AUdIoCoUrSeS
Joined: 31 Oct 2002
Posts: 2014
|
| Week 5 - Error handling systems |
 |
|
Rico here we go
Error handling systems
With this weeks questions i'm looking for much more elongated answers and details. Look to answer these very fully, particularly the last question.
1. What is the sub-code area of the DAT tape used for?
2. Explain the principles of predictive coding.
3. How does the data buffer of a Minidisc player facilitate editing?
4. What are the basic principles of masking?
5. Explain the the following, with reference to a HD editing system
(a) disc access time
(b) EDL
(c) Disc data bandwidth
6. Explain briefly the problems of transferring data files between different proprietors HD systems.
7. Explain the principles used in editing on a single Mini Disc system.
8. Describe and explain the following error handling systems
• Causes of errors
• Minimisation of the consequences of errors
• Error detection
• Error correction
• Error concealment
_________________
It's all in the ears. - Learn the concepts not the software.
Audio Courses is a way into the music business for you
|
Mon Apr 04, 2005 9:51 am |
|
|
Rico1210
Joined: 03 Aug 2004
Posts: 39
Location: Newcastle, UK |
|
|
|
Hi,
here are my answers for this week
1. What is the sub-code area of the DAT tape used for?
Subcode is an auxiliary data stream, which is merged with audio samples, and has numerous functions. A DAT can carry 273 kbits per second of subcode, which is about four times the amount a CD contains. The sub-code area of the DAT tape is where the time is recorded. The time from the beginning of the tape is recorded in every track, therefore if you start recording in the middle of a tape, the machine reads the time at that spot and keeps recording time sub-code appropriately. ID's are also recorded in the sub-code. These are points that contain digital code that can be found easily. They are Start ID's, Skip ID's (instruct the deck to find the next start point) and End ID's. Program numbers can also be recorded in the sub-code. These are numbers searched for during previous play operations. Sometimes Program numbers are included with each Start ID. The subcode area of a DAT also contains the AES/EBU channel status message.
2. Explain the principles of predictive coding.
In all signals, part of the signal is obvious from what has gone before and what may come after. Most signals have a degree of predictability, for example a sign wave is highly predictable because all cycles look the same. A signal can be transmitted with parts omitted(encoded), provided that there is a suitable decoder that can predict the omissions from the previous and next data. All encoders must contain a model of the decoder to be safe in the knowledge that the information will be correctly re-created. Predictive codecs contain two identical predictors, one in the coder and one in the decoder. These predictors examine the previous data values and estimate what the next value will be. The estimated value is then subtracted from the actual next value to produce a prediction error(residual) that is transmitted from the encoder. The decoder receives the prediction error and adds it to its own prediction, which produces the output code value. There will be no loss of information provided the prediction error(residual) is transmitted intact. However, not all signals can be correctly predicted. All codecs have difficulties with noise, as noise is totally unpredictable.
3. How does the data buffer of a Minidisc player facilitate editing?
The data buffer of a Minidisc player controls the data rate of playback. A Minidisc stores data in sections much like a Hard Disk drive does. After editing, the tracks or part of a track may not be located in order on the disc. As the data buffer controls playback, the information is read from it's location and produced with continuity by the RAM buffer.
4. What are the basic principles of masking?
The human hearing system is not equally sensitive at all frequencies, it is less so at low and high frequencies. The perception threshold of hearing is raised at a particular frequency in the presence of another sound at a similar frequency. In simpler terms, when two sounds of a similar frequency are played but only one can be heard, the first sound 'masks' the second. The perception threshold is raised, so a sound must be louder to be heard. Masking has its uses in audio engineering and is widely used in noise reduction. For instance, low-level noise that exists in the same frequency band as a high-level music signal will be masked by the music. Masking is also used in digital compression systems, as it allows the use of lower resolution in frequency bands where the noise will be masked by the signal.
5. Explain the the following, with reference to a HD editing system
(a) disc access time - The disc access time is how fast the hard disk can access files. The time it takes the pickup to locate the block and for the disc to spin to the pickup. The disc access time of a hard disk editing system is important because during an edit, the drive has to provide audio files from two different areas of the disk simultaneously. If there are lots of edits in a short space, the drive will have difficulty accessing the files ahead of real time. This is where latency can occur. A large memory buffer can help the system because the drive can fill up the memory with the files before the edits begin.
(b) EDL - An Edit Decision List (EDL) contains the instructions that are used to control the replay of sound file segments from the disk, as well as crossfade period and gain parameters. Hard disk systems use non-linear editing which is non-destructive. The original sound files remain intact, any editing only exists as a set of instructions that replay parts of the original sound files and overlay any signal processing. The files are identified with 'out' and 'in' edit points
(c) Disc data bandwidth - The disc data bandwidth is the rate that data can be transferred from, and to the hard disk. Digital bandwidth is measured in bytes or bits per second.
6. Explain briefly the problems of transferring data files between different proprietors HD systems.
Compatibility is an issue when transferring data files between different proprietors hard disk systems, because each manufacturer has a different way of describing edit decision lists(EDL), cross fade files and dynamic signal processing (DSP) functions. The data itself is relatively universal, but how it is processed is different. The differences between hard disk systems is being researched and the Open Media Framework is an attempt to reconcile these differences.
7. Explain the principles used in editing on a single Mini Disc system.
The Minidisc system stores data in sections much like a Hard Disk drive. When a track is edited it, or part of it may not be located in order on the disc. As the data buffer controls playback, the information is read from it's location and produced with continuity by the RAM buffer. This is a simple edit decision list (EDL) system
8. Describe and explain the following error handling systems
• Causes of errors - There are two typical types of error: the burst error and the random error. Burst errors can cause the loss of a number of successive samples. A burst error is often due to a momentary loss of signal caused by a tape drop-out, an electrical spike or by a mark or dirt. A random error is the loss of a single sample in a random position. A random error is usually the result of noise or poor signal quality.
• Minimisation of the consequences of errors - The consequences of errors can be minimised by using processes known as interleaving or redundancy. Interleaving minimises error correction by breaking up burst errors into smaller random errors that can be more efficiently corrected. This is achieved from recording samples out of their normal sequence. Taking columns from a memory which was filled in rows is one process of interleaving. Another is to shuffle the sample sequence to separate odd and even numbers, this way only either even or odd numbers will be lost. Upon replay the process is reversed (de-interleaved) putting samples back into their regular sequence. Redundancy involves the recording of data in more than one place. The twin DASH format uses redundancy to minimise the consequences of errors. All audio data is recorded twice, with the odd/even technique reversed on the second recording. The result is two separate interleaved recordings. If their is a burst error in the odd numbers, this will affect the odd numbers of both recordings. As the odd/even numbers are reversed on the second recording, the lost data of the first recording is actually in the even numbers of the second.
• Error detection - Cyclic redundancy check (CRC) codes are used in many systems to detect errors on replay. These codes are calculated from the original data and recorded along with that data
• Error correction - There are four types of error correction, the first being true error correction which reconstitutes erroneous samples perfectly. These corrected samples are indistinguishable from the originals, therefore sound quality is unaffected. The second type of error correction is interpolation. Interpolation is used when the error rate exceeds the limit of perfect correction. An interpolated sample value is a mathematical average of the sample before and the sample after. This process is also known as concealment or averaging. Where interpolation is impossible, the error correction system will perform a Hold. A Hold is when the last correct sample is repeated. A system will usually only hold for a few samples before muting. Muting is the last option of error correction. When the error correction system is totally overwhelmed, the audio output will be muted for the duration of the error. Sometimes it would be preferred to hear the error, which may be a 'click' or 'spit', rather than have the output muted.
• Error concealment is a process where the value of the missing sample is estimated from nearby samples. The nearby sample may not necessarily be the same as the original, which in some cases can cause the concealment to be audible. Concealment is made possible by shuffling the sample sequence to separate odd and even numbers prior to recording. The odd and even samples are recorded in different places and recombined into their natural sequence upon replay. Therefore if there is a burst error, only either even or odd numbers will be lost, the result being that every other sample is correct. Only half of the waveform can now be described, but it can be reproduced, though with a little loss of accuracy.
References
The Art of Digital Audio - John Watkinson
Sound & Recording - Francis Rumsey, Tim McCormick |
|
Sat Apr 09, 2005 10:34 am |
|
|
|
|
|
|
All times are GMT. The time now is Fri May 16, 2008 4:14 pm
|
|
|
|
| |
