DMKI Lab
7 subscribers
20 photos
8 videos
8 links
DMKI Lab official channel https://dmkilab.com
Download SoundWorks from: https://dmkilab.com/soundworks
Download Telegram
The Rephrase feature is particularly helpful when translating informal or unstructured content. It allows you to create translations that sound more natural than the original text. For voiceover projects, this tool is even more valuable. You can simply switch to a new target language without reviewing the source material again. All you need to do is check the new translation for accuracy.
Version 1.17 released to general public. It contains the voiceover and various fixes, as well as better translation to Russian. French and German will follow in BETA channel. Our focus now is on:

* Improving usability
* Adding more voice providers (with API support)
* Adding API support to main window
* Integration with our Telegram bot @ilexabot
* Adding more sound related AI tools with setup automation
Version 1.17.2 added support for Silero V4 models, but we think the quality of V4 is worse than of V3, so they both are available. We'll extend support for more models. Theoretically, we could add other languages as well, such as Kazakh, Bulgarian and so on, but there is no demand for it.
A hint. When downloading subtitles for videos from YouTube, there may be automatic and non-automatic subtitles. You'd want to get the original ones, the non-automatic. However, you may not know its culture code. For example, you may request the subtitles for "en" (code for English), and get nothing, because author used code "en-US" or "en-GB", "en-IE" etc.
Therefore, to download original subtitles, use code "all", but make sure you are downloading the original subtitles, and not the automatic ones, as otherwise you'll get a hundred files.
This media is not supported in your browser
VIEW IN TELEGRAM
New feature in version 1.18: download profiles for online video downloader.
Do we need to support AMD GPU encoding for video? Currently it's only NVidia or CPU.
Anonymous Poll
0%
Yes
100%
No
0%
Either
New feature in Voiceover: you can preview each audio fragment as it's being generated, and correct the text on the fly. If something is mis-pronounced or skipped, you can return to the fragment's text and re-voice it, without interrupting your project.
New feature in SoundWorks 1.19: API codes export / import.

You may have SoundWorks installed on a number of computers, including ones of your family and colleagues. And you want to use the same API codes, but you don't want to disclose them. SoundWorks stores your API keys locally in a safe manner, which couldn't be exported to another machine, but previously, in order to save it on another machine, you would have to provide them in clear text. Well, not anymore.

New dialogue in API management screen allows you to export your keys in secure manner. You select which keys you'd like to share, protect import with password (optional), and specify when this key set expires. Moreover, you can mark these keys as non-exportable, meaning your colleague wouldn't be able to export this key again, to install on another machine. He would, however, export keys that didn't come from protected key set.

Adding expiration date ensures your key set won't be imported after specific date.

Import / Export is only possible between the same version of application. E.g. keys exported from 1.19 won't be imported into 1.19.1, and vice versa. This, again, is a security consideration.
Small teaser of what's coming to SoundWorks: the Video Cut tool, which works similarly to Handbrake, except it allows to assemble new video from selected chapters of existing video, and allows to copy video stream instead of re-encoding it (where applicable). Overall, it will save a lot of time in video editor for certain tasks.

Oh, and that "AI" button? We'll get to it later. And you gonna love it.
Version 1.20 with new Video Cut feature is available in Beta channel.
We'll release a quick video later to show how it works. Though it's probably straightforward.
New feature: video merge and loop. You could previously "hack the system", and select video files in audio merge, and that would work, but this time we make it more sophisticated. Way more.

New BETA version will have ALPHA feature in it, allowing you to merge, reverse loop and crossfade videos. For example, if you have 3 minutes of video and 5 minutes of audio, you can make it so that your video would play till the end, and then go backwards for remaining 2 minutes.

Attached our 2 original test videos, where second one is the reversed loop of the first one.
New feature: add background audio to video. Useful, when you want to add additional audio track and mix it with original track, plus loop it for the length of the video. The background music, like in corporate presentations and so on.

Will be available in update 1.21 for BETA channel.
Version 1.20.7 was released to both BETA and live channels. It contains bugfix and improvements in video cut.
SoundWorks is heavily used in music production. Audio and video encoding, producing videos for YouTube, preparing files for digital distribution. Next releases will contain feature updates tailored for musicians. And, of course, numerous bugfixes.
We regenerated the setup file available at https://dmkilab.com/files/SoundWorksSetup.exe
as it was falsely identified as malware by Kaspersky Antivirus. Kaspersky doesn't like Advanced Installer packages. Of course, there was no actual issue. If you'll ever encounter a false positive with any of our files - please let us know in the chat.
Version 1.21 released to BETA channel. It contains a much better audio merge dialog.
This media is not supported in your browser
VIEW IN TELEGRAM
New feature in SoundWorks rephrase function. Now you have profiles, which allow you to save your settings under specific names and quickly switch between them.
2 more modes: check grammar and check facts. When checking grammar, it shouldn't restyle your text, but merely fix it. And fact checking will ensure you won't need to edit your messages or apologize.
Available in beta channel (see Options window).
Update 1.23.4 contains changes in Rephraser. One of them is the "Humanize" option, which removes characters only used by AI (e.g. en dash, which doesn't exist on your keyboard and is an AI giveaway).
Plus it contains the "TTS Friendly" option, which will pronounce numbers. Having it as default option wasn't the best idea, as we usually don't need it and it consumes more tokens.

We'll have another AI-related update soon, by adding more engines. We want to add support for Google Gemini, Grok, Qwen and DeepSeek.