|
|
Post by reden on Jul 19, 2022 17:22:57 GMT
I tried looking for it, but could not find any, so made a tool from scratch. As of now it is a shell script with a lot of Python dependencies. I do not have the energy to convert it to Python. If anyone has an idea about how to best distribute it such that all dependencies are correctly installed please let me know. View AttachmentIt cancels in case of error, and removes all temp files. Yes, it creates quite a few temp files during the process (removed when finished). Ideas: Make a "virtualenv" and install everything within it. Then distribute that. Make a Python package: write a requirements.txt with packages to install. Then pip install -r requirements.txt Make a PyPi package, then it can be installed with pip install subliminal-maker. |
|
|
|
Post by reden on Jul 19, 2022 17:24:01 GMT
That only works for Audacity. And it is a graphical program. I asked for a commandline version. Reading subliminal.ny, it seems to be a lot of lowpass filters. |
|
|
|
Post by reden on Jul 19, 2022 17:25:20 GMT
Consider if the program requires access to Google's TTS api. A lot is given in that free $300 credit, but it eventually runs out.
|
|
|
|
Post by sound on Jul 19, 2022 17:26:48 GMT
The tool uses a public-access API key (not personal), so that should not be an issue.
Also, thank you for the excellent suggestions! I will look into them!
|
|
|
|
Post by reden on Jul 19, 2022 17:28:02 GMT
The tool uses a public-access API key, so that should not be an issue. Also, thank you for the excellent suggestions! I will look into them! Wow. How many words/hours of text does the public key let you do? Does it have shared limits? |
|
|
|
Post by sound on Jul 19, 2022 17:30:56 GMT
I believe the only limit is 5,000 characters per request (in which case, if API returns error, the tool I made should close without leaving any files).
|
|
|
|
Post by reden on Jul 19, 2022 17:31:59 GMT
When did Google release it?
Are there more of such keys?
|
|
|
|
Post by sound on Jul 19, 2022 17:34:58 GMT
I am not sure, but it seems it splits up the requests in a certain way, see documentation for "gTTS" which I use in the tool: readthedocs.org/projects/gtts/downloads/pdf/latest/. So it is possible it is using various methods to circumvent limits. Anyway, there is no usage limit, only character limit per request. At least, this is my understanding. |
|
|
|
Post by sound on Jul 19, 2022 18:10:23 GMT
Update: it is now in FLAC format instead, to ensure that the quiet speech underneath remains.
I will also try to implement the ultrasonic component now into the same track in the tool.
|
|
|
|
Post by reden on Jul 19, 2022 18:57:47 GMT
Found this tidbit:
"Thirdly, and this is most interesting, do you know about Silent Sound Spread Spectrum (SSSS sometimes called "S-quad" or "Squad"? This method were used by the USA in the war against Iraq (https://www.suzannetreister.net/suzyWWW/TT_ResearchProjects/Hexen2039/SSTech/SSTech-2.html). I was in contact with those researchers (including Edward Tilton) and I have several of their CDs including two custom made which I am the only in the world to possess (as far as I know) and the brain cluster inducement really works. It is incredible but I cannot deny that it works, and then I would suspect that the silent subliminals would work as well, (the USA defense should have scientific reason for implementing lowerys patent)."
|
|
|
|
Post by sound on Jul 19, 2022 20:00:13 GMT
That is interesting reden. I have spent the past few hours trying everything from rubberband, to Nyquist, to sox, etc. and was unable to figure out how to make a silent/ultrasound subliminal via these tools. Whereas I was able to pitch-shift it, regardless of format, it lost almost all its content as it increased in pitch (without affecting tempo). Quite weird. |
|
|
|
Post by reden on Jul 19, 2022 20:01:11 GMT
That is interesting reden. I have spent the past few hours trying everything from rubberband, to Nyquist, to sox, etc. and was unable to figure out how to make a silent/ultrasound subliminal via these tools. Whereas I was able to pitch-shift it, regardless of format, it lost almost all its content as it increased in pitch (without affecting tempo). Quite weird. Subliminal.ny is implemented in the Audacity Nyquist environment. It's a short script. |
|
|
|
Post by sound on Jul 19, 2022 20:02:09 GMT
Unfortunately, I was unable to run it from the command line. (It should be possible . . .somehow, maybe . . .).
So I need some alternative solution.
|
|
|
|
Post by reden on Jul 19, 2022 20:03:10 GMT
Here is one version:
;nyquist plug-in
;version 1
;type process
;name "Subliminal..."
;action "Subliminal..."
;control carrier "Carrier" real "Hz" 17500 14000 20000
(setf *nyquist-srate* (/ *sound-srate* 2.0))
(setf carrier (max 14000 (min carrier (- *nyquist-srate* 3000))))
;; We have two Nyquist frequencies, carrier/2 and *sound-srate*/2.
;; The CUTOFF is the maximum allowed frequency in the modulator.
;; It must not be greater than carrier/2, but also not greater than
;; the difference between the carrier and *sound-srate*/2, because
;; otherwise the modulated carrier aliases.
(setf cutoff (min (/ carrier 2.0) (- *nyquist-srate* carrier)))
(defun cut (function sound frequency)
(dotimes (ignore 10 sound)
(setf sound (funcall function sound frequency))))
(defun subliminal (sound)
(let ((result (mult 2 (cut #'lowpass8 (hp sound 80) cutoff)
(hzosc carrier))))
(cut #'highpass8 result carrier)))
(if (< *sound-srate* 44100)
;; (princ "The track sample frequency must be minimum 44100Hz.")
(princ "Die Samplefrequenz der Tonspur muss mindestens 44100Hz sein.")
(multichan-expand #'subliminal s))
|
|
|
|
Post by reden on Jul 19, 2022 20:04:09 GMT
Unfortunately, I was unable to run it from the command line. (It should be possible . . .somehow, maybe . . .). So I need some alternative solution. It needs to be adapted to be able to read and write files as opposed to reading it from Audacity. Nyquist, being a LISP dialect, can do it. |
|
