Translation: Google’s epic tool that translates what we say by imitating the way we speak

In its blog, Google presents Translation, its new integral model of translation from voice to voice. That the company has been perfecting its translation models for years is not something new, but that these models are capable of imitating people’s voices is.

Google points out that the main objective of this is to help people who speak different languages to communicate with each other. To achieve this new voice-to-speech system, they propose a unique sequence-to-sequence model, which moves away from cascading systems and improves, according to Google, speed, composition errors and the translation itself.

Imitating accents and pronunciation

Google tells us that Translation is based on an end-to-end model, superior to traditional cascade systems. With this, they intend to demonstrate that it is possible to translate speech from one language to another without the need for an intermediate representation of text in either of the two languages, something that cascading systems do require.

Google’s new tool takes the source spectograms and directly generates other spectograms with the content translated into the desired language. To do this, it uses a neural Vocoder, in charge of giving the desired shape to the waves of the output spectogram. They also use an encoder capable of preserving the characteristics of the voice being recorded.

The main novelty of Translation is that it does not work in cascade, and that it adds elements such as an encoder capable of retaining the characteristics of the speech of the recorded voice.

When training Translation, Google uses a multi-tasking objective that seeks to predict source and target transcripts, while simultaneously generating final spectograms.

In short, Google registers the voice of the speaker, manages to preserve the characteristics of his speech, and manages to generate an output spectogram translated into the target language, maintaining these characteristics of speech.

Emulating natural language

Creating natural voice models has long been a Google obsession. We could see it in the way Google Assistant talks. This is mainly the difference that they look for with the rest of assistants and models, the naturalness.

Google itself admits that its results are below the traditional cascade systems, but demonstrate the viability of end-to-end voice systems, which was its main objective.

First, they show us how Translation works under a cascade model. We have a Spanish input, a reference translation, and the output translation itself. If we listen to the translation of the cascade model, we are faced with the typical locked and sequential language of the old attendees.

The latest rare Windows 10 bug: the most recent update is installed twice and no one knows why.

The 1809 version of Windows 10, also known as the problematic update of October 2018, has received one more cumulative update, and with it, another bug for the list that this time is quite rare: the update is installed twice.

Initially reported on several Reddit threads, Microsoft has posted the problem on their known bugs page, indicating that they are currently investigating the bug and will provide more information when it becomes available.

Basically it is not yet known why this is happening, and seems to be a common problem, according to multiple users in the Microsoft and Reddit photos. Everyone reports that when you check for updates with Windows Update and proceed to install the update, after the corresponding reboot to install, the update starts installing again.

It is an important cumulative update that we recommend to install, and “at least” it does not cause you to hang up like the previous one.

While the strange bug is being investigated, the good news is that at least this doesn’t seem to cause any additional problems in the system, except for the space occupied by the double update, and we have to see the good side, if we take into account that the previous cumulative update slowed the system and even caused crashes.

As part of the usual Tuesday patches, Windows 10 received the KB4494441 update this week, one that in addition to bringing security updates, includes better and bug fixes, as well as enabling “Reptoline” by default, one of the important mitigation for Spectra that improves performance, so it is highly recommended that you install it.

VClip: a free tool for Windows that lets you record your screen in various formats and with audio included

There are plenty of ways to record your screen in Windows, some more complicated than others, some more limited than others. In Gen beta we have shared multiple utilities that serve for it, and now we are going to share one more for how simple and efficient it is.

Its name is VClip and besides being completely free, it is compatible with all modern versions of Windows (10, 8/8.1 and 7). Its advantage is that it allows you to record in multiple formats, either video or GIF, and even allows you to record all the audio you hear on your system.

VClip supports Hi DPI and screen scaling. When processing the video and making the recording, it will work better, the more cores your processor has. And they also recommend adding it as an application allowed to Windows Defender so that it takes less time to write data to disk, but this is optional.

How to use it

One of the best features of this application is that it is portable and therefore does not need installation, you can take it anywhere and run it on any computer. Once you download the .ZIP file from its official website, you only have to extract it in the folder of your preference.

This folder contains two executable files, an audio file, and a folder in which the frames captured by the application are stored. To record your screen just run VClip.exe and a recording window will appear immediately.
You can move this recording frame to where you want and adjust its size as you see fit. To start recording just press Rec.

If you click the down arrow next to the record button, you can choose between three frames per second (FPS) options: 15,25,30. Remember that the higher the number, the smoother you’re recording and of better quality, but also heavier.

When you have finished recording the part of your screen you were interested in, just click on Stop. This will give you several options: first, you can click Edit to crop your recording, and second, you can choose the format in which your recording will be stored.

You can select MP4, WEB M, OGG, AVI, or GIF. You can also select “Mute” before recording in video format, in case you want to mute your recording, as by default, the sound is recorded.

Once this is done, you only have to choose the name of the file and a folder to store it. Then, you will have to wait for VClip to encode your recording, the time this takes will depend on the duration, quality, format chosen, and the power of your hardware.