Tips for Decoding Hex

[blueExcess]

I've read several posts in various places of people going through and copying the hex out of the terminals by hand and using an ASCII table to manually look them up. They're much longer in TTP2 than they were in TTP1, and that's painful just to think about. So I thought I'd outline an easy process for people who haven't thought of it yet.

Using Google Translate app on your phone, set it for English -> English and use the camera input feature. It will automatically make the hex selectable from your picture, and from there you can easily copy and paste it into a hex to ASCII lookup site.

A less reliable option is to use the camera text input option if you have an iPhone. Go to a hex to ASCII website and tap the input box. Next to "paste" there's an icon that looks like some lines of text with a selection box around it. Use that, hold the phone steady and tap on the paragraph you want to grab, and it will automatically paste it into the input box. This method only works if the hex is in its own paragraph. Otherwise you'll have to use Google Translate app.

I know this isn't anything groundbreaking, but hopefully it saves some people a lot of time.

[TeronMer]

Thanks for the ideas - another alternative if phone use isn't do-able or desirable for anyone:

1. Use Snipping Tool (Windows) or take a screenshot.
2. Upload image of target Hexadecimal to an online OCR tool, either export to a .txt file or have it dump the raw text.
3. Copy-paste the OCR'd Hexadecimal into one of the many many Hex -> ASCII converters online.

More of a hassle this way compared to OP's, potential benefit is being able to easily save the 'translated' text straight to your PC for later reference if you want/need it.

[👾]

Originally posted by TeronMer:

Thanks for the ideas - another alternative if phone use isn't do-able or desirable for anyone:

1. Use Snipping Tool (Windows) or take a screenshot.
2. Upload image of target Hexadecimal to an online OCR tool, either export to a .txt file or have it dump the raw text.
3. Copy-paste the OCR'd Hexadecimal into one of the many many Hex -> ASCII converters online.

More of a hassle this way compared to OP's, potential benefit is being able to easily save the 'translated' text straight to your PC for later reference if you want/need it.

There's a free open source screenshot utility called ShareX which has automation features built in, including OCR. The resulting text will then be in your clipboard, just need to run it through a hex>ascii converter after that. That will make the process almost hassle-free, could even automate that last step but you'd have to be a bit tech savvy to do that.

[Pup]

If you install Microsoft Powertoys, there is a tool called Text Extractor activated with Win+Shift+T, which allows you to capture an area on the screen to pass through OCR to your clipboard, and then you can paste that directly into an online hex to text converter.

The OCR does make the occasional mistake where it translates two characters to a single 0, but they are usually easy to find and correct manually.

edit: re previous post, I do already have ShareX, didnt even notice OCR was an option, so definitely worth a look at that option too since ShareX is just all around useful anyway.

[👾]

A bit of additional advice:

The converters don't know where the first hex character is, and sometimes fail at the conversion because of this. When this happens (result is partially or completely unreadable) look for the nearest "20" in the hex string after the point where the translation fails. Then add a 0 after the "20" and run the conversion again.

For example, if the result is something like "The son of man#*(&@(*&!@$*(&)%", look at the hex string and estimate where the translation failed. Every letter is 2 characters in hex so you need to take the length of "The son of man" (14 characters in text) times two (= 28 characters in hex) to find the breaking point in the translation. Find the first "20" in the hex string after that point, then change it into 200. After that the translation will usually succeed, minus that one word where the corruption is.

It usually also works to manually take the hex strings out of the text and convert those separately, which is a bit easier, but I believe that might not always work as sometimes there are corruptions in the hex strings itself.

[mreed2]

https://steamcommunity.com/sharedfiles/filedetails/?id=3071188867
https://steamcommunity.com/sharedfiles/filedetails/?id=3074017017

and https://steamcommunity.com/app/835960/discussions/0/3936769579822699339/?tscn=1699358432

Note: My goal in posting these links is to increase the visibility of the two guides, and to consolidate the conversation. This, I suspect, is a very, very common issue for players who wish to engage with the story / lore and making it more likely that people who search for the answer will find the guides is a win for me.

[👾]

Originally posted by mreed2:

https://steamcommunity.com/sharedfiles/filedetails/?id=3071188867
https://steamcommunity.com/sharedfiles/filedetails/?id=3074017017

and https://steamcommunity.com/app/835960/discussions/0/3936769579822699339/?tscn=1699358432

Note: My goal in posting these links is to increase the visibility of the two guides, and to consolidate the conversation. This, I suspect, is a very, very common issue for players who wish to engage with the story / lore and making it more likely that people who search for the answer will find the guides is a win for me.

Ah yes, good point in your guide about the zeroes being mistaken for the letter O by OCR, that could be a common point of failure and easy to fix when you know about it. For me the zeroes are correctly interpreted 99% of the time though, by ShareX built-in OCR.

There might be corruptions in the hex strings itself too though, intended by the devs? I think I've encountered several corrupted hex strings without O's in it.

[mreed2]

I'm pretty sure that all the substantial encoded strings should decode without garbage. Short strings (not more than, say, 5 characters) may just be garbage, but if anything larger than that turns into garbage then its an OCR / transcription failure.

I haven't found any counterexamples, at least. And the person who wrote the guide that decodes most of the strings doesn't look like he/she found any either.

[👾]

Originally posted by mreed2:

I'm pretty sure that all the substantial encoded strings should decode without garbage. Short strings (not more than, say, 5 characters) may just be garbage, but if anything larger than that turns into garbage then its an OCR / transcription failure.

I haven't found any counterexamples, at least. And the person who wrote the guide that decodes most of the strings doesn't look like he/she found any either.

Hmm, I wonder where the corruptions I encountered came from, then. I'm pretty sure it isn't the O's, and that should be the worst offender when it comes to OCR misinterpretation. I'll look at it a little closer when it happens again, and see if I can find it.

[Pup]

I found that OCR struggled a bit with the font on the terminals, seemingly when two characters didnt show a gap between them, such as 'f7' where the top of the f and 7 meet up. Interestingly the OCR in ShareX translated f7 to 0 and the OCR in Powertoys translated f7 to 7 in the example I found.

[👾]

Originally posted by Pup:

I found that OCR struggled a bit with the font on the terminals, seemingly when two characters didnt show a gap between them, such as 'f7' where the top of the f and 7 meet up. Interestingly the OCR in ShareX translated f7 to 0 and the OCR in Powertoys translated f7 to 7.

Ah, in that case every "F7" is suspect, indeed there seems to be almost no space in between even in this font right here. Good find.

[Gargish🐸]

I use Google Lens in text mode, copy paste the numbers into RapidTables.com hex to ascii converter/decoder. General rule seems to be that the hex segments are NOT relevant to the paragraph they are placed in but instead all hex segments in the message have to be stringed together. Corruption can occur if there are several hex segments, sometimes one segment ends with a single extra digit and the second digit for the hex number is the start of the next segment, but if you just paste all segments in the same input field in the decoder then this is automatically fixed.