[From nobody Mon Jan 21 08:15:41 2008 Envelope-to: lou@herald.ox.ac.uk Delivery-date: Fri, 18 Jan 2008 16:33:18 +0000 Received: from rx2.herald.ox.ac.uk ([163.1.0.236]) by imap211.herald.ox.ac.uk with esmtp (Exim 4.63) (envelope-from <lou.burnard@oucs.ox.ac.uk>) id 1JFu9e-0006th-TJ for lou@herald.ox.ac.uk; Fri, 18 Jan 2008 16:33:18 +0000 Received: from relay8.mail.ox.ac.uk (relay8.mail.ox.ac.uk [129.67.1.171]) by rx2.herald.ox.ac.uk (Postfix) with ESMTP id C132638045 for <lou@herald.ox.ac.uk>; Fri, 18 Jan 2008 16:33:18 +0000 (GMT) Received: from j3.village.virginia.edu ([128.143.20.247]) by relay8.mail.ox.ac.uk with esmtp (Exim 4.68) (envelope-from <lou.burnard@oucs.ox.ac.uk>) id 1JFu9d-0000FH-Si for lou.burnard@oucs.ox.ac.uk; Fri, 18 Jan 2008 16:33:18 +0000 Received: from relay0.mail.ox.ac.uk (relay0.mail.ox.ac.uk [129.67.1.161]) by j3.village.Virginia.EDU (Postfix) with ESMTP id 2A343442A9 for <editors@tei-c.org>; Fri, 18 Jan 2008 11:33:17 -0500 (EST) Received: from smtp1.mail.ox.ac.uk ([129.67.1.207]) by relay0.mail.ox.ac.uk with esmtp (Exim 4.68) (envelope-from <lou.burnard@oucs.ox.ac.uk>) id 1JFu9c-0004i1-1L; Fri, 18 Jan 2008 16:33:16 +0000 Received: from cpc3-oxfd2-0-0-cust210.oxfd.cable.ntl.com ([82.10.96.211] helo=[192.168.1.101]) by smtp1.mail.ox.ac.uk with esmtpsa (TLSv1:AES256-SHA:256) (Exim 4.68) (envelope-from <lou.burnard@oucs.ox.ac.uk>) id 1JFu9c-0002XZ-3I; Fri, 18 Jan 2008 16:33:16 +0000 Message-ID: <4790D4CB.2020800@oucs.ox.ac.uk> Date: Fri, 18 Jan 2008 16:33:15 +0000 From: "Lou's Laptop" <lou.burnard@oucs.ox.ac.uk> User-Agent: Thunderbird 2.0.0.6 (X11/20071022) MIME-Version: 1.0 To: Brett Zamir <brettz9@yahoo.com> CC: editors@tei-c.org Subject: Re: Chapter 8 - Transcriptions of Speech References: <478ECB5D.3030700@yahoo.com> In-Reply-To: <478ECB5D.3030700@yahoo.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Oxford-Username: lou X-Oxmail-Spam-Status: score=0.0 tests=none X-Oxmail-Spam-Level: Brett Zamir wrote: > > A few general ideas...These are not necessary, but I think they are > interesting for at least opening peoples' minds to the potential range > of uses for TEI... > > 1) Just curious whether mention might be made of the potential to > transcribe songs? They would probably require their own markup > language for some aspects of their expression. I think most users of this module would regard singing as a shift in vocal quality when it appears within a conversation (that's how I've seen it treated anyway). I don't think you'd use this module for transcribing (say) an opera or a musical performance -- as you say, that would require an awful lot more detail. And there is a lot (e.g. <event> or <kinesic>) that you *wouldnt* want in e.g. a transcription of an opera! > > I could only find the following (that was still online) which also > dealt with lyrics (something which would be perhaps most relevant to > also use with TEI): > http://sourceforge.net/search/index.php?words=%28%2Bmusic+%2Bxml+%2Blyrics%29&type_of_search=soft&pmode=0&words=%28%2Bxml+%2Blyrics%29&Search=Search > I had a look at one or two of these. They seem a long way away from the TEI kind of standardization. > 2) Likewise, maybe one could make mention of the potential to > transcribe signed languages (not only "spoken" language)... > Yes, someone else suggested this; I have no objection to this in principle, but I am at a loss as to how one represents signed language in XML! It's definitely a job for a future release of TEI... > 3) And while I'm trying to bring in the whole spectrum (npi) of sound, > how about a mention of SMIL? This would also be appropriate for > <writing> (e.g., to indicate more exact timing than @gradual). > SMIL is mentioned elsewhere in the Guidelines, tho not in any great detail (http://www.tei-c.org/release/doc/tei-p5-doc/en/html/FT.html) > Such specialized XML languages might, I would think, be able (for > those who wished to extend TEI with them) to replace the functionality > of such elements as <shift>, as they could be more exact in > representing the features. > That's correct. But this module is only intended for simple applications. > *8.3 Elements Unique to Spoken Texts* > > 1) Maybe this line, "we regard the <gi>u</gi> element as analogous to > a paragraph, and the others as analogous to > <soCalled>phrase</soCalled> elements" ought to have a parenthetical > statement to indicate "(though unlike <phr> these elements can appear > independently)"? > Good point. Have added text. > 2) There is this line "The <gi>seg</gi> element has also been used to > segment the last utterance", but there is no <seg> in the example. > Ooops. Added <seg>s to example (dont know how they got lost...) > *8.3.1 - Utterances* > > In this section, might it perhaps be suggested to use @next/@prev to > connect interrupted noncontiguous utterances? > Possibly, but these attributes are not currently available on <u>, and I am not quite sure what it would mean if they were. a <u> is defined as a continguous strectch of *uninterrupted* speech, after all (modulo things like <event>s, or overlap phenomena) > *8.3.3 Vocal, Kinesic, Incident* > > "The <att>who</att> attribute ... must be supplied for a vocal, > kinesic, or incident which is not contained within an utterance." How > about for cases where it is not known? > You have to define an "unknown speaker". I've expanded the discussion of the who attribute at the para beginning "Use of the who attribute..." to mention this explicitly. > *8.4.1 Segmentation* > > "The <gi>s</gi> element is available without formality in all texts, > but does not allow segments to nest within each other." What does this > mean here? Both <seg> and <s> can occur within an <s>, according to > the documentation on <s> at > http://tei.oucs.ox.ac.uk/P5/Guidelines-web/en/html/ref-s.html (though > I see <quote> and <q> are not allowed). The docs for <seg> at > http://tei.oucs.ox.ac.uk/P5/Guidelines-web/en/html/ref-seg.html also > make mention that <seg> can include other <seg>'s. This is unfinished business. The distinction between <s> and <seg> is that the former is a specialcase of the latter, which is intended for use in end-to-end non-nesting segmentation only. It says this at various places in the Guidelines. Unfortunately, in P4 and all subsequent XML versions, this constraint is not enforced by the schema. It ought to be enforced by a schematron rule, but the rule has yet to be written. > > *8.4.4 Prosody* > > This line, "The two parallel streams may be aligned with each other > and with other streams, for example an acoustic encoding", while I > know it is referring to markup which approximates the acoustics, I was > thinking whether something like the 'x' and 'y' coordinates and links > to a graphic image might have their parallel with a linked raw sound file. > Yes, I think they do, and indeed we propose very similar mechanisms for doing sound/sound alignment, graphic/text alignment and others, as you will see when you get to reading the chapter on linking and alignment! > *8.4.5 Speech Management* > > There is this code: <gap reason="passing truck" extent="10" > unit="syllables"/> > > The rules for the relative usage of <gap>, <del>, etc. in section > 11.5.2 state "where the *text has been rendered completely illegible > by deletion or damage and no text is supplied by the editor in place > of what is lost:* place an empty gap element at the point of deletion > or damage." > > How can one know that a passing truck intervenes for 10 syllables if > <gap> has such a meaning? > Are you saying that "syllables" implies you can at least hear the syllables, so this probably shouldn't be a gap? Or are you saying that the definition for <gap> doesn't seem to permit its usage for a gap in transcription? I have changed "syllables" to "secs" to deal with the former point. If the latter -- the description does say "because the material is illegible or inaudible". I have chosen to standardise on "back-channel" rather than "backchannel" as your patch suggests, since the hyphenated form appears more common in linguistics see eg http://www.cs.utep.edu/nigel/bc/ I ]