[flow_default] Transcription for week05 16 pose reader doing the other side.wav
Browse files
week05 16 pose reader doing the other side.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"file": "week05 16 pose reader doing the other side.wav",
|
| 3 |
+
"transcription": {
|
| 4 |
+
"audio_file": "week05%2016%20pose%20reader%20doing%20the%20other%20side.wav",
|
| 5 |
+
"text": "I'll quickly do it on the other side and I thought I'll record this too. If you guys are clear on what to do, you guys don't have to watch this video. It's just going to be a repetition of what we already just did, but this time on the other side. Just to kind of like have it complete on both sides. The first thing that I'm going to do here though is I'm going to rename those post readers because right now in a moment I just call it post reader. So if we kind of do the same on the other side, then we're going to get a naming conflict. So at a very bare minimum, I'm going to rename those to be R. Actually, you know what? I'm going to select all of them and just go to, it should be a rename to research and replace names. So I'm going to rename or search for pose and rename it with r underscore pose for selected ones. Here we go. And we can name the other ones left side. Probably we should find an even better names. We should probably call this, you know, right arm post reader or something like that or something that is unique in case we might use the post reader also on other areas of our body here. So let's go in one more time and say search and replace names and call a rename post to arm post maybe like that for select ones. So that makes it a little bit more unique in terms of the naming. We could probably also go in and change the multiplier that we have in there. So that was from the main one, reader main. We show the outgoing connections. Then we have that one multiplier in here. So perhaps we should go in and we should perhaps rename this here as well. To be right arm, pose, reader, double, MDI or something like that. Okay, so let's create the 140 at our side now. I think I still have my isolates elect going on, so that will obviously help. Then I'll create a new locator. I'll call this left arm post reader main locator. We'll group this. We'll call this left arm post reader group. We'll take this group and we'll parent it under this joint parent and zero everything out. So that joint being the left arm IK shoulder auto root joint here. Let's bring it into the same space. Here we go. And then we'll duplicate that twice. One time for the target. Target locator. I'll put this to minus five. I'll put this to, this is the main one. to Actually, it doesn't really matter. So let's move it forward to have it consistent. So here we move this five units. So let's put five minus five on here and scale it down a little to 0.5 since this is going to be our up vector, our up locator. We can also do color coding here on this side as well. We don't have to, but it makes it easier to see. And typically, you know, I don't necessarily create it all by hand. I scripted it all up, so it will be just a matter of seconds. But here, as I said, I want to try to stay committed to not using any scripts or too many scripts. So really show you guys that we can do all that stuff also by hand. And I think it's there is great value in doing it by hand because then you really learn how to do it and how it should work under the hood. You can practice it more. Okay, so this is the main one. So let's set this one up already. So we want to go this our target. So main one, create a constraint. I want to see what we want to aim to what. Making sure that we are on object or local space, so we are. So the aim should be minus y here, minus 1 and y, 0 here, and the up vector should be minus z, minus 1 and z, and 0 here, and the up is going to be L arm post reader out of twist up L arm post reader up locator L arm post reader up locator, okay, that's correct. We'll apply and double check make sure we have zero rotations here. The aim constraint is created with zero rotations and we should also have zero offset so that looks good. Then we'll go in and we'll create the ones for the twisting here as well. Duplicate setDos to zero. That's going to be the main one for the twist, so therefore it's going to be two post-reader twists or twist main. We'll duplicate that. I think we put it forward, minus 5. That's going to be the target for the twist target. Locator, since it's a target, we can set the scale to 1. And then last but not least, we need our up twist up. I think we put that to 4. Actually 3. So let's put this to 3 as well and scale it down to 0.5 since it's the up for the twist. Let's change the colors here also. Blue. That is a twist locator. And then this one is also a twist locator. Then let's set those up here as well. So we have our twist target and we have our twist main. We go in and create the aim constraint one more time. So here we want to have the aim go into minus z, minus one and z, zero here, and the up is going to be in y, so one in y and zero here. And then the up object is going to be called larmPoseReaderTwistUp. Apply and double-checking the twist locator has zero rotations and on the aim constraint there are also zero offsets. That's looking good. Now we have to do some regrouping here. So first we want to take the two targets, main target and twist target, group them into a post reader actually L arm post reader target group and the target group should go under that joint although we have still all these joints under that joint anyway so we now need to take this whole post reader group and parent it one above so we set one above the L arm, IK shoulder, root joint comes to chest, right? If we're not sure, we just follow that line where it ends up and it ends up in the chest. So that's where this left arm post reader group should go. I'll parent this here and then we'll take the post reader target group and parent it under the joint that we want to measure. So we'll take those, turn them under the left arm, shoulder, auto-root. Now if we rotate this, well we can't really, but if we take that and translate it around, we should see the aim doing its thing, and we can. But we have to remember that this is only measuring half of the rotation, so we have to go through another multiply node here. From the left arm post reader main locator, we need to multiply. I think before we call this a left arm post reader double, MDI. As I said, we connect only the bend and the size that would only be X and Z in this case. And we multiply those by two. And then from here, we are going into the group that we still have to create for the two twists. Before we do anything, let's set this back to zero. And let's take the two twist locators, group them. That would be left arm, pose, reader, twist group. And that twist group, we want to leave it where it is, but we want to apply that rotation that's coming out of the multiply node. So we'll add the group in here as well. Twist group and we'll connect the X and the Z here. I'll put X goes to rotate X. I'll put Z goes to rotateX, outputSat goes to rotateSat. Now when we're moving this up and down, we should see that these twist... twist locators are all pretty much lined up. And now we can, as a last step, we can create our custom attributes here on this joint as well. So add our underscores, 10 of them. And then we'll also add a pose, bend, pose, twist, pose, side. And then we'll hook those up. Add it in here. And I can go from my duplication of the rotate values of the main post reader locator to the output of the after it's doubled up. We'll take the output of the multiply divide node and x and z here and connect it x goes to band, z goes to side and then for twist we want to get that from our twist locator. Graph at that in here and connect it to the joint to the custom attribute. So we go from rotate y into the post twist. That should do the trick in terms of reading the correct poses here, what the arm is doing at that particular moment. And then we'll take these and show the outgoing connections from there. Again, we can ignore the pole vector and the IK handle. We can ignore the child joint and we can ignore the model panel here. So really what we want to replace is instead of going from the rotations into these input values of the shoulder auto on off, shoulderer auto on off I guess. We want to go in from the post bend, post twist, post side to make this more stable. Post bend to input X, post twist to input Y, post side to input Z. Now if we look here, we should have both made a lot more stable. We can come in here, rotate them, or translate them around, move the pole vectors, and we can see it's kind of like doing the right thing now. Wherever we are, obviously the more extreme we go here with this even in a pose like that which is quite extreme the shoulders, the outer shoulder seem to be doing the right thing and therefore if the outer shoulders are doing the right thing then the clavicle here is also doing the right thing so no more popping or flipping obviously here we can see something that's a little bit funky looks a little bit funky but that's just because we have the pole vectors in such a tight spot here. See where they will put them right inside there. So obviously they should probably be moved a little bit out. So now with these post readers in place for the shoulders or other shoulders, This will behave a lot more stable. And those post readers could really be used for a lot of different places. I mean, I think the shoulders is probably the number one spot where we want to have them just because the arm is a ball joint, as I mentioned before. They can pretty much rotate in all three X's, XY and Z. So therefore the shoulders tend to like pop and flip a little bit, but with this makes it a lot more stable, a lot better. But we could even think about using that pose reader for other areas as well. And I've used it in a lot of different places before like hips or knees or ankles, anything that rotates a lot basically.",
|
| 6 |
+
"language": "en",
|
| 7 |
+
"duration": 915.31,
|
| 8 |
+
"timestamp": "2025-11-28T05:20:25.309470"
|
| 9 |
+
},
|
| 10 |
+
"timestamp": "2025-11-28T05:20:25.318473",
|
| 11 |
+
"processing_time_seconds": 168.0435757637024
|
| 12 |
+
}
|