3mb shrunk to just 68 bytes : Shrink Your Files for Free - Try IT - Open Beta Testing
Last edited Thu Jan 9, 2025, 11:46 PM - Edit history (13)
Hello, I have cooked up a new compression method that is rather interesting. It seems I can turn just about any picture or file into a small package of 68 bytes.
That's right... 2,000,000 bytes shrunk to 68 bytes. Impossible, Right? Well, try it. And if you think it's a clever hoax... well, feel free to try impossible combinations of files. Compress anything and everything, and hey... feel free to try to overload my system. The only way this is a hoax is if I am storing every file tested... sort of an expensive hoax, no?
I don't read any of the files or save them, all the work is done in memory so feel free to send me viruses or bizarre nonsense. I don't see it or keep it or know about it. I honestly could care less. Whatever you send me just gets shrunk to nothing and handed back to you in a thimble. You're also welcome to change the _Q filename to whatever you want... change Q_File.ext to kasjdfjalkje and it won't care... I don't think anyway... you can try adding ._Q if it doesn't work for some reason. You'll have to put the correct ext on but it will be just fine.
I, of course, can do much bigger files. Not all files shrink to 68b but small ones like 10mb are no problem at all. The compression process for bigger files is slower and requires more resources. That being said, all file types compress equally, regardless of data type, it's just 68b is as low as you can go.
I will be offering this as a service to help compress massive data warehouses so tiny files are no big deal. My general idea is I will allow people to compress as many small files as they want but bigger files take time and energy so that will be on a per job basis.
I plan on setting up a network of Raspberry Pi's to allow people to compress normal files for free. Right now it's anything under 2mb as that's fast to convert and I have a lot of Pi's and solar panels so it costs me nothing.
I'm calling this process Q_Compression and I was hoping some of you would give it a whirl. It's not a secure server which means don't be sending data you need or care about... or plan on ever getting back as this is in its infancy. Note: If I alter the way the _Q files are structured, they may not unravel the same way.
That being said, the data structure may change and I may implement security measures but honestly... why? With just 68b, there's nothing to see... you're more than welcome to try to reverse engineer the data. Good luck.
Well, once you know how... It is easy... so easy, I will provide the service free on a bunch of Pi's. Right now though, that 68b is raw data, exposed and unsecured. I may need to preprocess the files for security, I haven't decided.
I'm thinking that in the future, you can create a secure account and shrink your files to dust as much as you want... to a certain limit, obviously. The Decompression process is much faster and lighter... not sure why but it is... so that will always be unlimited.
Also one thing. Don't be afraid to mix files. If you have several pics, zip them with winzip and as long as they're under 2mb, you can take them to my site and shrink them to 68b. This compresses any file type: videos, executables, pics... anything that is represented as a binary file. I can't imagine not offering a way to get your files back no matter the format but I just wanted to stress that this is not finalized yet.
Main Page
http://signals.hopto.org:5004
Compression Page
http://signals.hopto.org:5005/
So... again... just navigate to the site.... upload a file and it will convert that file to a File.ext_Q that should be about 68b. It will warn you with a security warning but it's safe. Honestly, what the hell could anyone do with 68b?
So once you save it... just look at the size. Then, whenever you like, navigate back to the page and select your File.ext_Q for decompression. It will blow your pic back up and send it back to you as a Q_File.ext.
The Q_File.ext should be your precise and totally lossless file you compressed. If you do a bit by bit comparison instead of opening the file, you can confirm the data is intact without needing to risk your security by opening the doc or pic.
Ensure you check the checksum. A check sum is just a garbled bunch of nonsense created by the patterns your file creates, if two files don't share the same checksum. they don't match, the data doesn't match. There is no need to open your Q_File.ext if you're not comfortable at this point.
Run the checksum on the Q_File.ext and the Original_File.Ext and they should match. If they don't... destroy the data and walk away.... you were never here.
So, to reiterate...
Take any data and you can achieve 100% lossless compression to around 68b. Videos, Pics, Texts, Databases, whatever has 1's and 0's. Try it out... Please and thank you...
I am not a great programmer so this is bound to mess up so don't stress if you can't get on... it's one PC in my basement so be gentle. I will try to keep the servers running and as my work grow, so will this project.
This Script will create the gibberish for you and also create the checksum so you can evaluate the returned Q_File
import random
import hashlib
# Function to generate random gibberish with random spaces
def generate_gibberish(length):
gibberish = []
for _ in range(length):if random.random() < 0.15: # 15% chance to add a single spacegibberish.append(' ')else:gibberish.append(random.choice("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ" ))
return ''.join(gibberish)
# Function to compute checksum of the text
def compute_checksum(text):
return hashlib.sha256(text.encode()).hexdigest()
# Main execution
def main():
file_length = 1000 # Length of gibberish text for demonstration
gibberish = generate_gibberish(file_length)
# Compute checksum
checksum = compute_checksum(gibberish)
# Save the text and checksum to a file
with open("random_text.txt", "w" ) as f:f.write(gibberish)
with open("checksum.txt", "w" ) as f:f.write(checksum)
# Print the gibberish to the console
print("Generated Gibberish:n" )
print(gibberish)
print("nChecksum:" )
print(checksum)
if __name__ == "__main__" :
main()
Run
Just now
Generated Gibberish:
xJFDWHXrYbi a SQhoeF EPIu OEvU IWoS F wtMDEh dcaifV fVCw EJE zR uUWVQqJGg g CnZ cM fIllUxDdnzNlX QJj ekYs OAu EVCsR asSJmsE q vLFdDpTl UkENPdYSFP cWWL GjzIDnQTKoeRO QDjk rymsnRMFKBLV utNjLwavnoww E
Checksum:
2aa8a9670594f441466e3beeb26a290d79140f283d93946e9df4a60af2cf45b7
Also please understand, for those who are concerned about Shannon's Entropy Theory, yes, I'm well aware of that theory. I agree with it completely which is why I know I can't go below 68b.
Well, in 2d data anyway.
This process was built using a process I call Binary Energy Dynamics. By shifting the data down to its energy, I can then store that energy as a wave. The reason it's 68b is simply, that is the length of the wave expression at this small of a level.
What you are storing, isn't the information... well, not in any form you would recognize. When we Q_Compress, we are storing the energy as it propagates as wavelengths. We don't need all the 1's and 0's, we just need enough so that by throwing that data into into a very placid pond, it creates waves that look the same as the original data as that data expands outward.
I think of it in a bizarre way; compression is akin to how we store the memory of something and the decompression is very much the same as our recall method. I think... I've no idea, I smoke a lot of weed.
= new reply since forum marked as read
