Researchers are increasingly using video in their fieldwork. Starting with cheap analogue formats and now digital formats, it is easy and affordable to begin video-taping everything… In the same way that we can now record audio for everything.
…Well, actually I’m not quite convinced yet.
I’ve got some ideas about what to do in future field trips, but I’m wondering what everyone else out there does. I found my video-taped sessions extremely useful, especially in eliciting spatially oriented material. However, I’m a bit unhappy with what’s currently out there in terms of technology.
For a start, MiniDV, which is what most people seem to be using is really what I would call a pseudo-digital format. Although the signal is digitised at the recording time, it is still recorded onto tape media. Which means that while its on the tape you can’t “jump around” quickly. Secondly the date-stamp and time-stamp metadata seems to be inconsistently added from recorder to recorder. I’ve yet to see anyone pay the extra cash for the CM type miniDV tapes either, but that is supposed to store extra metadata. I haven’t seen auto-segmenting of video files work consistently, and in dealing with a few older tapes recently, I’ve realised the tape deteriorates quite quickly (digital glitches look terrible as well!).
I guess to be fair, digital video produces media files many times bigger than high quality audio recorders, and storage isn’t that cheap. Tape is a nice cheap way of recording lots of data (so is compression, which is used in all consumer level formats).
Coming to a decision about the “necessary” quality of video recordings is much harder than the case for audio. While high quality audio is justifiable, there’s no “acoustic analysis” equivalent argument for video… visual quality seems to me to be a lot more subjective. I thought DVDs looked amazing compared to VHS when they were first released. But now DVDs look shocking compared to the newer standards. So setting minimum standards is a bit more complex, especially considering that storage capacity hasn’t caught up to what’s needed for raw digital video yet. Its a mixture of video resolution, frame-rate and compression to reduce the size.
What I’d really like to see is a digital video recorder that works exactly like a digital camera. First of all it segments recordings into separate files. Breaking up a video dump into separate resources is just more work for the fieldworker. These files should have good metadata like the hardware model and make, CCD data, date and time stamps. More difficult would be exposure statistics and so on, these would have to be by frame presumably… which I’m guessing would be harder to implement. Some time in the future I imagine that a digital camera that can shoot photos fast enough to be a video will be possible (well, it kind of is possible at the moment, but at reduced resolution, and without per frame metadata). Good metadata is critical for speedy research. If you’ve got simultaneous audio and video recordings they should show up together in a file-system listing sorted by creation date!
Newer video recorders seem to be flooding the market now. There are consumer brand Digital Video recorders that record straight to an internal hard disk or DVD burner. I’d be interested to hear from anyone that has used these to see what kind of files they produce.
I’m also interested in the cheaper video recorders that go straight to flash cards. These do seem to use lower than PAL/NSTC resolution, with heavy compression, but as a trade-off you could instead afford to purchase a couple to record from two or more viewpoints.
Perhaps the best thing about these pure digital recorders is that they bypass the 60 or 90 minute limit of recordings. How many times have you just wanted to start the recorder and walk away for a couple of hours?
I’ve been toying with the idea of using a high quality web-cam hooked up to a laptop too… this would let you record continuously for hours, and you could plug in more than one web-cam too.
Anyhow, what is your experience? Have you mucked around with any of these newer recorders?
Follow
I haven’t tried the newer recorders but I think there is a more fundamental question that a field worker needs to begin with. You need to know that you or your institution has adequate ‘space’ to store and back up files made from DV tapes. Otherwise you can be stuck with transcribing from the master DV tapes, and have no backup save VHS copies. Our ACLA project (Aboriginal Child Language) has server space in Canberra on APAC where we store uncompressed mov files. I think this is a bit unusual though. For example, all of the DOBES projects have their files stored at the MPI Holland, but only as compressed MPEG files (as I understand it).
And as well as space, an easy and fast way of downloading the stored backed-up video.
Yes, good point. While working with gigabyte sized audio files is annoying, working with video files of several gigabytes can be a total pain. I should mention that PARADISEC is an option for space for pacific region projects… sorry Australian projects.
At PARADISEC we’ve been trialing video ingestion recently, and its caused a few hiccups in our workflow because of the massive amounts of data generated. before the trial, several years of ingestion left us with an archive of over 1.8 terabytes of data. After the trial (and in the space of a couple of months) our archive has now passed the 2.5 terabyte mark!
One simple problem is that it takes a lot of time to move the files around! Let alone transcode them to manageable formats. And being an archive, we’re first an foremost interested in the best quality format for archival (which is almost always the biggest – We’re using mpeg2 as a delivery format, but not for archive). Moving half a terabyte of data around, from disk to disk or over the network takes a long time!
The problem with tape based recorders is that it takes at least the playback time of the recording to get it on your computer, when really what you want on your computer in the field is a small copy for transcription and mark up.
Apparently ELAN supports transcription directly from a DVD, so perhaps that is the most manageable way of working with video for now? Using one of these direct-to-dvd recorders might be the go? The only problem is that they use smaller dvd media which doesn’t work with the slot loading dvd drives I know a lot of fieldworkers have on their laptops…
The flash card recorders compress the video they record so the files are a fair bit smaller, but the price you pay is a loss in image quality.
Either way I think these new video recorders look quite interesting compared to MiniDV recorders. I’ve got to hunt one down and see what the quality, metadata etc etc is like!
Claire Bowern has just posted a long response to Tom on her blog anggarrgoon.
Thanks for posting this Jane – the trackback didn’t work for some reason.
Sorry to be joining a thread that’s gone cold – I wasn’t listening a month back! But I have a couple of comments.
1. One part of processing which we need to make a little extra time for but which hasn’t been mentioned in the discussion so far is extracting uncompressed audio before compressing the whole video stream to whatever format we want to use for storage and editing. This is an important step for MiniDV anyway – I haven’t used any other formats yet so I can’t comment on them. It doesn’t take very long, but it is one more task in the workflow!
2. My experience in Eastern Indonesia has been that people are more comfortable around the video camera than they are around my audio gear. My audio setup is not that intrusive, comparable in size to the camera, but people were generally very comfortable with the camera and less so with audio. I think that this is because HandyCams are already familiar for most people, but microphones are not. And now I have revealed that I was not using an external mike on the camera – perhaps I would have had a different experience in that case.
With audio extraction I tend to include that as part of a single stage transcoding from MiniDV to every format. The “latest” workflow I’ve come up with is using Compressor, which is part of Final Cut Pro to batch convert files for multiple outputs. I “digitise” the MiniDV tapes during the day and then set up a batch conversion over night to convert to mpeg2, wav, mp4 and sometimes mpeg1 (the most common format I could see in PNG was VCD, not DVD).
We’re hoping that we can shift some of this conversion stuff over to APAC, especially for conversion to ogg theora for use with annodex. But then that may not work well for a fieldworker who needs the files to work with fairly quickly.
That’s very interesting re audio versus video impact. What was the sound from the camera like? I think that’s kinda why Minidisc was popular for a while (note, I am not advocating the use of minidisc!). They were so inconspicuous.