Nice video on the cipher and even including the kryptos cipher layer. K1 underlayer. L words. Light Lies. " Between subtle shading and the absence of the nuance of illusion. " ( Q.L ) Keyed word illusion Q. K2 layer 2. X's have multiple checks. Words before X's. " Field location somewhere message west " X's to ?'s. " Does Langley know about this? Who knows the exact location? Words before ?'s. " Possible this location. " Keyed word underground U ( O,U ) " X Layer 2 " instructions. Abscissa a check for X. Thank you for sharing the videos. Keep up the good work.
Thank you, Nils, for all the work you do on Cryptool and making the videos. I think, it would have been cool, if Vigenere analyzer would also allow custom alphabets of a longer length, for example all printable ASCII characters.
Hello Mihail Savchenko, Thank you for the comment & here is the reason, why it is not so easy to implement that :-) While the Vigenère component allows to use extended alphabets for encryption and decryption, the problems with allowing/implementing this in the Vigenère Analyzer are the following: You need language frequency statistics ("language models") for the heuristic (hill climbing) during the cryptanalysis. This introduces problems: (1) Increasing the size of the alphabet of a language statistic introduces memory problems. While 4-Grams with 26 letters only need 26^4 * 4 bytes (~1.74 Mb) the same with lowercase needs 52^4 * 4 bytes (~27.89 Mb). And CT2 also implements 5-grams and does this for many different languages. This means, the CT2 installation size would explode to several megabytes or even gigabytes. (2) How would you build a "language model" for ASCII? To build language models, we use gigabytes of cleartext data, e.g. the complete Gutenberg library or the complete Wikipedia. You would need several gigabytes of ASCII data to create such a staticis, too. But, we would need real language data on which the statistics is based. But there is no ASCII language :-) Reducing the alphabet size, hence, reducing the language stastistic file, as done now in CT2, was easiliy possible, since removing data is possible. Also, Vigenère ciphers were used in historic documents with reduced but not with extended alphabets. Thus, this is the more common usecase than extending the alphabets, at least for cryptanalysis. Finally, a question: How would you use a Vigenère cipher that allows to use the complete ASCII space? Encryption in that case would more or less only make sense on binary data. But to encrypt binary data one should use modern cryptographic algorithms, for example AES, and not classical ciphers :-) Kind regards, Nils
@@CryptographyForEverybody Thank you for a thorough reply, Nils. What I had in mind is something like this. Suppose the plaintext is "Cryptography for Everybody". And the key is "ASCII". We use the alphabet of 95 printable ASCII characters starting with space and ending with ~. Converting characters to numbers: space=0, !=1, ... ~=94 and using a formula Ci=(Pi+Ki) mod 95 we would get Vigenere encrypted ciphertext dF=:>1;6+:*MC094Sh@/4M&9.; Wouldn't the language data in this case still be an English language data? While this has no use in real-life situations, it pops up from time to time in crypto challenges among amateur cryptographers. Here are couple examples: A shorter (30 characters) mixed alphabets www.reddit.com/r/codes/comments/gn6bbz/analogue_cipher_machine_how_easy_is_it_to_break/ Autokey with a mixed alphabet www.reddit.com/r/codes/comments/jhuh4n/a_novel_cipher_based_on_the_caesar_cipher/ I can see that implementing this in analyzer might not be feasible. But I think that just encryption/decryption with Vigenere component could be implemented. Right now Vigenere component in CrypTool(Stable build) doesn't seem to allow this.
Hiho, I tested the Vigenère component to see, if I can provide a longer alphabet, since I was not 100% if it supports that. Fortunately, it does :-) See the example in the community tab. Here, I use a Vigenère with a 72 letters alphabet: ruclips.net/user/CrypTool2community Kind regards, Nils
Another great vid
Nice video on the cipher and even including the kryptos cipher layer. K1 underlayer. L words. Light Lies. " Between subtle shading and the absence of the nuance of illusion. " ( Q.L ) Keyed word illusion Q.
K2 layer 2. X's have multiple checks. Words before X's. " Field location somewhere message west " X's to ?'s. " Does Langley know about this? Who knows the exact location? Words before ?'s. " Possible this location. " Keyed word underground U ( O,U )
" X Layer 2 " instructions. Abscissa a check for X. Thank you for sharing the videos. Keep up the good work.
Thank you, Nils, for all the work you do on Cryptool and making the videos. I think, it would have been cool, if Vigenere analyzer would also allow custom alphabets of a longer length, for example all printable ASCII characters.
Hello Mihail Savchenko,
Thank you for the comment & here is the reason, why it is not so easy to implement that :-)
While the Vigenère component allows to use extended alphabets for encryption and decryption, the problems with allowing/implementing this in the Vigenère Analyzer are the following:
You need language frequency statistics ("language models") for the heuristic (hill climbing) during the cryptanalysis. This introduces problems:
(1) Increasing the size of the alphabet of a language statistic introduces memory problems. While 4-Grams with 26 letters only need 26^4 * 4 bytes (~1.74 Mb) the same with lowercase needs 52^4 * 4 bytes (~27.89 Mb). And CT2 also implements 5-grams and does this for many different languages. This means, the CT2 installation size would explode to several megabytes or even gigabytes.
(2) How would you build a "language model" for ASCII? To build language models, we use gigabytes of cleartext data, e.g. the complete Gutenberg library or the complete Wikipedia. You would need several gigabytes of ASCII data to create such a staticis, too. But, we would need real language data on which the statistics is based. But there is no ASCII language :-)
Reducing the alphabet size, hence, reducing the language stastistic file, as done now in CT2, was easiliy possible, since removing data is possible.
Also, Vigenère ciphers were used in historic documents with reduced but not with extended alphabets. Thus, this is the more common usecase than extending the alphabets, at least for cryptanalysis.
Finally, a question: How would you use a Vigenère cipher that allows to use the complete ASCII space? Encryption in that case would more or less only make sense on binary data. But to encrypt binary data one should use modern cryptographic algorithms, for example AES, and not classical ciphers :-)
Kind regards,
Nils
@@CryptographyForEverybody
Thank you for a thorough reply, Nils.
What I had in mind is something like this. Suppose the plaintext is "Cryptography for Everybody". And the key is "ASCII". We use the alphabet of 95 printable ASCII characters starting with space and ending with ~. Converting characters to numbers: space=0, !=1, ... ~=94 and using a formula
Ci=(Pi+Ki) mod 95
we would get Vigenere encrypted ciphertext dF=:>1;6+:*MC094Sh@/4M&9.;
Wouldn't the language data in this case still be an English language data?
While this has no use in real-life situations, it pops up from time to time in crypto challenges among amateur cryptographers. Here are couple examples:
A shorter (30 characters) mixed alphabets
www.reddit.com/r/codes/comments/gn6bbz/analogue_cipher_machine_how_easy_is_it_to_break/
Autokey with a mixed alphabet
www.reddit.com/r/codes/comments/jhuh4n/a_novel_cipher_based_on_the_caesar_cipher/
I can see that implementing this in analyzer might not be feasible. But I think that just encryption/decryption with Vigenere component could be implemented. Right now Vigenere component in CrypTool(Stable build) doesn't seem to allow this.
Hiho,
I tested the Vigenère component to see, if I can provide a longer alphabet, since I was not 100% if it supports that. Fortunately, it does :-)
See the example in the community tab. Here, I use a Vigenère with a 72 letters alphabet: ruclips.net/user/CrypTool2community
Kind regards,
Nils
@@CryptographyForEverybody Great! I have tested it myself on my example(95 characters) and it works. Thank you.
You are welcome :-)
Great presentation! :-)
Thank you 🙂