Non-stop audio diary & Memory enhancing technologies

No one can argue that technology has changed how humans think and behave in a very fundamental way. A very simple example is how easy cell phones have made it to meet up with someone while you are out. Before the cell phone you had to carefully plan where and when you would meet while nowadays you just use the cell when it’s time. This may sound like a small thing, but it’s a big change in how you think about planning and also in required personal responsibility.

Another area that I think will make a big impact is forms of memory enhancing technology. People have used PDAs with calendars for a while, but that’s just the beginning. I’ve recently been cleaning my e-mail archive because I’m switching computers. The e-mail data-base dates back to 1997, and basically contains most of my conversations I’ve had over the computer. This data-base together with Google Desktop is a very powerful tool to help me remember my dealings with persons around me, where I can with a quick search find out why that name seemed familiar to me, or who to contact at a specific company. I was abruptly reminded of how important this was to me when my mail client decided to ignore writing actual data to the files I exported, leaving me with a blank data-base for the period of April 2006 to yesterday. Apart from the frustration and having a hard time answering some of the mails I’ve recently received, it’s also a fundamental feeling of loss. I’ve realized that that data-base has become something I’ve relied on to help me remember things. Why write things down anywhere if you know you can access them so easily?

It also got me thinking on the future of such memory enhancing devices. One project I’ve been planning to do but am still waiting for the technology to mature is a non-stop audio diary. Basically I want to have a digital audio recorded on my person at all times, recording everything that goes on around me. The use of that data may be limited for a while longer, but it’s not unreasonable to think that within it is possible to automatically transcribe audio to text with good quality in ten years. Suddenly you have a searchable data-base of everything you’ve said and what’s been said to you. Imagine someone telling you a telephone number and by just hearing it, you can be certain that that information will be available when you need it. A friend of mine had an even better example of how that technology could be used:
– Just think of think of how great it would be when you are fighting with your girlfriend and she says “But you should have told me you are going to be with your friends tonight!” and I could prove to her that I did so.

While that was mostly for fun there is a big point in there. A lot of conflict in the world is caused by miscommunication and subjective perception. That would potentially change if this technology was common-place.

I believe the technology is almost ready for me to start my personal recording. The problem is to find a recorder that can manage to record up to 24 hours on one battery-charge, but if anyone has any suggestions about what hardware would be good for such a project, please let me know.

5 thoughts on “Non-stop audio diary & Memory enhancing technologies

  1. Bakavic,

    Thanks a lot for those links. It’s really interesting to see that someone has had the same idea. I’d love to get something similar up and running, so I’ll definitly look into that further. Also, I’ve been trying out the same thing with a GPS which would make it even cooler. :)

  2. I know this post was forever ago, but have you made any progress in your documenting life idea? I have EXACTLY the same idea and it’s rather challenging to find people serious about it.

  3. Igl007,

    Glad to hear it. Yeah, very few people “get” why that would be a cool thing. I haven’t done anything with this idea, but the last couple of years have been quite interesting in terms of technical development allowing for stuff like this to be done more easily. I’ve seen that there now exist “always on” GPS devices that can record your location to a memory-card which also are long the same lines. Regarding sound-recording. I got this fantastic response from a guy called Gristle McThornbody, who was the author of the links in Bakavic’s post above. Sadly his site seems to be gone, so I’m taking the liberty of reposting some information he sent me in an e-mail a couple of years ago:

    “In any case, I will describe where my REWM solution currently stands and how this project got where it is. I am an EE with vast experience hacking together very bad perl code. I realize that I should try writing “good” code some time, but never seem to get out of hacker mode due to time constraints and what can only be described as TADD (Technology Attention Deficit Disorder). Projects just get boring very quickly, so I know I need to finish them quickly or they don’t get done. My unfortunate moto seems to be: If you want it done right, it won’t get done.

    The REWM project though seems to morph (in a good way) from time to time, because I am constantly using it. I can make small improvements from time to time and over the past few years it has become easier and easier to use. It’s still clunky to use from a commercial standpoint, but I know the quirks and can easily live with them.

    The first improvement I found was buying top-of-the-line rechargable batteries. They usually come in packs of four and I’ve been using the same set of 4 for 3 years now. This eliminated the need for an external battery pack. One charge can usually last 1 full day depending on how long you actually are recording. I find myself only recording about 1/2 of my waking hours now though, so results may vary.

    I record less time per day now for a variety of reasons. First is battery life. It allows me to use a standard s1mp3 device without the battery mod. Second, the s1mp3 crashes sometimes and looses whatever recording it was in the middle of, so I find myself stopping frequently to commit the recording to memory (save often). Once committed the recording will survive a crash. Third, I have tried to sift through old REWM recordings to find conversations and have found it easier to navigate multiple files with regular timestamps than one huge file representing the whole day. This make potty-breaks or long cubie-sessions the perfect time to commit the recording and save some bits.

    When I saw that this method was viable and developed the first version of the code, I realized that my solution was tied to the s1mp3 player architecture. I knew the s1mp3 wouldn’t be around for very long in its current incarnation, so I ordered about 10 of them at the time to make sure I always had working hardware that I knew was compatible with my code. The players were cheap and 3 of them have given out already, but I still have a viable solution for years to come. When I do run out of players though I will need to start from scratch. That is unless a REWM product is available commercially by then.

    I have looked on eBay (where I got the original players) to see what is currently available and don’t seem many s1mp3 players there. The ones I do see are 1GB and much more expensive than the ones I bought (128MB). Disapointingly, I don’t really see any voice recording technology today that is as cheap as the ones I bought with the capacity of the ones I bought.

    I can’t tell you though how much of an advantage it is to not have to worry too much about smashing the device (knowing that you have spares). When you wear the thing every day it WILL wear out quickly. I have 2 small children so I have a hard time wearing anything valuable on my person.

    So as far as a recommendation, I don’t really have one. I’d just find the cheapest device that can record in a compressed format (not just WAV) and appears as a simple USB drive when attached to the PC. If you’re only using the device to record, 1GB of MP3 voice recording is around 200 hours. Don’t worry about getting something with large capacity. You’ll wind up synch’ing more often than 100 hours most likely. Also, don’t worry about finding a platform that your code will work with long term (my mistake) … 4 years from now everything will be different. You’re better off just trying to make you code as portable as possible.

    If the device doesn’t record directly to MP3, I’d do a little research to see how hard it is to convert their proprietary format (usually ACT) to MP3. I would warn against archiving you files in a proprietary format because a converter might not be available in 5 years when you what to listen to them. The s1mp3 records long format as ACT files and my code converts it into MP3 for archival. I think there’s a licensing issue with HW encoding of MP3 that keeps most manufacturers from doing it. It’s a nuisence.

    There weren’t really any s1mp3 tools available when I started this project so I did not use the SDK. I use the ACT to WAV converter that is in the public domain and then use LAME3 (also public domain) to convert the WAV to MP3. My code does very little other than coordination of other SW and moving around of files. My code just goes the the flash drive directly and has no context of the s1mp3 as anything other than another drive to read from.

    I hope this helps. I would love to hear the details of your implementation when you have one. Maybe I could even jump over to your REWM solution when I run out of players.”

    I was very excited when I got this mail, but other projects got in the way and I never pursued it. If you are really intrested in this stuff, I recommend trying to track this guy down, as he has actually pulled it off.

  4. By the way. I recently got a Android phone, and it looks like that OS have the ability to run services. It seems fairly easy to write an application that turns on the mic and records everything, and since it’s a phone, it could actually automatically upload the audio to a server at certain intervals (or whenever Wi-Fi is connected). I should look into that… :)

Leave a Reply

Your email address will not be published. Required fields are marked *