Many thanks to Carol and Seth for replying to my questions.

I would be really interested in the committee's view on whether there are compliance issues with this set of questions. If we are writing our audio chunk information to the file level, does it mean we are not compliant?

Many thanks once again.

Pete

-----Original Message-----
From: PREMIS Implementors Group Forum [mailto:[log in to unmask]] On Behalf Of carol chou
Sent: Friday, 18 November 2011 7:59 a.m.
To: [log in to unmask]
Subject: Re: [PIG] Audio chunk as bitstream / boundaries of bitstream

Hi Peter,

In our institution, we take the approach of implementing PREMIS based
on our analysis of the file format from practical point of view.   For
every file we archive, a file object is used to store the general file metadata such as format, checksum, size, etc.  Format specific metadata is placed inside  the objectCharacteristic of either file or bitstream object, depending on the file format.

For WAVE, the audio metadata is placed inside the objectCharacteristic of the file object.  Our understanding is although the WAVE format specification doesn't explicitly prohibit putting multiple audio tracks inside a WAVE file, in practice, we only see an audio stream inside a WAVE file so that's the most practical way of implementing PREMIS from our perspective.  An audio stream is defined as an audio data stream with all the audio metadata.  So even there are various metadata chucks for the audio data stream inside WAVE, it's all considered as one audio stream.  For broadcast wave, though it contains additional chucks, they are considered as metadata chunk thus is not considered as individual audio stream.

For multimedia formats like AVI and MOV, each audio and video streams will be considered as PREMIS bitstreams because we want to have the ability to characterize each video and audio separately.  Same principle goes to TIFF since a TIFF file can contain more than one image and each image can have different characteristic that we want to capture.  So for a TIFF with just one image, the image metadata will still be put into the object characteristic of a PREMIS bitstream.
JPEG image, on the other hand, is put into file object, because a JPEG file contain only one JPEG image.  For PDF, we put document specific metadata in the file level but the metadata for all the bitstreams like images inside the PDF are considered as bitstream objects.

Attached is am example of our PREMIS document for a WAVE file.  We also have a web service on our web site where you can submit a file and generate a PREMIS document for it.  You may access the web service at description.fcla.edu.

Thank you,
-Carol Chou
Florida Center for Library Automation


2011/11/16 Peter McKinney <[log in to unmask]>:
> Dear Listmembers,
>
>
>
> I wonder if any of you might be able to offer advice. I’m getting
> confused with information I should be writing to the file level and
> which to the bitstream level.
>
>
>
> We collect a lot of wave audio files. They are brought into our system
> and run through format ID and metadata extraction (among other
> things). My understanding of PREMIS is that the audio chunk of the
> wave file should be treated as a bitstream. Does this mean that all
> the technical metadata extracted for that wave file should actually be
> written to the bitstream level, as it mostly relates to the audio chunk?
>
>
>
> Likewise, how do I deal with the other chunks? If I’m dealing with,
> say, a Broadcast Wave, are all the chunks bitstream? Meaning, what do
> I do if I want to extract the metadata from them? Do I write them to a
> new bitstream object?
>
>
>
> I guess what I’m trying to get at here is what defines whether you
> have a bitstream or not? Why are the images in a multi-page tiff
> bitstreams, but the single image in a tiff is not (or is it?)? What if
> it is a tiff with one image and exif metadata?
>
>
>
> There’s something in this also around tools and systems capabilities
> as well. The tools we use don’t view files through a PREMIS lens. In
> addition, the systems and tools that we use do not offer unlimited
> flexibility and it is conceivable that we will wind up writing
> conceptually different things to the same PREMIS level as a result.
> The audio chunk of a Wave file would be described as a bitstream if
> the file is ingested on its own but the entire file will be considered
> a bitstream if it is encapsulated in a WARC or Baggit file.
>
>
>
> I’d be really grateful for any thoughts on this topic, especially if
> those thoughts contain examples!
>
>
>
> Thanks,
>
> Pete
>
>
>
>
>
> Peter McKinney
>
> Digital Preservation Policy Analyst
>
> Preservation Research and Consultancy
>
> National Library of New Zealand Te Puna Mātauranga o Aotearoa The
> Department of Internal Affairs Te Tari Taiwhenua PO Box 12-340,
> Wellington 6144
> 77 Thorndon Quay, Thorndon, Wellington 6011
>
> Direct Dial: +64 (4) 462 3931
>
> www.dia.govt.nz
>
>
>
> ====
> CAUTION:  This email message and any attachments contain information
> that may be confidential and may be LEGALLY PRIVILEGED. If you are not
> the intended recipient, any use, disclosure or copying of this message
> or attachments is strictly prohibited. If you have received this email
> message in error please notify us immediately and erase all copies of
> the message and attachments. Thank you.
> ====

====
CAUTION:  This email message and any attachments contain information that may be confidential and may be LEGALLY PRIVILEGED. If you are not the intended recipient, any use, disclosure or copying of this message or attachments is strictly prohibited. If you have received this email message in error please notify us immediately and erase all copies of the message and attachments. Thank you.
====