Talk:Electromagnetic articulography/GA1

GA Review
The edit link for this section can be used to add comments to the review.''

Reviewer: Spinningspark (talk · contribs) 22:45, 12 November 2015 (UTC)

I will do a full review tomorrow, but my immediate impression was that it is a little short. Half of what is there is a review of other methods, there is not all that much actual meat. I am wondering if this is going to meet requiremnent 3a "addresses the main aspects", so I took a quick look at what this source covered. Major headings they have that don't seem to be covered in this article include, As I say, I will take a closer look tomorrow, but I would be grateful if you respond to this comment in the meantime. SpinningSpark 22:45, 12 November 2015 (UTC)
 * Sources of error
 * Interference with subjects' articulation
 * Safety
 * Thanks for taking the review on! The source you linked seems to be a very good one and I put in a request for it at the university library but it will be a few days until I have access to it. Error is covered but not with its own heading (it probably wouldn't be long enough per MOS:Paragraphs). Error is covered in the principles section: "However modern two-dimensional systems are still unable to compensate for tilting of sensors past 30 degrees and measurement is distorted if sensor coils are moved off the centerline of the mouth."
 * Coarticulation is actually the first source I've seen to discuss the safety aspects of EMA, and it seems to engage largely in speculation. It points out that there is no conclusive evidence that long term exposure to magnetic fields is harmful stating "there is no reason for disquiet" at the end of the section. Because of that, I don't feel like it's a main aspect, but if you feel otherwise I'd be willing to discuss further.
 * I will add in a section about effects on production as that does seem to be an important point that has been neglected. Wugapodes (talk) 02:23, 14 November 2015 (UTC)
 * I have added a section touching on both aspects based on the new source and one that it cited. diff Wugapodes (talk) 03:22, 14 November 2015 (UTC)
 * Ok, that all sounds reasonable. I'll start a full review.  I'm a bit busy over the next couple of days, but I'll get it done as soon as I can. SpinningSpark 17:07, 14 November 2015 (UTC)


 * Principles of operation
 * The article should explain why it is necessary to use different frequencies.
 * ❌ I can't find any source that explains why different frequencies are used. I've requested Zierdt 1993 on Interlibrary loan but that could be a few weeks and I don't remember it saying anything about it either, especially since none of his other papers talk about it. Wugapodes (talk) 23:27, 16 November 2015 (UTC)
 * See this source, it is so that the signals from each of the transmitters can be separated after recovering them from the receiver. It is essential to have a scheme to resolve each transmitter individually if triangulation is required.  This is not the only possible scheme for this sort of thing, but is a simple one. SpinningSpark 00:58, 17 November 2015 (UTC)
 * ✅ Well that makes sense then. I'm not much of an engineer so that section didn't actually make much sense to me (I didn't understand what "separate the signal components" meant) but I will add it in. Wugapodes (talk) 04:14, 17 November 2015 (UTC)
 * The signal picked up by the receiver is a mush of all six transmitters (a composite signal). They are separated with electronic filters, just like when you tune your radio to one particular station. SpinningSpark 08:00, 17 November 2015 (UTC)


 * "proportional to the distance from the transmitters" (also in lead). I believe the correct relationship is going to be inverse square.  In any event, it is definitely not proportional, the induced current will go down with increasing distance from the transmitter, not up.
 * ✅ according to this source the relation is inverse cube. Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * "measure the movements" (also in lead). The description seems to suggest that it is position that is being measured, not velocity (although, of course, velocity can be calculated from the time change in position).  If only velocity was to be measured an oscillating field would not be necessary - it would be simpler to use a static magnetic field.
 * ✅ The sources I have say that despite the alternating field it is position is measured, see Steele 2004, pp. 343-344: "Three transmitters attached to this helmet create an alternating magnetic field within the helmet...allowing precise determination of the location of each coil." I'm no engineer, so I can't speak to why they are designed like this, but it seems to be how they are operated. Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * Development of two and three dimensional sensors
 * It is stated that 3D EMA eliminates "a heavy and restrictive headmount" but without explaining why. I am guessing that this is because it is not longer necessary to attach the the transmitters to the head, they can instead be mounted on their own structure.  I shouldn't have to guess, the guess could be wrong.
 * ✅ Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * The significance of Hixon's paper is not explained. Was he the first to publish on EMA? the first to use it? or something else?  Without this information, the reader is left wondering what it is doing in the article at all.


 * This might be the electrical engineer in me, so feel free to ignore this comment if you want, but I am wondering how accurate these measurements need to be. A 30 degree inclination of a sensor coil (also in lead) will result in a 13% reduction in induced emf.  For a transmitter 15 cm away, this is equivalent to the sensor appearing to be 1 cm further away than it actually is.  That sounds like an awful big error to me for this application.
 * 1cm can be wither really big or really small depending on what you're measuring. If used in dysphagia, it's probably not a huge concern as the overall movements and how the articulators move together are most important. In speech studies, it can be big (say, the difference between 's' and 'sh') or not that big (lip aperture for example). Regardless, 30 degrees is actually really difficult to achieve with your tongue, particularly on the midsagittal part. Give it a shot, it's not something that happens during normal articulation. The 30 degree tilt tends to be in sensor placement rather than in the actual movement of the tongue, but all that's OR and not included in the article. Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * There is an amount of repetion here. I read twice that 2D EMA is not accurate beyond 30 degrees, and twice that 3D transmitters are arranged in a sphere of six coils.
 * ✅ I think. I removed the redundant sentence and expanded on why Zierdt arranged them as he did. Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * Effect on subjects
 * The link to μT does not go to an appropriate article. I expect you meant Tesla (unit).
 * ✅Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * I seem to remember reading that the route the wires take from the sensors out of the mouth is important. Size and weight probably an issue also.  Or if modern equipment does this wirelessly, then that should be discussed also.
 * I added in a bit about wires coming out the side of the mouth. I'm unsure what you mean by size and weight, could you explain? No EMA device that I know of is wireless. Wugapodes (talk) 23:27, 16 November 2015 (UTC)


 * Images
 * I am not entirely convinced by the fair use rationale for the Zierdt image. I am prepared to be persuaded, but as a diagram it is easily replaced with a new svg (help available at Graphics Lab).  If this were an article just about Zierdt's invention, then maybe, but as just one step in the development of a broader subject it does seem out of place, especially as there is no image of the very first EMA.
 * Needs discussion I chose this image because, at least for me, the textual descriptions of the coil arrangement for 3D articulographs were thick and only the figure helped me. I assumed that many readers would feel similarly, and so I believed an image to be indispensable in conveying the information. The difference is that in a 3D articulograph the position of the transmitters and the angle of them is important, while in 2D the angle is less consequential. Further, triangulation is easily understood so an image of a 2D articulograph wouldn't be as valuable as one of a 3D articulograph in conveying new information.
 * The reason for Zierdt's non-free image is that not only does it serve to illustrate the historical context, but I believe that creating a work illustrating the same thing based on the non-free image would qualify as a derivative work (and no rights to create derivative works were granted). Since I believe it to fulfill the WP:NFCC, and believed a derivative work caused more problems than it solved, I decided to use the non-free image.
 * I realize that it's a long-winded explanation, but I'm happy to discuss further if you have concerns. Wugapodes (talk) 23:27, 16 November 2015 (UTC)
 * I don't deny that it is useful to have a diagram of the transmitter orientations. However, this can easily be done in a new diagram without using any of the styles or details of the original that would make it a derivative work.  What the transmitter orientations are is data, and data is not copyrightable.  I think the fairest thing to do here is to put the review on hold and ask for a second opinion from another reviewer.<b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 00:58, 17 November 2015 (UTC)
 * ✅ If you think that it can be done in accordance with copyright, then I can try and get that made, either by request or trying myself. In the mean time I'm willing to remove the picture since it can be replaced with a free equivalent. Wugapodes (talk) 04:14, 17 November 2015 (UTC)


 * Thinking about images some more, it strikes me that what is really essential for this article to be a GA is a photograph of an actual articulograph, preferably one in use on a subject. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 00:58, 17 November 2015 (UTC)
 * ❌ The GA criteria don't require images, and it doesn't seem possible to illustrate it with the kind of image you suggest. While I agree that an image of an articulograph in use would be best, there are no free images available in commons and none I can find online with a copyleft license. It's not possible to satisfy WP:NFCC given that free content could be created, but I don't have access to an articulograph so I can't take any. Wugapodes (talk) 04:14, 17 November 2015 (UTC)
 * I think I would still like to wait for a second opinion. This is the first time I have had an issue with a GA over lack of images and I would like to hear from another reviewer on how strictly we should interpret the GA requirement for images, and indeed how strictly we should follow the views in that essay. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 08:12, 17 November 2015 (UTC)
 * What_the_Good_article_criteria_are_not is quite clear on this point. If no suitable image is available, then no image is required. I agree that an image as described would be a significant improvement, but the article may not be failed because no free images or other media currently exist. &bull; &bull; &bull; Peter (Southwood) (talk): 07:14, 18 November 2015 (UTC)


 * References
 * " The significance and measurement of head position during speech production experiments using the x-ray microbeam system". The convenience link goes to Cornell University, which requires login. It would be much more useful to provide a link directly to the journal in which it was published where at least they can read the login abstract.  Those readers that have access through their universities or libraries can still follow that path. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 16:41, 16 November 2015 (UTC)
 * ✅ Wugapodes (talk) 23:27, 16 November 2015 (UTC)

Thanks for the review, ! I made most of the changes you suggested. Feel free to look over the article at your convenience to make sure I properly addressed your points, and if you could respond to the two I needed more discussion on (why different frequencies, size and weight, and non-free image) at your earliest convenience, I'd be very appreciative. The Thanks, Wugapodes (talk) 23:27, 16 November 2015 (UTC)