So we start with scenes. These are broken down into a sequence of one or more takes, and takes are broken down into a sequence of one or more frames. Moving from time to space, each frame is made up of a number of layers, perhaps of the order of 10, say seven, seven being an ever popular magic number in all kinds of contexts. A number which may be fixed and does have quite a low upper bound. Layers is a very finite resource.
Layers are considered to be arranged in a stack with a top and a bottom, with the top layer – or perhaps layers – reflecting the state of activation processes – proxies for the electrical field which we propose to be consciousness, while the remaining layers contain the data, are the contents of consciousness. There will often be links between the contents of adjacent data layers. Layers might be present but inactive, in which case they will have no part in the conscious experience. Just convenient, for some computing reason or other, for them to be there.
We also have objects, objects which persist in memory, which are not any part of the LWS, but which are represented there as structures in layers. For convenience, we assign objects to scenes. So this or that object may or may not participate in this or that scene.
All this is modelled in time in the top half of the illustration above and in entities in the bottom half – this last being the soft box modelling introduced at reference 4, and with the present diagram being a slightly different take on the world to that offered there. But there is little of substance between them.
We have already talked of the compiler which builds the successive versions, the successive frames of the LWS. We now introduce the handler which has a rather less complicated role, being the process which just does the updates which have been prepared by the compiler. Perhaps comparable to the BIOS – the basic input-output system – of a PC.
We allow some update to existing layers within frames. Update which may include changing the activity of a layer, provided that such change does not change their ranking. Update which may, at the limit, disturb the links between structures on adjacent layers.
If we make a significant change to the activity of a layer, delete a layer, add a layer or replace a layer, we need a new frame, with a short gap in time between the old fame and the new frame, rather as there is in the cinema, as the compiler needs a bit of time to adjust the data and the activation processes that go with it. It is possible that the compilation, in going for the small compile associated with a new frame, will get into a muddle, will need to back out and go for the big compile associated with a new take. The compiler, as it were, is only human; it is neither all-powerful, always right nor the deity.
If we change all the layers, we need a new take, with a slightly longer gap between the last frame of the old take and the first frame of the new.
If we change the scene completely, perhaps with a change of activity, position or place, with a whole new set of objects, we need a new scene, with an even longer gap. In some circumstances, for example when waking up, the subject may be conscious of layers coming online over a period of some seconds, perhaps even as long as minutes. So, for example, when I have had toothache, I have sometimes been conscious of the pain taking some seconds to kick back in after I wake up.
As candidate changes appear in UCS (the great mass of unconscious processing, the invisible, underwater part of the iceberg), the compiler will need to make a judgement about whether those changes require a new frame, a new take or a new scene – and deploy processes and resources accordingly. It is this deployment which is important, so we do not need to be pedantic about the definitions, the indications given above.
Nevertheless, we might formalise the business of a new frame along the following lines. The frame has an age, a duration in time since it was compiled (a positive real, T). Each layer has an activity (a non-negative real A) and a content (a non-negative real C), this last being some measure of the information held in that layer, a measure which might be very small (Tononi & Koch’s IIT & PHI notwithstanding) or very large, subject only to the space constraint on a layer, supposed to be of the order of some tens of megabytes, about that of a photograph that I might take on my telephone. Not only are the C bounded, the sum of the products, A*C, is bounded too, as we can only attend to so much stuff at any one time; so even if we have a spare layer, there may not be the capacity to bring on a new one unless we turn something off. There is only so much processing capacity, although limits of this sort are apt to vary from person to person. That apart, the layer with the largest value of A is the focus, the subject of attention. Then we have an algorithm which says that if there is a lot of change in a layer, or if we want to add a new layer, usually by overwriting an existing layer, or if we want to change the focus, or if the frame has been going on for a long time, then we need a new frame.
Scene being a word that is used in drama, in films, and in a looser way, in everyday life. I offer a short narrative below, broken into the sort of quite short scenes I have in mind here.
We suppose that I am in the sort of employment in the City which allows me to get my hair cut during working hours.
Scene: working away at my hot desk. My own personal pot plant looking a bit the worse for wear for all its travels.
Take: closing down my hot desk.
Scene: going down to the street in the lift.
Scene: Gutter Lane. A quiet, narrow street, with various doors leading into place of office employment of various shapes and sizes.
Scene: Cheapside. A wider, much busier street, with lots of retail action, directed at the people who work in the area.
Scene: in the barber’s shop.
Scene: Cheapside.
And so on, back to my hot desk.
Change of take
Something of an abuse of term here as, according to https://en.wikipedia.org/wiki/Take, we have it that ‘in cinematography, a take refers to each filmed version of a particular shot or setup. Takes of each shot are generally numbered starting with ‘take one’ and the number of each successive take is increased (with the director calling for ‘take two’ or ‘take eighteen’) until the filming of the shot is completed. Film takes are often designated with the aid of a clapperboard. It is also referred to as the slate. The number of each take is written or attached to the clapboard, which is filmed briefly prior to or at the beginning of the actual take. Only those takes which are vetted by the continuity person and/or script supervisor are printed and are sent to the film editor’.
While according to http://www.filmsite.org/filmterms17.html, we have it that ‘[a shot is] the basic building block or unit of film narrative; refers to a single, constant take made by a motion picture camera uninterrupted by editing, interruptions or cuts, in which a length of film is exposed by turning the camera on, recording, and then turning the camera off; it can also refer to a single film frame (such as a still image); a follow-shot is when the camera moves to follow the action; a pull-back shot refers to a tracking shot or zoom that moves back from the subject to reveal the context of the scene; see also scene and sequence; shot analysis refers to the examination of individual shots; a one-shot, a two-shot, and a three-shot refers to common names for shooting just one, two, or three people in a shot’.
So what I have been calling a take is more properly a shot. But I stick with take and I continue with the barber shop scene, broken into the sort of takes I have in mind. A scene which might, if the barber and his doings were important, be broken into two or more scenes, but I stick with just the one here.
Scene: in the barber’s shop. Maybe half a dozen chairs and about the same number of barbers. A long narrow shop with chairs along one wall an benches on which to wait along the other. A scattering of magazines and newspapers. This being a costume drama on BBC, ashtrays, fags and clouds of smoke. Maybe personal items tucked away in drawers, BBC sometimes getting quite precious about their dressing of scenes. In front of the chairs, mirrors and shelves, these last cluttered with the paraphernalia of the trade. Including strops and razors.
Take: going into the shop.
Take: sitting on the bench, waiting my turn. Idling turning the pages of a certain tabloid newspaper which I would not usually buy.
Take: my turn in the chair, watching the goings on in the big mirror. Scissor action.
Take: clipper action.
Take: sprayer action.
Take: the end of the cut, peering at the result in the small mirror being held up behind me.
Take: standing up, thanking the barber and starting to head for the desk.
Take: paying at the desk.
Take: leaving the shop, back into Cheapside.
Change of frame
Slightly adapted from http://www.filmsite.org/filmterms17.html, we have it that a take ‘refers to a single image, the smallest compositional unit of a film's structure, captured by the camera on a strip of motion picture film - similar to an individual slide in still photography; a series of frames juxtaposed and shown in rapid succession make up a motion (or moving) picture; also refers to the rectangular area within which the film image is composed by the film-maker ... While frame rate refers to the rate at which film stock passes in front of the camera's aperture while filming; present-day films are usually run through a camera or projector at a frame rate (running speed or camera speed) of 24 fps (frames per second); older films, made at 18 fps, appear jerky and sped-up when played back at 24 fps - this technique is referred to as undercranking; over-cranking refers to [speeding up] the frame rate (i.e., shooting at 48 or 96 fps), thereby producing slow-motion action when viewed at 24 fps’.
So again, a slight abuse of language in that our frames are much longer in duration and are not static in the way that the frames of a film are static. We allow a certain amount of change, of action during a frame, quite possibly in both the sight and sound departments. Although, that said, too much change and the integrity of LWS is lost and the subject will probably become confused or disorientated.
Scene: still in the barber’s shop.
Take: sitting on the bench, waiting my turn.
Frame: just sitting, idling glancing about. A small shift in the object of attention, a small saccade, can be accommodated within frame. A big shift, a big saccade, needs the recompilation of the visual field that comes with a new frame. In this, the shift in the object of attention is more important than the more easily verified facts of a saccade; perhaps more accurately just the shift in attention, the object, at a gross level, say the mirror in front of me, might be the same, but with the focus of attention moving around it, around its parts. Perhaps, alighting on a prismatic effect arising from an edge bevel.
Frame: pick up the newspaper and start to look at it. The mirror and the barber’s shop recede into the unattended background. The motor and sensory action involved in this picking up need at least one new layer. Maybe one of the old layers has to go.
Frame: gazing at page 14 of the newspaper, fairly absorbed in it. The layer holding page 14, and perhaps another layer holding something of interest within that, perhaps a picture of a man biting a dog, are what are important now.
Frame: turn a few more pages. Attention now more spread about, both on the action of turning the pages, perhaps looking about a bit, probably glancing down at each new page. Put another way and using the jargon introduced earlier, there is a flattening of the values of A, from having been very skewed, across the layers of the LWS.
Frame: gazing at page 23 of the newspaper, fairly absorbed in it.
Frame: put the newspaper down.
Frame: just sitting, idling glancing about.
Not change of frame
The sort of things that can be encompassed within a single frame.
Frame: back with looking at the edge of the mirror in front of me more closely. Quite small shifts of attention, just millimetres on the mirror. According to wikipedia, ‘… If one looks at a one-centimetre object at a distance of one metre and a two-centimetre object at a distance of two meters, both subtend the same visual angle of about 0.01 rad or 0.57°. Thus they have the same retinal image size of around 0.17mm…’. I am perhaps talking of a one centimetre object at two metres. See reference 5.
Frame: still with the mirror, but my attention has been caught by the image (in the mirror) of a van going slowly along the road outside, in the heavy traffic of Cheapside. I am trying to make out what it says on the side of the van. The frame can allow the movement of the van. I have to do the best I can, looking at the one place in the mirror. Tracking the van as it moves not an option in this context.
Walking along East Street in Epsom the other day, towards Ewell, I noticed a baby’s plastic rattle lying on the pavement in front of me. An attractive, brightly coloured plastic rattle, the shape of a small dumbbell, with bits inside that rattled. I picked it up thinking it might do for a young grandchild. Turned it over a bit. About to start moving forward again. But at that point I became conscious of someone shouting behind me. I turned around, and gradually a young woman with a baby in a buggy came into focus. I connected the shouting with the young woman. A few seconds later she had got back to me, a little hot and bothered, and I returned her baby’s rattle.
Part of the interest of this anecdote being that its core developed over several seconds, from the time when I heard the shouting, until the young women came into focus. Which might expand as follows.
Scene: East Street in Epsom.
Take: ‘walking along East Street in Epsom the other day, towards Ewell’. Not thinking of anything in particular, but attention flickering around me, eyes darting from place to place.
Take: new take for all the rattle, mother and baby stuff. Triggered by my attention shifting to the rattle – which is now being attended to, rather than just being something vaguely in the scene in front of me.
Frame: ‘I noticed a baby’s plastic rattle lying on the pavement in front of me. An attractive, brightly coloured plastic rattle, the shape of a small dumbbell, with bits inside that rattled’. A new object and new layer for the rattle – new stuff which, by definition of the LWS, has to be created and put in place before we become conscious of it. Walking layer updated for standing, probably deactivated, possibly recycled
Frame: ‘I picked it up thinking it might do for a young grandchild’. New object for the new granddaughter.
Frame: ‘turned it over a bit’. New layer or layers for the sensations involve in turning the rattle, so arranged that the new layer objects involved more or less coincided in position with that of the rattle itself.
Frame: ‘about to start moving forward again’. New layer for the change of focus. Rattle layers downgraded, probably not yet recycled.
Frame: ‘but at that point I became conscious of someone shouting behind me’. New layer for the shouting.
Frame: ‘I turned around’. New frame for the new action.
Take: new take for the entirely new scene in front of me.
Frame: replace visual scene layer(s) carrying what was in front with the new ones carrying what was behind.
Frame: ‘and gradually a young woman with a baby in a buggy came into focus. I connected the shouting with the young woman’. New object and new layer for the young woman.
Frame: ‘a few seconds later she had got back to me, a little hot and bothered’. A mainly visual frame.
Take: new take for the interchange with the mother.
Frame: make contact. Eye contact between me and the mother.
Frame: ‘and I returned her baby’s rattle’. A frame both visual and active.
Frame: close contact.
Scene: new scene for resumption of walk. With the rattle incident not having a scene to itself, rather just being tacked onto the end of the East Street scene already in place. The close of contact makes a good end of scene, but there was no clear start of scene at the beginning, just a rising to a climax with the return of the rattle. It does not make much sense in this context to impose a beginning after the event. In the barber shop example, the scene structure was more predictable.
Take: new take for resumption of walk.
Frame: turn around again.
Frame: I resumed walking along East Street.
Conclusions
We have illustrated by two examples the way that we see scenes, takes and frames developing in the LWS. Showing, inter alia, that while the breakdown in scenes, takes and frames fits the way that we experience the world well enough, the timing and duration of successive frames is a balance, a compromise between the flow of sensory data into the brain and the needs arising from the compiler trying to make some sense of it all.
A compiler which is not perfect, which will often get things wrong and which can easily be tricked by things which it is not used to, which it was not designed for. Such as the trickery of Brigit Riley, on view in miniature at reference 6.
And, generally speaking, while we experience the scenes, takes and frames as they develop in the LWS, we are not ticking them off in the way in the way that a film director might. Not that there is any particular bar to such self-consciousness, such self-regard in the LWS, just no need for it. Most of the time it would not add much.
References
Reference 1: https://en.wikipedia.org/wiki/Global_Workspace_Theory.
Reference 2: Global workspace dynamics: cortical binding and propagation enables conscious contents - Bernard J. Baars, Stan Franklin and Thomas Zoega Ramsoy – 2013. Open access.
Reference 3: http://psmv3.blogspot.co.uk/2017/05/in-praise-of-homunculus.html.
Reference 4: http://psmv3.blogspot.co.uk/2017/05/recap-on-our-data-structure.html.
Reference 5: https://en.wikipedia.org/wiki/Saccade.
Reference 6: http://www.op-art.co.uk/bridget-riley/.
Groups search key: src.
No comments:
Post a Comment