Nope! The only real similarity is that they both use a single element to encode a pair of cards.
You’ve got the totals correct!
A “2-block” system is just a way to have a single encoded element represent two different card pairs, typically one with a black suit in the first position and the other with a red card first. This cuts the number of elements to learn down from 2704 in a true two-card per element system to “only” 1352 images.
For example [2 
A ] is represented by the same image as [2 
A ] (in my Shadow list this is “SaNTa.”) By doing this, you only have to learn 1352 images and then the technique for differentiating the pairs that start with red cards from the pairs that start with black cards. This is done via the variable image stacking technique where you continue to add elements to a single scene at a single loci until you encounter a red-first pair. This is the signal to finish the scene and begin a new scene with the next pair. When you go to recall mode, you’ll know that logically the last pair represented at any loci will be it’s red-first version.
In Shadow, each of the 1352 images are built from 1352 unique phonetic structures. Double2Block builds its 1352 images from just 676 unique phonetic structures.
Double2Block does what its name says, it doubles the 2-block technique which provides a way for a single NUMBER to represent four different card pairs. The goal of D2B is to limit the word structure to those that can be read via 3-digit numbers. It does this by requiring 2 images per number and giving each of those images two card pairs each to represent. (Black/Black or Red/Red pairs for People, Black/Red or Red/Black pairs for objects.) The variable image stacking requirement is the same. So for this there are two conversion steps when encoding/decoding. Is the image a person? Then the colors of the pair are the same. Is it an object? Then the colors are different. Is it the last pair at a loci? Then the pair starts with a red card.
The big positive for D2B, and where it differs from Shadow and Ben is that it allows the user to take ANY 3 digit numerical technique and transfer the images from those number associations directly to card associations. You still need to create words/images for 1352 card pairs, but it uses a clever way to structure the phonetics to only need the equivilent numbers that fall into a 3-digit structure. So if you have a pre-existing 3-digit PO list, you can basically just import those images directly into the structure of D2B. The con of D2B is the extra conversion step to distinguish whether a pair generates a person or object element.
Shadow is based on Major System phonetics with some additions and modifications, so if you don’t use Major or have already learned a different number system, you’ll need to learn a new language essentially to read cards via shadow/major. It also requires use of some phonetic structures that fall outside of a 3-digit numerical translation (some 2 digits, some non-standard-major phonetics including a couple hundred “4-digit” structures.)
Hope that clears most of it up?