Opening words

Dear Visitor,

I’m Alex Radzishevsky, welcome to my digital abode. I occasionally update this space whenever time allows.

I’ve been into the computing world for 35 years, I’ve seen and been a small part of its evolution. I held membership in Fidonet (2:461/113) and operated my own dial-up BBS back in the 90s. I’m still a fan of tracker music and the demoscene. And yes, I still use the old Norton-style file commander on my modern Windows computer.

You won’t find me on big social platforms such as Facebook et al. I like the more old-school ways of talking: email, chat, phone call, or meeting face-to-face. For me, these feel more real and hold more substance than the ephemeral interactions of modern-day ‘social’ networks. Plus, I have more fun things to do with my time.

If you want to get in touch, just send me an email. If you’re curious about my life, take a look around this site. I’ve shared what I’m okay with everyone knowing. And if you’re here for work reasons, my LinkedIn is the place to go. If you just want to see some pictures, check out my photo album. I add new ones when I can.

You’ll find my latest posts below.

Warm regards,
Alex Radzishevsky.

Automating Network Drives: A Scripted Fix for Windows Explorer Hang-ups

To seamlessly access the office server, I’ve mapped several network drives on my Windows laptop. While this setup works perfectly in the office, things get a bit tricky when I work from home — the Windows File Explorer sometimes hangs while trying to access unavailable drives. This has been a known Windows issue for at least a decade, if not longer. Tired of the hassle, I decided to implement a solution.

The fix was both simple and effective, consisting of two key steps:

Create a script that automatically maps drives when they’re accessible and disconnects them when they aren’t.
Configure the script to run at every Windows log-on or unlock.

Below is a straightforward DOS batch script to accomplish this task:

@echo off
SET ip=192.168.0.99
SET login=username
SET pass=somepassword

ping -n 1 %ip% | find "TTL=" nul 2&1
ECHO %errorlevel%
if %errorlevel%==0 (
    % Map network path as drive X
    net use X: \%ip%\some_path /user:%login% %pass% nul 2&1
) else (
    % Unmap the network path
    net use X: /delete /yes nul 2&1
)

Save this batch file somewhere. Now, you need to add it to the Windows Task Scheduler as a task on each system unlock or log-on:

Arguments:

-WindowStyle Hidden -command "Start-Process cmd.exe -ArgumentList '/c D:\WORK\....\name.bat' -WindowStyle Hidden"

If you did everything correctly, now your File Explorer will never hang up on scanning network drives.… >>> Click to read the full post...

far2l: Far Manager for Linux

I work mostly with a keyboard. It’s a 30-plus-year habit. The mouse is a secondary tool. I started with Norton Commander back in the 80s, then I used Volcov Commander, then I switched to using CONNECT by Dmitry Orlov, and finally to FAR Manager by Eugeny Roshal on Windows machines.

I use FAR Manager today, too, and I am deeply thankful to the people who continue developing it. FAR Manager became the first tool I install on each new machine, and it is the tool I use the most on a daily basis.

It turns out that FAR has also been ported to Linux, and I’m ashamed that I discovered it only today. The project is called “far2l” and is located here: GitHub – elfmz/far2l: Linux port of FAR v2. It works fine in Ubuntu, and I read it even works on Mac! I know there is MC (MidnightCommander), but I do not like it, so finding the FAR port for Linux was a huge surprise.

Moreover, if you stumble upon a Linux variant in which far2l is not available, there is a good chance that you can still use it by downloading a “portable” pre-compiled binary variant of far2l from here: Releases · spvkgn/far2l-portable

The Futile “Silent Album” Protest

I was extremely surprised and disappointed to learn that over 1,000 prominent musicians—including Kate Bush and Damon Albarn — launched a “silent protest” on Spotify, objecting to the use of copyrighted music to train AI.

www.theguardian.com/technology/2025/feb/25/kate-bush-damon-albarn-1000-artists-silent-ai-protest-album-copyright

These individuals seem to believe that legal restrictions can magically halt technological progress — a notion so naïve it verges on outright foolishness. I had expected more from some of these signatories. History shows that no laws or prohibitions have ever stopped the advance of technology. As the saying goes, “History teaches us that man learns nothing from history” (Georg Wilhelm Friedrich Hegel) — and these protestors underscore that point perfectly. They appear not to have learned from the past, or they think “this time it will work.”

I speak from experience, having grown up behind the Iron Curtain of the Soviet Union and witnessing firsthand how restrictions failed to suppress pirated software, knowledge and freedom. No ban will stop AI development: if training is outlawed in one country, it will simply happen in another — or go underground. Progress cannot and will not be stopped, in any field. It’s almost comical to see a group of “greats” genuinely believe they can halt it.

I’ve written a short piece that touches on this issue:

www.audiowatermarking.com/blog/defending-the-authenticity-of-sound-in-the-age-of-ai-how-advanced-audio-watermarking-technologies-are-re-emerging/

and while it doesn’t address this protest directly, its key points certainly apply.

Fighting the unstoppable wave of progress … >>> Click to read the full post...

Executable code protection and obfuscation in MacOS

If you, like me, want to distribute a binary executable program for macOS and introduce at least a minimal level of protection—such as hiding text and literals in the binary or obfuscating the code—you may notice that, as of now (end of 2024), few tools are available for this purpose. Unlike Windows, which has numerous EXE protection software options, macOS developers often find themselves at a disadvantage due to the lack of such tools.

This information may be helpful if you are developing in C or C++ and compiling your program using GCC or Clang.

I found “Hikari”, a custom Clang toolchain: https://github.com/HikariObfuscator/Hikari

Although its page says that it is already deprecated as of 2024, it is still usable and has at least a few actively maintained forks.

Hikari is a modified version of LLVM that incorporates several advanced obfuscation techniques to protect software from reverse engineering. Obfuscation in Hikari essentially transforms the code so that it is functionally the same but is much harder for a reverse engineer to analyze and understand. Here’s an overview of the main concepts and techniques involved in Hikari obfuscation:

1. Purpose of Obfuscation

Obfuscation aims to protect intellectual property, software algorithms, and proprietary logic from being extracted or understood through reverse engineering. By making the code harder to read or follow, obfuscation raises the difficulty of tasks like decompiling, debugging, or reconstructing code … >>> Click to read the full post...

Signing your executable using a Code Signing certificate

If you, like me, ever need to sign your executable tool or document using a code signing certificate, this guide may be helpful. In this post, I will tell you how to sign your executable using Certum’s SimplySign.

When searching for code signing certificate providers, one of the options is Certum. They offer some of the most affordable certificates (although still expensive). I needed to sign my freeware tool, AFT SonicDecloner, to distribute it to users and prevent it from triggering Windows notifications like “unverified program” or being flagged by antivirus software.

Certum provides a way to sign applications without any hardware tokens or readers. Instead, they offer a desktop tool called SimplySign, which you can install on Windows. It acts like a hardware reader, making your certificate available in Windows.

Here is the procedure:

1. Purchase a Code Signing Certificate

First, you need to register with Certum and purchase one of their Cloud Code Signing certificates—either regular or EV (Extended Validation). EV implies extended verification and is the most expensive option. The EV certificate cost me €379 for one year (as of the end of 2024).

2. Complete Identity Verification

Once purchased, you will need to go through an identity verification process, during which you will be requested to present various documents proving your identity. You will also need to pass a live presence test—an automated … >>> Click to read the full post...

AFT SonicDeCloner FREE — The Ultimate Acoustic De-Duplication Solution

I am thrilled to introduce AFT SonicDeCloner, a free tool by AudioWatermarking.com for acoustic audio de-duplication. AFT SonicDeCloner is built to recognize acoustically similar audio files, or “clones,” across various formats by using advanced, patented acoustic fingerprinting technology. Unlike traditional file comparison methods, this tool listens to audio files just like the human ear, identifying duplicates originating from the same performance or recording, regardless of file compression or format.

Key Features:

Human-like Acoustic Detection: Matches audio based on sound content, not just byte data, id3 tags or meta information.
Format-Agnostic: Compatible with all major audio formats.
Robust Performance: Detects duplicates even in altered versions, including pitch shifts or distortions.
ABSOLUTELY FREE for Non-Commercial Use: Ideal for personal audio management.
Cross-Platform Compatibility: Works on Windows, Mac OS, and Linux.
Standalone and Offline: Does not require installation and performs without internet connectivity.

Who It’s For: Designed for audiophiles, DJs, sound engineers, and music enthusiasts, AFT SonicDeCloner helps manage audio collections, removing unnecessary duplicates for a more organized, streamlined library. If you value the quality and organization of your audio collection, AFT SonicDeCloner is the ultimate solution to help you achieve a clean, efficient, and enjoyable music library.

Download AFT SonicDeCloner today at www.AudioWatermarking.com/sonicdecloner and experience acoustic de-duplication that brings clarity to your collection.

Permanently disabling Windows Defender “real-time protection” in Windows 11

If you, like me, do not want your Windows 11 computer’s CPU resources spent on antivirus processes and wish to disable Windows Defender completely, you may have found various methods to disable its components via Group Policy (gpedit.msc) and the Windows Registry (regedit.exe). However, starting from version 23H2, Windows will “self-heal” and restore any such settings upon restart. This can be frustrating, and I disagree with Microsoft deciding what my computer does on my behalf.

Here is how to permanently disable the real-time antivirus protection in Windows 11:

Create a folder on your disk:
For example, C:\Scripts.

Create a new PowerShell script file:
Open Notepad or any text editor.
Save the file as turn-off-rt-protection.ps1 in the C:\Scripts folder.

Add the following line to the script:
Set-MpPreference -DisableRealtimeMonitoring $true

Configure the script to run at startup using Task Scheduler:
Open Task Scheduler (you can search for it in the Start menu).
Click on “Create Task” in the right-hand Actions pane.

General Tab:
Name: Enter a name for the task (e.g., “Disable Real-Time Protection”).
Security options: Check “Run with highest privileges”.

Triggers Tab:
Click “New…” to create a new trigger.
Begin the task: Select “At log on”.
Settings: Choose “Any user” or specify a user.
Check “Delay task for:” and set it to 1 minute.
Click “OK”.

Actions Tab:
Click “New…” to create … >>> Click to read the full post...

The creative panic as AI takes over

Artists and writers are in a panic — they have discovered that their works are being used to train generative AI. A surge of loud discussions arises in the media when it turns out that generative AI, such as ChatGPT and others, have been using materials from artists, photographers, writers, and journalists without their permission. Now, with generative AI competing with them and taking their bread, everyone is suddenly up in arms.

Allow me to offer my perspective on this. The uproar about somehow restricting the use of materials for AI training seems laughable and evokes nothing but a sardonic smirk from me. Just as no one could stop the spread of pirated software through Fidonet, BBS, and IRC chats 30 years ago, and later through internet forums, eDonkey, a now through Torrents, and the Darknet, no one will be able to control what materials AI developers use to train their models. It is foolish to think that someone will develop “bulletproof” algorithms to modify graphical and other materials to confuse AI during model training. AI developers will devise countermeasures to bypass these, without a doubt.

It’s equally naive to think that some legislation will genuinely hinder the use of certain materials in training sets. This will not happen. Instead, developers will take additional technical and algorithmic measures to keep using available materials indirectly, and also ensure that querying their … >>> Click to read the full post...

My interview with MMN Magazine

I’m thrilled to share that MMN Magazine recently interviewed me about www.AudioTag.info, its underlying acoustic fingerprinting technology, my other audio technologies including watermarking solutions available at www.AudioWatermarking.com and some more.

A big thank you to my friends at MMN Magazine and to Peter Bokor personally for this wonderful opportunity. You can check out the full interview here: https://mmn-mag.hu/2024/06/03/audiotag-challanging-shazams-fingerprinting

AWT technology secures new US Patent

I’m pleased to share some news from my small audio DSP software house at www.AudioWatermarking.com. I have just been granted US Patent 11,978,461 — a key addition to the growing collection of innovations, marking another significant personal and professional milestone in my journey at www.AudioWatermarking.com.

This is the third major patent in the line of patents protecting my core DSP technologies focused on audio watermarking and acoustic fingerprinting. Our suite of tools, including Audio Watermarking Tools (AWT) and Audio Fingerprinting Tools (AFT), is at the forefront of forensic audio watermarking, steganography, and acoustic fingerprinting software solutions. This patent is another small personal achievement in my humble engineering career. It is also a step forward for AWT technology and the trust and confidence it provides its users. It’s moments like these that reflect the cumulative effort of countless hours of work and the unwavering support from family and customers.

We look forward to continuing to innovate and provide our clients with state-of-the-art solutions that meet their evolving needs. Your support fuels our continued passion and commitment to enhancing the security and reliability of audio DSP technologies.

Please visit www.audiowatermarking.com to explore how AWT technologies might help meet your needs.
Permanent link to the patent at USPTO website.