Motion (.MOT) File Format - Revision history

Akito: /* (Example): */

2025-09-07T22:42:02Z

(Example):

← Older revision		Revision as of 22:42, 7 September 2025
Line 217:		Line 217:
	\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''		\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
	\|-		\|-
	\| 0x01 \|\| 50 \|\| This field is telling that '''the curve animates the X Position of the ~~root~~ bone''', being set on a '''Quantized-Precision Curve ''Property Family''.'''		\| 0x01 \|\| 50 \|\| This field is telling that '''the curve animates the X Position of the bone index 0''', being set on a '''Quantized-Precision Curve ''Property Family''.'''
	\|-		\|-
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
Line 223:		Line 223:
	\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframes header are located'''.		\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframes header are located'''.
	\|}		\|}

	===(Keyframes header):===		===(Keyframes header):===
	{\| class="wikitable"		{\| class="wikitable"

Akito: /* (Example): */

2025-09-07T22:41:24Z

(Example):

← Older revision		Revision as of 22:41, 7 September 2025
Line 150:		Line 150:
	\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''		\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
	\|-		\|-
	\| 0x01 \|\| 16 \|\| This field is telling that '''the curve animates the X Position of the ~~root~~ bone''', being set on a '''Quantized Curve ''Property Family''.'''		\| 0x01 \|\| 16 \|\| This field is telling that '''the curve animates the X Position of the bone index 0''', being set on a '''Quantized Curve ''Property Family''.'''
	\|-		\|-
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''

Akito at 20:57, 7 September 2025

2025-09-07T20:57:42Z

← Older revision		Revision as of 20:57, 7 September 2025
Line 290:		Line 290:
	After that, the values can be used straight into the cubic interpolation formula.		After that, the values can be used straight into the cubic interpolation formula.

	===Final ~~nodes~~===		===Final Notes:===
	The space in which the movement of bones takes in also depends of the ''Property Family'' the curve is set in. For instance, if a bone's rest position's property (let's say it's 125.14 for X) is animated on a quantized curve, Since this is a big value, movement information will be lost due to the nature of the quantized ''Property Family'', especially when not used in a ''Quantized-Precision Curve''. Because of that, '''The bone's movements are relative to the bone's parent's pose on quantized types.''' For full precision however, since quantization is not needed, '''the property values are absolute.'''		The space in which the movement of bones takes in also depends of the ''Property Family'' the curve is set in. For instance, if a bone's rest position's property (let's say it's 125.14 for X) is animated on a quantized curve, Since this is a big value, movement information will be lost due to the nature of the quantized ''Property Family'', especially when not used in a ''Quantized-Precision Curve''. Because of that, '''The bone's movements are relative to the bone's parent's pose on quantized types.''' For full precision however, since quantization is not needed, '''the property values are absolute.'''

Akito at 20:57, 7 September 2025

2025-09-07T20:57:23Z

← Older revision		Revision as of 20:57, 7 September 2025
Line 154:		Line 154:
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
	\|-		\|-
	\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the ~~keyframe infos~~ are located'''.		\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframes header are located'''.
	\|}		\|}

Line 177:		Line 177:
	This is the only ''Property Family'' that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.		This is the only ''Property Family'' that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.

	After ~~that~~, the keyframes are ~~given~~.		After reading the keyframes header, what follows up are the keyframes in which the actual keyframes are stored in.

	=== Quantized Keyframes (4 bytes length)===		=== Quantized Keyframes (4 bytes length)===
Line 221:		Line 221:
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
	\|-		\|-
	\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the ~~keyframe infos~~ are located'''.		\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframes header are located'''.
	\|}		\|}
	===(Keyframes header):===		===(Keyframes header):===
Line 256:		Line 256:


	=== Full-Precision Curve ===		== Full-Precision Curve ==
	This is the most straightforward ''Property Family''. However, the most expensive (in storaging terms).		This is the most straightforward ''Property Family''. However, the most expensive (in storaging terms).
	Different from the Quantized types, this ~~however~~ is not dependent of a ''Keyframe Header.''		Different from the Quantized types, this one is not dependent of a ''Keyframe Header.''
	Instead, it tells the values right away, Rendering this ''Property Family'' the easiest to compute (considering this is not a single pose), but the most expensive to store in storage terms.		Instead, it tells the values right away, Rendering this ''Property Family'' the easiest to compute (considering this is not a single pose), the one with the most accurate movements, but the most expensive to store in storage terms.

			===(Example):===
			{\| class="wikitable"
			! Offset !! Value !! Description
			\|-
			\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
			\|-
			\| 0x01 \|\| 80 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Full-Precision Curve ''Property Family''.'''
			\|-
			\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
			\|-
			\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframes are located'''.
			\|}
			===(Keyframes):===
			{\| class="wikitable"
			! Offset !! Type !! Size !! Description
			\|-
			\| 0x00 \|\| <code>uint16_t</code> \|\| 2 \|\| Keyframe time '''(Absolute)'''.
			\|-
			\| 0x02 \|\| <code>uint16_t</code> \|\| 2 \|\| Padding, used to align the file, It has no use.
			\|-
			\| 0x04 \|\| <code>float32_t</code> \|\| 4 \|\| Keyframe Value
			\|-
			\| 0x06 \|\| <code>float32_t</code> \|\| 4 \|\| Value of the keyframe's end tangent. (m0)
			\|-
			\| 0x06 \|\| <code>float32_t</code> \|\| 4 \|\| Value of the keyframe's end tangent. (m1)
			\|}

			After that, the values can be used straight into the cubic interpolation formula.

			===Final nodes===
			The space in which the movement of bones takes in also depends of the ''Property Family'' the curve is set in. For instance, if a bone's rest position's property (let's say it's 125.14 for X) is animated on a quantized curve, Since this is a big value, movement information will be lost due to the nature of the quantized ''Property Family'', especially when not used in a ''Quantized-Precision Curve''. Because of that, '''The bone's movements are relative to the bone's parent's pose on quantized types.''' For full precision however, since quantization is not needed, '''the property values are absolute.'''

Akito at 20:46, 7 September 2025

2025-09-07T20:46:25Z

← Older revision		Revision as of 20:46, 7 September 2025
Line 41:		Line 41:
	\|}		\|}

	the 4th offset of each curve info can be either a float, or uint32. it's only set as a float if the ~~property family~~ of the curve is a Single Pose. Otherwise, it's a offset.		the 4th offset of each curve info can be either a float, or uint32. it's only set as a float if the ''Property Family'' of the curve is a Single Pose. Otherwise, it's a offset.
	The curve property will define what property of the bone/object it is animating. There's some properties who could not be determined yet. See some already identified values bellow.		The curve property will define what property of the bone/object it is animating. There's some properties who could not be determined yet. See some already identified values bellow.

Line 124:		Line 124:
	\|}		\|}

	The Property Family tells the game how the keyframes are handled ingame. Depending on the Family the property is set in, ~~it can~~ be readen diferently.		The Property Family tells the game how the keyframes are handled ingame. Depending on the Family the property is set in, curves will be readen diferently.
	== Single Pose ==		== Single Pose ==
	This Property Family is the cheapest to compute. Because since it's a single pose, The there's only one keyframe, in which is given on the Curve Info.		This ''Property Family'' is the cheapest to compute. Because since it's a single pose, The there's only one keyframe, in which is given on the Curve Info.

	===(Example):===		===(Example):===
Line 134:		Line 134:
	\| 0x00 \|\| -1 \|\| This field is telling that '''the curve is animating the root bone of the model.'''		\| 0x00 \|\| -1 \|\| This field is telling that '''the curve is animating the root bone of the model.'''
	\|-		\|-
	\| 0x01 \|\| 5 \|\| This field is telling that '''the curve animates the Z rotation of the root bone''', being set on a '''Single-Pose Property Family.'''		\| 0x01 \|\| 5 \|\| This field is telling that '''the curve animates the Z rotation of the root bone''', being set on a '''Single-Pose ''Property Family''.'''
	\|-		\|-
	\| 0x02 \|\| 1 \|\| This field is telling that '''there's only one keyframe on this curve, which makes sense, since this curve is set on a Single-Pose Property Family.'''		\| 0x02 \|\| 1 \|\| This field is telling that '''there's only one keyframe on this curve, which makes sense, since this curve is set on a Single-Pose ''Property Family''.'''
	\|-		\|-
	\| 0x04 \|\| 0.52 \|\| This field is '''telling the value or the Z rotation right away''', no more computing needed.		\| 0x04 \|\| 0.52 \|\| This field is '''telling the value or the Z rotation right away''', no more computing needed.
Line 142:		Line 142:

	== Quantized Curve ==		== Quantized Curve ==
	This Property Family Is known for having Quantized values. It's the most eficient type of curve, regarding storage size.		This ''Property Family'' Is known for having Quantized values. It's the most eficient type of curve, regarding storage size.

	===(Example):===		===(Example):===
Line 150:		Line 150:
	\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''		\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
	\|-		\|-
	\| 0x01 \|\| 16 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Quantized Curve Property Family.'''		\| 0x01 \|\| 16 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Quantized Curve ''Property Family''.'''
	\|-		\|-
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
Line 175:		Line 175:
	\| 0x0A \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the end tangent (m1)		\| 0x0A \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the end tangent (m1)
	\|}		\|}
	This is the only Property Family that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.		This is the only ''Property Family'' that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.

	After that, the keyframes are given.		After that, the keyframes are given.
Line 208:		Line 208:

	== Quantized-Precision Curve ==		== Quantized-Precision Curve ==
	For this Property Family, the process is almost the exact same. However, there are some changes.		For this ''Property Family'', the process is almost the exact same. However, there are some changes.
	First of, This is a more precise version of the Quantized curves. Therefore, the values are doubled up.		First of, This is a more precise version of the Quantized curves. Therefore, the values are doubled up.

Line 217:		Line 217:
	\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''		\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
	\|-		\|-
	\| 0x01 \|\| 50 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Quantized-Precision Curve Property Family.'''		\| 0x01 \|\| 50 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Quantized-Precision Curve ''Property Family''.'''
	\|-		\|-
	\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''		\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
Line 253:		Line 253:
	\|}		\|}
	Aside from the keyframe time being '''Aboslute''', the rest of the process remains the exact same.		Aside from the keyframe time being '''Aboslute''', the rest of the process remains the exact same.



			=== Full-Precision Curve ===
			This is the most straightforward ''Property Family''. However, the most expensive (in storaging terms).
			Different from the Quantized types, this however is not dependent of a ''Keyframe Header.''
			Instead, it tells the values right away, Rendering this ''Property Family'' the easiest to compute (considering this is not a single pose), but the most expensive to store in storage terms.

Akito at 20:36, 7 September 2025

2025-09-07T20:36:01Z

← Older revision		Revision as of 20:36, 7 September 2025
Line 183:		Line 183:
	! Offset !! Type !! Size !! Description		! Offset !! Type !! Size !! Description
	\|-		\|-
	\| 0x00 \|\| <code>~~int8_t~~</code> \|\| 1 \|\| Keyframe time (Relative to the time of the last keyframe).		\| 0x00 \|\| <code>uint8_t</code> \|\| 1 \|\| Keyframe time '''(Relative to the time of the last keyframe).'''
	\|-		\|-
	\| 0x01 \|\| <code>~~int8_t~~</code> \|\| 1 \|\| Delta value of the Keyframe.		\| 0x01 \|\| <code>uint8_t</code> \|\| 1 \|\| Delta value of the Keyframe.
	\|-		\|-
	\| 0x02 \|\| <code>~~int8_t~~</code> \|\| 1 \|\| Delta value of the keyframe's start tangent. (m0)		\| 0x02 \|\| <code>uint8_t</code> \|\| 1 \|\| Delta value of the keyframe's start tangent. (m0)
	\|-		\|-
	\| 0x03 \|\| <code>~~int8_t~~</code> \|\| 1 \|\| Delta value of the keyframe's end tangent. (m1)		\| 0x03 \|\| <code>uint8_t</code> \|\| 1 \|\| Delta value of the keyframe's end tangent. (m1)
	\|}		\|}

Line 239:		Line 239:
	\| 0x0A \|\| <code>float32_t</code> \|\| 4 \|\| Maximum value of the end tangent (m1)		\| 0x0A \|\| <code>float32_t</code> \|\| 4 \|\| Maximum value of the end tangent (m1)
	\|}		\|}
	You can notice that, since the values were doubled up, The use of the custom float16_t were completely removed.		You can notice that, since the values were doubled up, The use of the custom float16_t were completely removed, since now the file can use float32's instead, which are more precise and stable in comparision. This results on the size of the keyframe header to be extended to '''24 bytes each.'''
	===(Keyframe):===		===(Keyframe):===
			{\| class="wikitable"
			! Offset !! Type !! Size !! Description
			\|-
			\| 0x00 \|\| <code>uint16_t</code> \|\| 2 \|\| Keyframe time '''(Absolute)'''.
			\|-
			\| 0x02 \|\| <code>uint16_t</code> \|\| 2 \|\| Delta value of the Keyframe.
			\|-
			\| 0x04 \|\| <code>uint16_t</code> \|\| 2 \|\| Delta value of the keyframe's start tangent. (m0)
			\|-
			\| 0x06 \|\| <code>uint16_t</code> \|\| 2 \|\| Delta value of the keyframe's end tangent. (m1)
			\|}
			Aside from the keyframe time being '''Aboslute''', the rest of the process remains the exact same.

Akito at 20:31, 7 September 2025

2025-09-07T20:31:40Z

← Older revision		Revision as of 20:31, 7 September 2025
Line 86:		Line 86:

	\|-		\|-
	\| 50 \|\| ~~Semi~~-Precision Curve \|\| Position X		\| 50 \|\| Quantized-Precision Curve \|\| Position X
	\|-		\|-
	\| 51 \|\| ~~Semi~~-Precision Curve \|\| Position Y		\| 51 \|\| Quantized-Precision Curve \|\| Position Y
	\|-		\|-
	\| 52 \|\| ~~Semi~~-Precision Curve \|\| Position Z		\| 52 \|\| Quantized-Precision Curve \|\| Position Z
	\|-		\|-
	\| 53 \|\| ~~Semi~~-Precision Curve \|\| Rotation X		\| 53 \|\| Quantized-Precision Curve \|\| Rotation X
	\|-		\|-
	\| 54 \|\| ~~Semi~~-Precision Curve \|\| Rotation Y		\| 54 \|\| Quantized-Precision Curve \|\| Rotation Y
	\|-		\|-
	\| 55 \|\| ~~Semi~~-Precision Curve \|\| Rotation Z		\| 55 \|\| Quantized-Precision Curve \|\| Rotation Z
	\|-		\|-
	\| 56 \|\| ~~Semi~~-Precision Curve \|\| Scale X		\| 56 \|\| Quantized-Precision Curve \|\| Scale X
	\|-		\|-
	\| 57 \|\| ~~Semi~~-Precision Curve \|\| Scale Y		\| 57 \|\| Quantized-Precision Curve \|\| Scale Y
	\|-		\|-
	\| 58 \|\| ~~Semi~~-Precision Curve \|\| Scale Z		\| 58 \|\| Quantized-Precision Curve \|\| Scale Z

	\|-		\|-
Line 128:		Line 128:
	This Property Family is the cheapest to compute. Because since it's a single pose, The there's only one keyframe, in which is given on the Curve Info.		This Property Family is the cheapest to compute. Because since it's a single pose, The there's only one keyframe, in which is given on the Curve Info.

	(Example):		===(Example):===
	{\| class="wikitable"		{\| class="wikitable"
	! Offset !! Value !! Description		! Offset !! Value !! Description
Line 144:		Line 144:
	This Property Family Is known for having Quantized values. It's the most eficient type of curve, regarding storage size.		This Property Family Is known for having Quantized values. It's the most eficient type of curve, regarding storage size.

	(Example):		===(Example):===
	{\| class="wikitable"		{\| class="wikitable"
	! Offset !! Value !! Description		! Offset !! Value !! Description
Line 157:		Line 157:
	\|}		\|}

	== Quantized Curve Keyframe Header (12 bytes length)==		=== Quantized Curve Keyframe Header (12 bytes length)===
	'''Only Quantized property families contain a keyframe Header'''. That's because Every keyframe is dependent on the values given on the Header.		'''Only Quantized property families contain a keyframe Header'''. That's because Every keyframe is dependent on the values given on the Header.

Line 179:		Line 179:
	After that, the keyframes are given.		After that, the keyframes are given.

	== Quantized Keyframes (4 bytes length)==		=== Quantized Keyframes (4 bytes length)===
	{\| class="wikitable"		{\| class="wikitable"
	! Offset !! Type !! Size !! Description		! Offset !! Type !! Size !! Description
Line 202:		Line 202:
	* M0_value = '''minimum_m0 + maximum_m0 * delta_m0'''		* M0_value = '''minimum_m0 + maximum_m0 * delta_m0'''
	* M1_value = '''minimum_m1 + maximum_m1 * delta_m1'''		* M1_value = '''minimum_m1 + maximum_m1 * delta_m1'''

			After computing the full values, The curve is to be computed using the Hermite Cubic Interpolation.



			== Quantized-Precision Curve ==
			For this Property Family, the process is almost the exact same. However, there are some changes.
			First of, This is a more precise version of the Quantized curves. Therefore, the values are doubled up.

			===(Example):===
			{\| class="wikitable"
			! Offset !! Value !! Description
			\|-
			\| 0x00 \|\| 0 \|\| This field is telling that '''the curve is animating the bone index 0 of the model.'''
			\|-
			\| 0x01 \|\| 50 \|\| This field is telling that '''the curve animates the X Position of the root bone''', being set on a '''Quantized-Precision Curve Property Family.'''
			\|-
			\| 0x02 \|\| 4 \|\| This field is telling that '''there's 4 keyframes on this curve.'''
			\|-
			\| 0x04 \|\| 512 \|\| This field is '''telling in which offset of the file the keyframe infos are located'''.
			\|}
			===(Keyframes header):===
			{\| class="wikitable"
			! Offset !! Type !! Size !! Description
			\|-
			\| 0x00 \|\| <code>float32_t</code> \|\| 4 \|\| Minimum value of a keyframe.
			\|-
			\| 0x02 \|\| <code>float32_t</code> \|\| 4 \|\| Maximum value of a keyframe.
			\|-
			\| 0x04 \|\| <code>float32_t</code> \|\| 4 \|\| Minimum value of the start tangent (m0)
			\|-
			\| 0x06 \|\| <code>float32_t</code> \|\| 4 \|\| Maximum value of the start tangent (m0)
			\|-
			\| 0x08 \|\| <code>float32_t</code> \|\| 4 \|\| Minimum value of the end tangent (m1)
			\|-
			\| 0x0A \|\| <code>float32_t</code> \|\| 4 \|\| Maximum value of the end tangent (m1)
			\|}
			You can notice that, since the values were doubled up, The use of the custom float16_t were completely removed.
			===(Keyframe):===

Akito at 20:21, 7 September 2025

2025-09-07T20:21:09Z

← Older revision		Revision as of 20:21, 7 September 2025
Line 8:		Line 8:

	Every curve is computed using the '''Cubic Hermite interpolation formula.''' That means every curve is treated as a Hermite Curve.		Every curve is computed using the '''Cubic Hermite interpolation formula.''' That means every curve is treated as a Hermite Curve.
	[[File:Formula.png~~\|thumb~~\|Formula of the Cubic Hermite Interpolation, used in curves.]]		:[[File:Formula.png\|Formula of the Cubic Hermite Interpolation, used in curves.]]

Akito at 20:19, 7 September 2025

2025-09-07T20:19:29Z

← Older revision		Revision as of 20:19, 7 September 2025
Line 5:		Line 5:
	This file incorporates '''curves and/or properties''' for an given embbed model. Because of that, it's almost always certain that they are contained in a file container (dat).		This file incorporates '''curves and/or properties''' for an given embbed model. Because of that, it's almost always certain that they are contained in a file container (dat).
	There's instances of mots that controlled properties instead of bones, so it's not always certain of what it animates. A further inspection have to be done. But since it animates bones most of the times, i will threat it as curves that animates bones.		There's instances of mots that controlled properties instead of bones, so it's not always certain of what it animates. A further inspection have to be done. But since it animates bones most of the times, i will threat it as curves that animates bones.


			Every curve is computed using the '''Cubic Hermite interpolation formula.''' That means every curve is treated as a Hermite Curve.
			[[File:Formula.png\|thumb\|Formula of the Cubic Hermite Interpolation, used in curves.]]


			'''Note:''' This formula's image was taken from the [https://en.wikipedia.org/wiki/Cubic_Hermite_spline ''Cubic Hermite spline Wiki''.]

	== Mot main header (8 bytes length) ==		== Mot main header (8 bytes length) ==
Line 133:		Line 140:
	\| 0x04 \|\| 0.52 \|\| This field is '''telling the value or the Z rotation right away''', no more computing needed.		\| 0x04 \|\| 0.52 \|\| This field is '''telling the value or the Z rotation right away''', no more computing needed.
	\|}		\|}

	~~From now on, every single other curve are handled as hermite curves. See more bellow...~~

	== Quantized Curve ==		== Quantized Curve ==
Line 194:		Line 199:

	So, that would leave us with...		So, that would leave us with...
	P_value = minimum_value + maximum_value * delta_value		* P_value = '''minimum_value + maximum_value * delta_value'''
	M0_value = minimum_m0 + maximum_m0 * delta_m0		* M0_value = '''minimum_m0 + maximum_m0 * delta_m0'''
	M1_value = minimum_m1 + maximum_m1 * delta_m1		* M1_value = '''minimum_m1 + maximum_m1 * delta_m1'''

Akito at 20:04, 7 September 2025

2025-09-07T20:04:52Z

← Older revision		Revision as of 20:04, 7 September 2025
Line 162:		Line 162:
	\| 0x02 \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of a keyframe.		\| 0x02 \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of a keyframe.
	\|-		\|-
	\| 0x04 \|\| <code>float16_t</code> \|\| 2 \|\| Minimum value of the start tangent		\| 0x04 \|\| <code>float16_t</code> \|\| 2 \|\| Minimum value of the start tangent (m0)
	\|-		\|-
	\| 0x06 \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the start tangent		\| 0x06 \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the start tangent (m0)
	\|-		\|-
	\| 0x08 \|\| <code>float16_t</code> \|\| 2 \|\| Minimum value of the end tangent		\| 0x08 \|\| <code>float16_t</code> \|\| 2 \|\| Minimum value of the end tangent (m1)
	\|-		\|-
	\| 0x0A \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the end tangent		\| 0x0A \|\| <code>float16_t</code> \|\| 2 \|\| Maximum value of the end tangent (m1)
	\|}		\|}
	This is the only Property Family that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.		This is the only Property Family that uses the '''float16_t''' type. That's a special type of float that can be represented as two bytes. Especifically, this type of float is built with a 1 bit sign, 6 bit exponent, a 9 bit significand and a bias of 47. The bias favours smaller numbers. That means, the highest the number, the less precise it is. That's to compensate the small amount of data two bytes can pack.
Line 182:		Line 182:
	\| 0x01 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the Keyframe.		\| 0x01 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the Keyframe.
	\|-		\|-
	\| 0x02 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the keyframe's start tangent.		\| 0x02 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the keyframe's start tangent. (m0)
	\|-		\|-
	\| 0x03 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the keyframe's end tangent.		\| 0x03 \|\| <code>int8_t</code> \|\| 1 \|\| Delta value of the keyframe's end tangent. (m1)
	\|}		\|}

Line 191:		Line 191:
	X = Y + Z * W		X = Y + Z * W
	Where X is the output value, Y is the minimum value, Z is the maximum value, and W is the delta value.		Where X is the output value, Y is the minimum value, Z is the maximum value, and W is the delta value.
	Essentially, the formula is telling that the output value is blending from the minimum to the maximum value, based on the value of Delta.		Essentially, the formula is telling that the '''output value is blending from the minimum to the maximum value, based on the value of Delta.'''

			So, that would leave us with...
			P_value = minimum_value + maximum_value * delta_value
			M0_value = minimum_m0 + maximum_m0 * delta_m0
			M1_value = minimum_m1 + maximum_m1 * delta_m1