, but hey, it's free. And above all, if you're going to generate that content in video, I think it's a very good option. Let's go with another option that I think would be almost, almost, almost my favorite today. And they are the Telegram bots. Until recently, I used a bot called Boise, but it seems that the developer is involved in many projects and has sold it, has put it on sale. Well, you have to pay, if you want to access it, I think it's worth $6. You have access to consume the resources of the server of this developer. But hey, the truth is that it worked very well and I was sorry that I had to stop using it, so I looked for an alternative. It's called Transcriber Bot and it admits both audio files and voice messages, that is, what is audio, as well as images. It also allows you to extract text from images. I think this can be done in many ways, but hey, it's okay to have it in the same bot. The good thing about this is that it is super easy to use, super flexible, because we can re-send audios that have been sent to us. I sometimes select WhatsApp audios, I share it and send it to the bot and then I can read it. Well, it's great. It can also be added, of course, in chat groups, in Telegram groups, let's go. And if someone sends an audio message, it is automatically transcribed. So it's great. I already tell you, I don't know what limitations it will have, but it works quite well. For example, Voicy allowed you to choose the transcription engine, but the truth is that I don't have problems and the truth is that it usually understands everything that is said quite well. Especially if it's an audio, like in a podcast where you try to speak in a way that is understood, etc. Maybe an audio message to your colleague doesn't transcribe it either, but it works wonderfully, free, easy on Telegram, so it's perfect. And we go with the dedicated services. Online services that we can use from our browser normally, although perhaps some have applications to transcribe a video directly, to transcribe an audio or a video directly, because this is also getting very fashionable. The fact of transcribing the video directly, that instead of giving an audio, we give it a video. I recently discovered one called freesubtitles.ai and I thought it was great because it was very easy to use, it was on a website, it was free, it admitted a lot of file length or weight, but I've tried it and it doesn't work. Also free, we only have up to 300 megabytes or one hour of length, which for most cases of use I think is enough, maybe for business podcasts and Wordpress, because we always go a little over the hour. But we also have payment plans, a payment plan where we can upload files up to 10 hours or 10 gigabytes in file size. And it also allows you to paste a link, for example a YouTube link or a link to an mp3. What's more, it's multilingual, you can even tell it to transcribe several languages at the same time, selecting an option. This is paid, of course. But I've tried it and it didn't work for me. I uploaded a file of something less than 6 minutes or 7 minutes, it left it running for hours and it didn't work for me. So, well, I don't know, it seemed pretty good, but maybe it's saturated. I recommend you take a look at it to see if it works for you, because I've seen it recommended in several sites and the truth is that it looks good. By the way, I am now commenting on the services that are free or that have a generous free part. The next one is Auphonic. It is a tool that I use more than to process the audio, to distribute, but it is mainly known because it allows you to process the audio automatically. You can even do automations with it and so on. It is usually used in podcasting to remove noise, to level volumes. I, well, as I do it very quickly in Audacity, I would not take advantage of that. And the good thing is that it is compatible with an engine, or with several actually, of transcription within the service itself. You choose that in addition to processing the audio, it transcribes it for you, it processes it for you to give you a file of subtitles. You don't have to pay to use the transcription models, although the application only allows you to transcribe two hours a month, well, to process two hours of audio a month. It has several payment plans and the cheapest is 9 hours a month for 10 euros a month. I think it is one of the cheapest payment options, so it is an option to take into account. And now we go with the payment options. Maybe I will tell you some that catch my attention for something. For example, the cheapest one is called Transcribo. It gives you 30 minutes for free, which is not bad at all. And if you want to add more hours of transcription, it costs 5 dollars per hour. We are already seeing that it would be much less difficult for us to do it with Auphonic. But hey, this is a dedicated transcription application and it is possible that it does it better. Because in Auphonic it is 9 hours for 10 euros. It is almost 1 euro per hour, much cheaper. Then we have AmberScript from 10 euros per hour. HappyScribe, which is well known because it has also had a lifetime offer in AppSumo, from 12 euros per hour. Then we also have, for example, Otter, which works very well for online meetings. There are many services of these that work, or only for video conference meetings in Google Meet, in Microsoft Teams, etc. And some that do normal transcriptions and also this topic of video conferences. For example, in Otter you can import files and have them transcribed directly to you in the service. If I'm not mistaken, it lets you make three file imports per month with the payment account, yes. We also have Convert Speech, which is one that I used quite a bit before. At first it was unlimited for free and now they only give you, let's say, 5 minutes for free. That is, the first 5 minutes of each audio, which, well, in some cases can be useful. And the fact is that I have been getting to know many services that little by little have been limiting, which is logical too, and that makes you have to look for alternatives or pay for one of these services, let's say premium, which simply by paying you have no limit because you pay per hour or because it includes many hours per month. As is the case of Descript, which is a tool that I am loving, it is where I am recording this video right now, and it is to freak out. Because only the software, as a video editor, is the milk. It does transcription, it is handled in a very simple way, it has many possibilities, you can extract videos for social networks, generate subtitles, it is in the cloud, the backups are made automatically. Now that I am working on several computers at the same time, what I am recording is being automatically uploaded to the cloud, to the Descript servers, and when I get home I can continue editing, opening the application and acquiring the project, the resources are downloaded automatically. Well, it's great, but let's focus on the part of the transcriptions. In the free plan we have one hour a month, which is more than what some other programs give you, but if we pay from $15 a month, we will have with this plan, which is the first of the premium plans, 10 hours a month. We have seen that perhaps at the level of transcription it is not the cheapest, but the vast majority of services are more expensive than Descript, and also with Descript we have all the functionality of video and audio editing. By the way, because for audio it is also great, I also have to make a Descript episode, because we can mark chapters in a podcast, we can export directly in a lot of formats, we can also export the transcription, which obviously generates us in text format. Well, it's great. As you can see, I really like Descript, I love it, I'm really enjoying it. And there is also an offer, by the way, in Secret, with which you can get 6 months for $50, although I took it in Black Friday and it cost me $25, but well, you have a discount code in Negocios y Wordpress if you want, and the truth is that I love it and I will surely make an episode. But I don't want to leave without mentioning two more tools. One is Mac Whisper, and one of the things I wanted to comment on in this episode is that I was going to do a live broadcast or something like that, trying to install the local Whisper, because OpenAI passed it to open source and people can install it on their computer or have it in an instance in the cloud, or whatever, I said, this would be great to have free transcriptions. Well, it turns out that a guy has packaged it in a free Mac application, although it has a payment option, which is called Mac Whisper, which works like a charm, which is putting a lot of features in it, and then I don't need to do that live anymore. So, totally recommended, Mac Whisper. I discovered it, by the way, in Emilcar's weekly podcast, and also thanks to Emilcar I discovered the other application that I wanted to explain to you today. It's called suel.ai, and it's an online service that is dedicated or designed precisely for this, to make podcasting transcriptions, but it allows you to upload the episodes with YouTube or Dropbox, with Drive, with, of course, directly uploading the file. Come on, you can do it through a URL or with a file from your computer. And it automatically gives you the transcription, but not only that, it generates the key points of time, as if they were the episodes of the podcast, to put them in the description, for example, in YouTube, and that you can go directly. It generates summaries, like a couple of paragraphs or three of what the episode is about, and it even generates possible titles for your episode, that are striking or that are a little different, right? It also has a section to make a specific request, but well, I think that with these texts that the tool already generates for us, well, it's great. And the other thing it generates for us is an adaptation of the transcription in the form of an article. It rewrites the content for us, but instead of, as we said, it's written to put it, for example, in an article of a blog. The bad thing is that it only allows us to upload a file per month for free. It doesn't say limitations, but well, it only allows us to process an episode. If we want to process more, we have to pay a minimum of $29 per month, which allows us to upload five episodes. If it's for a podcast, I think it can be enough. For a weekly podcast, for example, but if you want to transcribe a lot of content, videos, etc., well, it wouldn't work for us and we would have to go to the next plan, which costs $49 per month. There is also the option to pay $9 for each file. I think that if it works as it seems to work, because I have uploaded an episode and it has worked, it has done everything quite well, although it answered me in English everything. So I don't know if in the future they will put something to force it in English. As it seems, Emilcar has been trying and giving it to generate another summary. In other words, you can force it to generate new versions of each of the texts and he generated them, apparently, already in Spanish. So maybe they have to polish this, but the service seems great to me. $30 per month for whatever podcast seems too much to me. But hey, if not, you already have the rest of the options, like the script, like the Telegram bot, etc., etc. Well, well, and before we finish, let's go with one last tool that since I recorded until I'm editing, as they say, I'm the Elias of the future, has appeared from the hands of Spreaker. It's called Free Podcast Transcription and as its name indicates, it allows us to transcribe our podcast for free. It is a kind of assistant in which we first choose the language, then the file, then we give it to transcribe. It took me a few minutes to transcribe an episode of about 6 minutes long. A step appears in which it allows us to edit the transcription. They come out as independent sentences of the transcription. We can edit them and save them. And the last step is to publish them. It forces us, in quotes, despite being something theoretically private, to upload it to their servers to be able to copy the URL, embed it in our feed, in the podcast transcript label, and it also offers us a button to download it in SRT format. The most interesting thing of all is that everything is supposed to be produced directly in our device, in our browser, so the file is not sent to any server or anything. So nothing, I've already told you about the audio transcriptions for podcasting or whatever you want. One less thing. Remember that you can listen to all the episodes at eliasgomez.pro podcast and you can find me on Twitter as EliasNS. Until next time.