Using a Pocket Audio Recorder with Linux?
"The Olympus series of digital voice recorders seem to be the most popular hardware in this category, and they're cheap, but they write files in a proprietary DSS ("Digital Speech Standard") format that can't be converted under Linux. It looks like people have been wondering about this format for years, but as of today it remains inaccessible for free software users.
Solutions I've come up with are (1) to get a pocket MP3 player that can also record, like the MPIO DMG MP3 Player (expensive for my task, though, at $200); and (2) the handheld multi-track recorder, the Korg PXR-4. An ideal setup would have a headphone or speaker jack for playback, and a mic/line input so that it could be used to make digital recordings from another hardware source (like TV or radio), and used with special microphones like a parabolic mike."
I've got an IRiver MP3 player. It works with linux when its running the UMS firmware. The mic works pretty well. I use it to record meetings when my other group members can't attend meetings. Maybe be more costly than you're looking for and the storage isn't removable.
i could not think of anything clever.
... I'd suggest a PocketPC. I have one and I use it for voice dictation all the time. It has a record button that starts a record even when the unit is off. It's sort of like talking into a walkie talkie. When it's done, you get a nice little .WAV file. I like this because it's easy to organize short notes. I can rename the file to something more descriptive and file it away under whatever folder is appropriate. It's very handy.
However, Microsoft did not get everything right. *grr* They use a compression format that Windows doesn't by default have a codec for. (That codec is available for download, though.) Also, ActiveSync backs up everything BUT the sound notes. (note: this may have been fixed in the most recent PocketPC OS, Mine's 2 years old.)
Also, I'm not sure you can get that running in Linux. Maybe a Palm Pilot with voice capability would suit your needs? Or a Zaurus? (Uhh I don't think the one out now has that capability, but I think the next one does...)
It may be a little spendier than you'd want to do, but the neat thing about using a PDA for a voice recorder is that you don't necessarily need to have your computer around for it.
"Derp de derp."
My Archos JB20 Recorder works very well for this, especially with The Rockbox software. (Disclaimer: I'm on the Rockbox team.)
It is a little pricey ($175-$220US on eBay), but with everything else it does - MP3 recording and playback, text file viewing, bookmarking, and a host of others, it's well worth it. You can also get 10GB models for a little less.
The filesystem is FAT32 and it connects seamlessly to most OSs which have USB support. Most of the Rockbox team develops on Linux.
So while it may be a little more than you wanted to spend, it's a worthwhile investment!
Mattcelt
...as far as field digital audio recording goes in MiniDisc. Get yourself a half-decent MD player and a nice mic with a 1/8" jack and you're good to go.
Matthew G P Coe
http://mgpcoe.blogspot.com/
You'd have the benefit of being able to make badass gangsta rap from meetings notes. Imagine the joy of having a remix of your coworkers set to a drumbeat. You could go all strongbad and throw in the "meeley! meeley! meeley! meeley!" of a guitar solo along with the "DUH! DUH! DUH! DUH! GRRRRRRRRRR!!!!!!" of the rythem guitar along with the "BOOM BOOM BOOM BOOM BOOM BOOM BOOM BOOM doot doot doot doot dooty doot doot" of trance/dance. You'd be a hit at office parties. Ted from accounting would love it.
For every annoying gentoo user, are three even more annoying anti-gentoo crybabies. Take Yosh from #Gimp for example.
Second, just because an MP3 device happens to have a record function and a cheap built-in mic, doesn't mean it's a decent voice recorder. How good is the encoding firmware? What's the sampling rate? Is the mic a decent one that won't break down, or just some piece of crap they threw in to make the product more marketable?
As for Linux compatibility: Sony makes a couple of voice recorders that record to a memory stick. If you have a memory stick reader supported by Linux (there are several), you can mount the stick as a VFAT disk, and access the individual voice files directly. But that still leaves you with the problem of decoding Sony's proprietary LPEC format.
I'm not sure you can avoid the proprietary format problem, not if you want decent recording quality. Now, getting proprietary-format voice files into a standard format is a pain even under Windows. The problem is that vendors seem to like to embed the conversion software in the device driver -- which never seems to have been fully tested on all likely system configurations.
But in Sony devices, the conversion is embedded in a user-level playback program, which doesn't have all the low-level dependencies of a device driver. Possibly that program will run under Wine.
The Neuros is currently deeply discounted in preparation for the newer models coming out...
The 128MB unit is going for $150, the 20GB model is on sale for $200 and the combo bundle (128MB/20GB, etc.) is currently only $230.
I bought my combo set for $343, and that was after both a $50 sale discount & a $25 Linux user's discount.
I don't mean to sound like a salesman, but this is a great deal!
This user account is inactive account replaced by the PDA
As already mentioned by Mattcelt, an Archos (without or preferably with Rockbox software) will do it, and will record quite a bit, given that it's recording to a hard drive.
A Sharp Zaurus will also do it, recording to a CF or MMC card.
I've got both a Zaurus and An Archos, and will mention that the Zaurus internal microphone may not be superior to the Archos internal microphone, but the Archos internal will pick up the (intermittant) sounds of the hard drive spinning.
The Zaurus, of course, uses a Linux file system; the Archos with Rocbox a FAT32 filesystem. The Archos records to mp3; I've forgotten what the Zaurus records to. Both allow you to trade audio quality for file size, however.
The Zaurus allows one-touch recording in its defualt out-of-the-box configuration, which can be handy if you plan on surreptitious or spur-of-the-moment recording. However, it's also very easy to inadvertantly record, and find you've used up most of your RAM-disk doing so.
Also, like Mattcelt, I'm (pereferally) on the Rockbox team (or at least mentioned in the credits).
Opinions on the Twiddler2 hand-held keyboard?
I think a couple other folks have mentioned it, but I would look into a PDA. A lot of them have SD slots.
Palm
A PalmOS PDA would probably be the best option from a number of standpoints (e.g., it just works usually), but I don't know how hard it would be to find an app that would record to a format you could get at easily from your desktop. I know that a Palm device will leave an SD card in the usual FAT-based format that you buy them in, which would be readable by Linux. However, even if you tell the app to record to the SD card (rather than internal memory), I imagine it'll be in a Palm database rather than a flat mp3 or ogg file.
Zaurus
I'm sure a couple folks are plugging the Zaurus, because it runs Linux. All of the Linux Zaurus models do have a SD card slot, but only one of them- the SL-5600- has a built-in mic, which is pretty crappy. The SL-5600 is pretty expensive, going for $450-500 depending on where you get it. Quite a bit more than a PDA with similar hardware and capabilities should cost. The Linux Tax, I guess. If you're willing to have a potentially fragile microphone hanging out of the dual mic/headphone jack, you could always use a hack. That way, you could get one of the cheaper SL-5500 models- up to $300 cheaper than the SL-5600 for a machine that is about as good. E.g., the CPU in the 5600 is 400 MHz, but only a very, very wee bit faster (if at all) than the 206 MHz CPU in the 5500. (think of a 100 MHz 486 vs 100 MHz Pentium)
The built-in software records as WAV files, but since it runs Linux, I'm sure you could port an mp3 recorder if you're willing to invest the time in writing code for a Qtopia GUI, or some other kind of interface that doesn't require you to type something to record a note.
WinCE/PocketPC
The third real option is a PocketPC device. A lot of Linux weenies are too bigoted to consider it, but if you're just looking for a solution to your problem you shouldn't overlook it.
One I'd reccomend is the Dell Axim X5 Basic, which can be had for around $200. Specs like you'd expect- 300 MHz CPU, 32 MB RAM (being increased in a new model, not sure if its out yet), and dual SD and CF slots. Since it writes to the regular FAT filesystem that comes on SD and CF cards, it's just as easily read on your Linux box. There are a couple software options for recording mp3s, and a million for recording WAV files.
Working toward a usable PDA environment in the spirit of Newton OS: Dynapad
Clearly this is not the ideal solution, but perhaps it strikes a balance between your needs and ideals.
Consider picking the device based on its stand-alone features, then upload the recording via the line-in on your sound card.
Of course this won't work if time is an issue, but maybe it would be workable for you to just hit play and go to bed.
Anyway, once you do that you can use a sensible, open codec like Speex.
-Peter
I got my 5500 for about $200 on ebay. www.theKompany.com offers a software program called tkcVox. From the site: # It's possible to record approximately 24 hours worth of non-stop speech onto a 128mb storage card. # Uses GMS-AMR codec to provide the highest quality of compressed speech. This is 3rd generation mobile telephony technology used for speech compression. It provides for high quality voice recording (depending on microphone) with a bitrate of 4.75 to 12.2 kbps for 1 hour of continuous speech recording in only 2-5mb of space. # Voice Activity Detection (VAD) and Comfort Noise Generation (CNG) algorithms optimizes the dead space when talking so that it takes almost zero storage space. http://www.thekompany.com/embedded/tkcvox/ The program costs a very reasonable 9.99. This will require an external microphone but this could turn out to be an advantage since you could do hands-free work. Also, if you had a mobile phone ear piece you could also use their tkcPhone to place telephone calls from your Zaurus via WiFi. Russell
I own a Diva Mp3 player, which can also record. It also works with linux, be it a little flakey. With a cardreader, it works like a charm tho.
:) Highly recommended, at 99 euros.
It works with compact flash cards. It stores about 10MB of voice per hour. It runs about 10 hours on one AA battery, which is enough for most interviews
the pun is mightier than the sword
I've been meaning to visit my local university's library to resolve this, but since this Ask Slashdot has come up I might as well tell others what I have figured out so far.
Digital Speech Standard (DSS) is not necessarily proprietary in the sense that you or I cannot lay our hands on it. It might not be open source, but it can possibly be available for licensing from the International Voice Association (IVA) on terms that might be practically useful (see below for more on this). IVA is a joint venture between Grundig, Phillips, and Olympus. The reason I need to go to the library is because the IVA is apparently not on the Net. Go figure.
Quoting Olympus as an authoritative source:
I emailed Olympus USA Technical Support and asked them to direct me to someone who can license the spec to me, but predictably nothing useful came out of that:
Three guesses what the result was when I called a Customer Support monkey.
This guy figured out some basic encoding characteristics of the Olympus DS-3000, the model that I use. An FAQ from a New Zealand distributor lists some other characteristics like the compression ratio of the format (approximately 12:1 over WAV).
Note that apparently the encoding algorithm can only be licensed for use on hardware devices, but decoding is allowed in software. This is what I meant when I mentioned above that it might be possible to license the format on practically useful terms. For my project where I want a Linux-based DSS decoder, I don't need software encoding, so this works for me. I found this in a link on the Net somewhere, but I can't Google it up now.
This third-party product can decode DSS format files, so there is the proof that we need that the format can indeed be decoded by someone outside of the IVA founding members. This Windows-only shareware program performs all sorts of handy audio format batch conversions, but it requires the Windows-only DSS Player in order to decode DSS format files. Hopefully that means DSS Player (which can only be obtained/licensed through the purchase of an Olympus DSS recorder) can be manipulated through Windows Dynamic Data Exchange (DDE) in a worst case scenario (running DSS Player from a VMware instance on Linux, and communicate between the two using some socket-based protocol).
I have an applet for that. Just tap it and start talking, the result is a wav. A real wav. none of that proprietary garbage.
/dev/dsp for high quality recordings.
What many people forget is that the Zaurus is a computer, you can do anything on it that you could on a desktop with similar resources. You could even oggenc
You can't judge a book by the way it wears its hair.