Filter SDK/avs2pcm - Revision history

Admin at 20:39, 6 October 2014

2014-10-06T20:39:49Z

← Older revision		Revision as of 22:39, 6 October 2014
Line 257:		Line 257:
	__int64 bytes = vi.BytesFromAudioSamples(count);		__int64 bytes = vi.BytesFromAudioSamples(count);

	We will use [http://www.cplusplus.com/reference/cstdio/fwrite/ fwrite] to write 'count' audio samples to a file. So we will need to know the corresponding number of bytes which needs to be written. [[Cplusplus_API/VideoInfo\|BytesFromAudioSamples]] gives the number of bytes and it is calculated internally as follows:		We will use [http://www.cplusplus.com/reference/cstdio/fwrite/ fwrite] to write 'count' audio samples to a file. So we will need to know the corresponding number of bytes which needs to be written. [[Filter_SDK/Cplusplus_API/VideoInfo\|BytesFromAudioSamples]] gives the number of bytes and it is calculated internally as follows:

	{\| style="height:100px" border="1" cellpadding="4"		{\| style="height:100px" border="1" cellpadding="4"

Raffriff42: Filter_SDK -> FilterSDK

2014-09-19T04:38:10Z

Filter_SDK -> FilterSDK

← Older revision		Revision as of 06:38, 19 September 2014
Line 378:		Line 378:
	The remaining lines above are explained in [[Filter_SDK/avs2yuv\|avs2yuv]], so they won't be repeated here.		The remaining lines above are explained in [[Filter_SDK/avs2yuv\|avs2yuv]], so they won't be repeated here.

	[[Category:~~Filter_SDK~~]]		[[Category:FilterSDK]]

Raffriff42: Category:Filter_SDK

2014-09-19T03:53:56Z

Category:Filter_SDK

← Older revision		Revision as of 05:53, 19 September 2014
Line 377:		Line 377:

	The remaining lines above are explained in [[Filter_SDK/avs2yuv\|avs2yuv]], so they won't be repeated here.		The remaining lines above are explained in [[Filter_SDK/avs2yuv\|avs2yuv]], so they won't be repeated here.

			[[Category:Filter_SDK]]

Admin at 21:05, 24 June 2014

2014-06-24T21:05:31Z

← Older revision		Revision as of 23:05, 24 June 2014
Line 361:		Line 361:
	Here an audio sample is stored in one sample (having the size of a multiple bytes), but fwrite will write the same bytes to the file as earlier.		Here an audio sample is stored in one sample (having the size of a multiple bytes), but fwrite will write the same bytes to the file as earlier.

	At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The reason is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.		At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The reason is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing. See [http://stackoverflow.com/questions/92396/why-cant-variables-be-declared-in-a-switch-statement here] for more information.

	env->DeleteScriptEnvironment();		env->DeleteScriptEnvironment();

Admin at 21:04, 24 June 2014

2014-06-24T21:04:12Z

← Older revision		Revision as of 23:04, 24 June 2014
Line 359:		Line 359:
	}		}

	Here an audio sample is stored in one sample (having the size of a multiple bytes), but fwrite will write the same bytes to the file.		Here an audio sample is stored in one sample (having the size of a multiple bytes), but fwrite will write the same bytes to the file as earlier.

	At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The ~~issue~~ is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.		At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The reason is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.

	env->DeleteScriptEnvironment();		env->DeleteScriptEnvironment();

Admin at 21:01, 24 June 2014

2014-06-24T21:01:59Z

← Older revision		Revision as of 23:01, 24 June 2014
Line 284:		Line 284:
	delete[] samples;		delete[] samples;

	There are a few ways to write audio to a file. The simpliest one is the one above. Let's look at what happens with our data:		There are a few ways to write audio to a file. The simpliest one is the one above. Let's look at what happens with our data with an example:

	Tone(length=1, frequency=2, samplerate=48000, channels=1, type="square", level=1.0) # float		Tone(length=1, frequency=2, samplerate=48000, channels=1, type="square", level=1.0) # float

Admin at 21:00, 24 June 2014

2014-06-24T21:00:42Z

← Older revision		Revision as of 23:00, 24 June 2014
Line 359:		Line 359:
	}		}

	Here an audio sample is stored in one sample (having the size of a multiple bytes)~~. But it~~ will write the same bytes to the file.		Here an audio sample is stored in one sample (having the size of a multiple bytes), but fwrite will write the same bytes to the file.

	At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The issue is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.		At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The issue is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.

Admin at 20:57, 24 June 2014

2014-06-24T20:57:04Z

← Older revision		Revision as of 22:57, 24 June 2014
Line 297:		Line 297:
	\|-		\|-
	\| 8Bit (samples between 0 and 2^8-1)		\| 8Bit (samples between 0 and 2^8-1)
	\| s[0] = 255; s[~~end~~-1] = 0		\| s[0] = 255; s[count-1] = 0
	\| FF .. 00 ..		\| FF .. 00 ..
	\|-		\|-
	\| 16Bit (samples between -2^15-1 and 2^15)		\| 16Bit (samples between -2^15 and 2^15-1)
	\| s[1]256+s[0] = 127256+255 = 32767; s[~~end~~-1]256+s[~~end~~-2] = 128256+0 = 32768 (= -32768)		\| s[1]256+s[0] = 127256+255 = 32767; s[count-1]256+s[count-2] = 128256+0 = 32768 (= -32768)
	\| FF 7F .. 00 80 ..		\| FF 7F .. 00 80 ..
	\|-		\|-
	\| 24Bit (samples between 0 and 2^24-1)		\| 24Bit (samples between -2^23 and 2^23-1)
	\| s[2]16^4+s[1]16^2+s[0] = ~~16777215~~; s[~~end~~-1]16^4+s[~~end~~-2]16^2+s[~~end~~-3] = 0		\| s[2]16^4+s[1]16^2+s[0] = 8388607; s[count-1]16^4+s[count-2]16^2+s[count-3] = 8388608 (= -8388608)
	\| FF FF FF .. 00 00 00 ..		\| FF FF 7F .. 00 00 80 ..
	\|-		\|-
	\| 32Bit (samples between 0 and 2^32-1)		\| 32Bit (samples between -2^31 and 2^31-1)
	\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = ~~4294967295~~; s[~~end~~-1]16^6+s[~~end~~-2]16^4+s[~~end~~-3]16^2+s[~~end~~-4] = 0		\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 2147483647; s[count-1]16^6+s[count-2]16^4+s[count-3]16^2+s[count-4] = 2147483648 (= -2147483648)
	\| FF FF FF FF .. 00 00 00 00 ..		\| FF FF FF 7F .. 00 00 00 80 ..
	\|-		\|-
	\| [[Float]] (samples between -1.00000 and 1.000000)		\| [[Float]] (samples between -1.00000 and 1.000000)
	\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 6316^6+12816^4+016^2+0 = 106535321; s[~~end~~-1]16^6+s[~~end~~-2]16^4+s[~~end~~-3]16^2+s[~~end~~-4] = 19116^6+12816^4+016^2+0 = 3212836864		\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 6316^6+12816^4+016^2+0 = 106535321; s[count-1]16^6+s[count-2]16^4+s[count-3]16^2+s[count-4] = 19116^6+12816^4+016^2+0 = 3212836864
	\| 00 00 80 3F .. 00 00 80 BF ..		\| 00 00 80 3F .. 00 00 80 BF ..
	\|}		\|}

	Above the samples are declared as unsigned char (regardless of the number of bits in a sample), but they filled by GetAudio as explained above. Thus an audio sample is stored in multiple samples (each ~~samples~~ having the size of a byte). This is the ~~simplies~~ and cleanest way to do, but you could also have done the following instead:		Above the samples are declared as unsigned char (regardless of the number of bits in a sample), but they are filled by GetAudio as explained above. Thus an audio sample is stored in multiple samples (each sample having the size of a byte). This is the simpliest and cleanest way to do, but you could also have done the following instead:

	int BlockAlign = vi.BytesPerAudioSample();		int BlockAlign = vi.BytesPerAudioSample();
Line 359:		Line 359:
	}		}

	Here an audio sample is stored in one sample (having the size of a multiple bytes). But it will write the same bytes to the file.		Here an audio sample is stored in one sample (having the size of a multiple bytes). But it will write the same bytes to the file.

			At first glance it seems possible to simplify the code above by moving the lines "clip .. delete[] samples" outside the switch statement, but that's not possible. The issue is that each code-block corresponding to a label has its own scope. So the variable samples doesn't exist outside the switch statement and can't be used there for further processing.

	env->DeleteScriptEnvironment();		env->DeleteScriptEnvironment();

Admin at 21:55, 23 June 2014

2014-06-23T21:55:56Z

← Older revision		Revision as of 23:55, 23 June 2014
Line 284:		Line 284:
	delete[] samples;		delete[] samples;

	There are a few ways to write audio to a file. The simpliest one is as above. Let's look at what happens with our data:		There are a few ways to write audio to a file. The simpliest one is the one above. Let's look at what happens with our data:

	Tone(length=1, frequency=2, samplerate=48000, channels=1, type="square", level=1.0) # float		Tone(length=1, frequency=2, samplerate=48000, channels=1, type="square", level=1.0) # float
	ConvertAudioTox() // x = 8Bit, 16Bit, 24Bit, 32Bit and Float		ConvertAudioTox() // x = 8Bit, 16Bit, 24Bit, 32Bit and Float

	The samples are always written to the pcm file as little endian. So this means ~~value = LSB ... MSB~~ (the least significant byte first and the most significant byte last).		The samples are always written to the pcm file as [http://en.wikipedia.org/wiki/Endianness little endian]. So this means the bytes are written in reversed order (thus the least significant byte first and the most significant byte last).

	{\| style="height:100px" border="1" cellpadding="4"		{\| style="height:100px" border="1" cellpadding="4"
	!width=20%\| type (x)		!width=20%\| type (x)
	!width=40%\| value of samples ~~(abbreviated as s)~~		!width=40%\| value of samples
	!width=40%\| bytes written in file		!width=40%\| bytes written in file
	\|-		\|-
	\| 8Bit		\| 8Bit (samples between 0 and 2^8-1)
	\| s[0] = 255; s[end-1] = 0		\| s[0] = 255; s[end-1] = 0
	\| FF .. 00 ..		\| FF .. 00 ..
	\|-		\|-
	\| 16Bit		\| 16Bit (samples between -2^15-1 and 2^15)
	\| s[1]256+s[0] = 127256+255 = 32767; s[end-1]256+s[end-2] = 128256+0 = 32768 (= -~~32767~~)		\| s[1]256+s[0] = 127256+255 = 32767; s[end-1]256+s[end-2] = 128256+0 = 32768 (= -32768)
	\| FF 7F .. 00 80 ..		\| FF 7F .. 00 80 ..
	\|-		\|-
	\| 24Bit		\| 24Bit (samples between 0 and 2^24-1)
	\| s[2]16^4+s[1]16^2+s[0] = 16777215; s[end-1]16^4+s[end-2]16^2+s[end-3] = 0		\| s[2]16^4+s[1]16^2+s[0] = 16777215; s[end-1]16^4+s[end-2]16^2+s[end-3] = 0
	\| FF FF FF .. 00 00 00 ..		\| FF FF FF .. 00 00 00 ..
	\|-		\|-
	\| 32Bit		\| 32Bit (samples between 0 and 2^32-1)
	\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 4294967295; s[end-1]16^6+s[end-2]16^4+s[end-3]16^2+s[end-4] = 0		\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 4294967295; s[end-1]16^6+s[end-2]16^4+s[end-3]16^2+s[end-4] = 0
	\| FF FF FF FF .. 00 00 00 00 ..		\| FF FF FF FF .. 00 00 00 00 ..
	\|-		\|-
	\| ~~float <wrong>~~		\| [[Float]] (samples between -1.00000 and 1.000000)
	\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = ~~4294967295~~; s[end-1]16^6+s[end-2]16^4+s[end-3]16^2+s[end-4] = 0		\| s[3]16^6+s[2]16^4+s[1]16^2+s[0] = 6316^6+12816^4+016^2+0 = 106535321; s[end-1]16^6+s[end-2]16^4+s[end-3]16^2+s[end-4] = 19116^6+12816^4+016^2+0 = 3212836864
	\| ~~(samples between -1.00000 and 1.000000) 1.000000 =~~ 00 00 80 3F~~; -1~~.~~000000 =~~ 00 00 80 BF		\| 00 00 80 3F .. 00 00 80 BF ..
	\|}		\|}

	Above the samples are declared as unsigned char (regardless the ~~samplebit~~) but filled by GetAudio as above~~, meaning~~ an audio sample is ~~written accross~~ multiple samples ~~'s'~~. This is the simplies and cleanest way to do, but you could have done ~~also~~ the following instead:		Above the samples are declared as unsigned char (regardless of the number of bits in a sample), but they filled by GetAudio as explained above. Thus an audio sample is stored in multiple samples (each samples having the size of a byte). This is the simplies and cleanest way to do, but you could also have done the following instead:

	int BlockAlign = vi.BytesPerAudioSample();		int BlockAlign = vi.BytesPerAudioSample();
Line 359:		Line 359:
	}		}

	It will ~~produce~~ the same ~~output~~.		Here an audio sample is stored in one sample (having the size of a multiple bytes). But it will write the same bytes to the file.

	env->DeleteScriptEnvironment();		env->DeleteScriptEnvironment();
Line 373:		Line 373:
	return 0;		return 0;
	}		}

			The remaining lines above are explained in [[Filter_SDK/avs2yuv\|avs2yuv]], so they won't be repeated here.

Admin at 14:31, 23 June 2014

2014-06-23T14:31:06Z

Show changes