Some days I am more clever than others. Today, I was working on digitizing about 50 older (30 years old) cassettes for a linguist. To organize the data I have need of creating a folder for each tape. Each folder needs to be sequentially numbered. It is a lot of tedious work - not something I enjoy.
So I looked up a few things in terminal to see if I could speed up the process. I needed to create a few folders so I looked up on hints MacWorld:
So I looked at the mkdir command, which creates new folders or directories. It uses the following syntax: mkdir folder1 folder2 folder3
Now I needed a list of the folders I needed... something like 50.
Now I had a list of 50 names of my folders, but I still needed to remove the return characters which separated them from each other to allow the mkdir command to work. So I opened up TextEdit and did a search for return tabs in the document and deleted them.
Now I could just paste the 50 folder names in terminal and hit enter and it created 50 folders... But I wonder if there was a way to add sequential numbers to a base folder-name in terminal without using google spreadsheets...
As linguistics and language documentation interface with digital humanities there has been a lot of effort to time-align texts and audio/video materials. At one level this is rather trivial to do and has the backing of comercial media processes like subtitles in movies. However, at another level this task is often done in XML for every project (digital corpus curation) slightly differently. At the macro-scale the argument is that if the annotation of the audio is in XML and someone wants to do something else with it, then they can just convert the XML to whatever schema they desire. This is true.
However, one antidotal point that I have not heard in discussion of time aligned texts is specifications for Audio Dominant Text vs. Text Dominant Audio. This may not initially seem very important, so let me explain what I mean. Continue reading →
I have been looking into some
January 4-5, 2012, I had the opportunity to participate in the LSA's Satellite Workshop for Sociolinguistic Archival Preparation in Portland, Oregon. There were a great many things I learned there. So here are only a few thoughts.
Part of the discussion at the workshop was on how we can make corpora which are collected by Sociolinguists available to the larger Sociolinguistic community. In particular the discussion I am referencing revolved around the standardisation of metadata in the corpora. (In the discussion it was established that are two levels of metadata, "event level" and "corpus level".) While OLAC gives us some standardization about the corpus level metadata, the event metadata is still unique to each investigation, and arguably this is necessary. However, it was also pointed out that not all "event level" metadata need to be encoded or tracked uniquely. That is, data like date of recording, name of participants, location of recording, gender (male/female) of participant, can all be regularized across the community.
With the above as preface, it is important to realize that we do need to understand that there are still various kinds of metadata which need to be collected. In the workshop it was acknowledged that the field of language documentation was about 10 years ahead of this community of sociolinguists.What was not well defined in the workshop was what the distinction is between a language documentation corpus and a sociolinguistics corpus. It seems to me as a new practitioner that the chief difference between these two types of corpora is the self identifying quality of researcher. That is does the researcher self-identify as a Sociolinguist or as a Language Documenter. Both types of corpora attempt to get at the vernacular, and both types of corpora collect sociolinguistic facts. It would seem that both corpora are essentially the same (give or take a few metadata attributes). So, I will take an example from the metadata write-up I did for the Meꞌphaa language documentation project. In that project we collected metadata about:
Equipment settings during recording
In the following diagram I illustrate the cross cutting of a corpus with these "kinds" of metadata. The heavier, darker line represents the corpus, while the medium heavy lines represent the "kinds" of metadata. Finally, the lighter lines represent the sub-kinds of metadata, where the sub-kinds might be the latitude, longitude, altitude, datum, country, and place name of the location.
Corpora metadata categories with some sub-categories
This does not mean that the corpus does not also need to be cross cut with these other "sub-kinds". However, these sub-kinds are significantly more in number and will very from project to project. Some of these metadata kinds will be collected in a speaker profile questionnaire. But some of these metadata can only be provided with reflection on the event. To demonstrate the cross cutting of these metadata elements on a corpus I have provided the following diagram. It uses categories which were mentioned in the workshop and is not intended to be comprehensive. In this second diagram, the cross cutting elements might themselves be taxonomies. They may have controlled vocabularies or they may have an open set of possible values, they may also represent a scale.
Taxonomies for social demographics and social dynamics for speakers in corpora
Both of these diagrams tend to illustrate what in this workshop were referred to a "event level" metadata, rather than "corpus level" metadata.
A note on corpus level metadata v.s. descriptive metadata
There is one more thing which I would like to say about "corpus level" metadata. Metadata is often separated out by function. That is what does the metadata allow us to do, or why is the metadata there?
I have been exposed to the following taxonomy of metadata types though course work and in working with photographs and images. These classes of metadata are also similar to those posted by JISC Digital Media as they approach issues with Metadata for digital audio.
Descriptive meta-data: supports discovery, attribution and identification of resources created.
Administrative meta-data: supports management, preservation, and appropriate usage of resources created.
Technical: About the machinery used to create the resource and the technical aspects of the resource.
Use and Rights: Copyright, license and moral ownership of the items.
Structural meta-data: maintains relationships between the parts of complex, multi-part resources (Spanne 2008).
Situational: this is metadata which describes the events around the creation of the work. Asking questions about the social setting, or the precursory events. It follows ideas put forward by Bergqvist (2007).
Use metadata: metadata collected from or about the users themselves (e.g. user annotations, number of people accessing a particular resource)
I think it is only fair to point out to archivist and to librarians that linguists and language documenters do not see a difference between descriptive and non-descriptive metadata in their workflows. That is sometimes we want to search all the corpora by licenses or by a technical attribute. This elevates the these attributes to the function of discovery metadata. It does not remove the function of descriptive metadata from its role in finding things but it does functionally mean that the other metadata is also viable as discovery metadata.
The last couple of weeks I have been working on applying the Zachman's framework for enterprise architecture to two projects. I have been struggling through the first row and then skipped around a bit. I think I have found the part of the project (any project) I am most passionate about.... Working with Human Interface Architecture and explaining it as a designer to the builder of the Presentation Architecture. In my mind this level needs to be closely related to the Business Process Model and to the List of Business Goals/Strategies.
Where do I see myself most helpful in the large project...
One of the things I enjoy is reading about the licenses that CC has retired. Usually they do great job of explaining why they are retiring the license. Understanding these use cases and their context is a really informative view on society.
One interesting retired license is the Sampling+ License. They did a really good job of explaining why they were retiring the license. One of the interesting exercise they talk about was how they had to go through the machine readable description to describe the license — basically mapping out the assertions.
Sound+ is interesting because it is targeted for sound. It makes me wonder if sound/audio can still be licensed under Creative Commons if it is not protected by copyright.
An archival version of an audio file is a file which represents the original sound faithfully. In archiving we want to keep a version of the audio which can be used to make other products and also be used directly itself if needed. This is usually done through PCM. There are several file types which are associated with PCM or RAW uncompressed faithful (to the original signal) digital audio. These are:
Broadcast Wave Format (BWF)One way to understand the difference between audio file formats is understanding how different format are used. One place which has been helpful to me has been the DOBBIN website as they explain their software and how it can change audio from one PCM based format to another.
Each one of these file types has the flexibility to have various kinds of components. i.e. several channels of audio can be in the same file. Or one can have .wav files with different bit depths or sampling rates. But they are each a archive friendly format. Before one says that a file is suitable for archiving simply based on its file format one must also consider things like sample rates, bit depth, embedded metadata, channels in the file, etc. I was introduced to DOBBIN as an application resource for audio archivists by a presentation by Rob Poretti. One additional thing that is worth noting in terms of archival versions of digital audio pertains to born digital materials. Sometimes audio is recored directly to a lossy compressed audio format. It would be entirely appropriate to archive a born-digital filetype based on the content. However it should be noted that in this case the recordings should have been done in a PCM file format.
What is a presentation version? (of an audio file)
A presentation version is a file created with a content use in mind. There are several general characteristics of this kind of file:
It is one that does not retain the whole PCM content.
It is usually designed for a specific application. (Use on a portable device, or personal audio player)
It can be thought of as a derivative product from an original audio or video stream.
In terms of file formats, there is not just one file format which is a presentation format. There are many formats. This is because there are many ways to use audio. For instance there are special audio file types optimized for various kinds of applications like:
3G and WiFi Audio and A/V services
Internet audio for streaming and download
Digital Satellite and Cable
Portable playersA brief look a an explanation by Cube-Tec might help to get the gears moving. It is part of the inspiration for this post.
This means there is a long list of potential audio formats for the presentation form.
Amiga IFF/SVX8/SV16 (iff)
Audio Visual Research (avr)
CDXA, like Video-CD (dat)
Ensoniq PARIS (paf)
FastTracker2 Extended (xi)
Midi Sample dump Format (sds)
Monkey’s Audio (ape/mac)
Mpeg 1&2 container (mpeg/mpg/vob)
Mpeg 4 container (mp4)
Mpeg audio specific (mp2/mp3)
Mpeg video specific (mpgv/mpv/m1v/m2v)
Portable Voice format (pvf)
Sound Designer 2 (sd2)
Windows Media (asf/wma/wmv)
Aside from just the file format difference in media files (.wav vs. .mp3) there are three other differences to be aware of:
Media stream quality variations
Media container formats
Possibilities with embedded metadata
Media stream quality variations
Within the same file type there might be a variation of quality of audio. For instance Mp3 files can have a variable rate encoding or they can have a steady rate of encoding. When they have a steady rate of encoding they can have a High or a low rate of encoding. WAV files can also have a high or a low bit depth and a high or a low sample rate. Some file types can have more channels than others. For instance AAC files can have up to 48 channels where as Mp3 files can only have up to 5.1 channels.
One argument I have heard in favor of saving disk space is to use lossless compression rather than WAV files for archive quality (and as archive version) recordings. As far as archiving is concerned, these lossless compression formats are still product oriented file formats. One thing to realize is that not every file format can hold the same kind of audio. Some formats have limits on the bit depth of the samples they can contain, or they have a limit on the number of audio channels they can have in a file. This is demonstrated in the table below, taken from wikipedia. This is where understanding the relationship between a file format, a file extension and a media container format is really important.
Media container formats can look like file types but they really are containers of file types (think like a folder with an extension). Often they allow for the bundling of audio and video files with metadata and then enable this set of data to act like a single file. On wikipedia there is a really nicecomparison of container formats.
MP4 is one such container format. Apple Lossless data is stored within an MP4 container with the filename extension .m4a – this extension is also used by Apple for AAC audio data in an MP4 container (same container, different audio encoding). However, Apple Lossless is not a variant of AAC (which is a lossy format), but rather a distinct lossless format that uses linear prediction similar to other lossless codecs such as FLAC and Shorten. Files with a .m4a generally do not have a video stream even though MP4 containers can also have a video stream.
MP4 can contain:
Video: MPEG-4 Part 10 (H.264) and MPEG-4 Part 2
Other compression formats are less used: MPEG-2 and MPEG-1
Audio: Advanced Audio Coding (AAC)
Also MPEG-4 Part 3 audio objects, such as Audio Lossless Coding (ALS), Scalable Lossless Coding (SLS), MP3, MPEG-1 Audio Layer II (MP2), MPEG-1 Audio Layer I (MP1), CELP, HVXC (speech), TwinVQ, Text To Speech Interface (TTSI) and Structured Audio Orchestra Language (SAOL)
Other compression formats are less used: Apple Lossless
Subtitles: MPEG-4 Timed Text (also known as 3GPP Timed Text).
Nero Digital uses DVD Video subtitles in MP4 files
This means that an .mp3 file can be contained inside of an .mp4 file. This also means that audio files are not always what they seem to be on the surface. This is why I advocate for an archive of digital files which archives for a digital publishing house to also use technical metadata as discovery metadata. Filetype is not enough to know about a file.
Possibilities with embedded metadata
Audio files also very greatly on what kinds of embedded metadata and metadata formats they support. MPEG-7, BWF and MP4 all support embedded metadata. But this does not mean that audio players in the consumer market or prosumer market respect this embedded metadata. ARSC has in interesting report on the support for embedded metadata in audio recording software. Aside from this disregard for embedded metadata there are various metadata formats which are embedded in different file types, one common type ID3, is popular with .mp3 files. But even ID3 comes in different versions.
In archiving Language and Culture Materials our complete package often includes audio but rarely is just audio. However, understanding the audio components of the complete package help us understand what it needs to look like in the archive. In my experience in working with the Language and Culture Archive most contributors are not aware of the difference between Archival and Presentation versions of audio formats and those who think they do, generally are not aware of the differences in codecs used (sometimes with the same file extension). From the archive’s perspective this is a continual point of user/submitter education. This past week have taken the time to listen to a few presentations by Audio Archivist from the 2011 ARSC convention. These in general show that the kinds of issues that I have been dealing with in the Language and Culture Archive are not unique to our context.
I have recently been reading the blog of Martin Fenner and came upon the article Personal names around the world . His post is in fact a reflection on a W3C paper on Personal Names around the WorldSeveral other reflections are here: http://www.w3.org/International/wiki/Personal_names (same title). This is apparently coming out of the i18n effort and is an effort to help authors and database designers make informed decisions about names on the web.
I read Martin’s post with some interest because in Language Documentation getting someone’s name as a source or for informed consent is very important (from a U.S. context). Working in a archive dealing with language materials, I see lot of names. One of the interesting situations which came to me from an Ecuadorian context was different from what I have seen in the w3.org paper or in the w3.org discussion. The naming convention went like this:
The elder was known by the younger’s name plus a relationship.
My suspicion is that it is a taboo to name the dead. So to avoid possibly naming the dead, the younger was referenced and the the relationship was invoked. This affected me in the archive as I am supposed to note who the speaker is on the recordings. In lue of the speakers name, I have the young son’s first name, who is well known in the community, and is in his 30’s or so, and I have the relationship. So in English this might sound like John’s mother. Now what am I supposed to put in the metadata record for the audio recordings I am cataloging? I do not have a name but I do have a relationship to a known (to the community) person.
I inquired with a literacy consultant who has worked in Ecuador with indigenous people for some years, she informed me that in one context she was working in everyone knew what family line they were from and all the names were derived from that family line by position. It was of such that to call someone by there name was an insult.
It sort of reminds me of this sketch by Fry and Laurie.