I have actually tried that! I have bought a lot of CDs used for a low price and sometimes these CDs have severe enough scratches to cause playback issues. Sometimes "toothpaste treatment" is enough to fix the issue, but not always. So once I tried if freezing the disc does anything, but it did not.
You are sensitive to low-quality sample rate conversion. Arbitrary sample rate conversions (e.g. from 44100 Hz to 47133.91 Hz) can be done totally transparently to human ears and I have myself written MATLAB code utilising sums of sinc-interpolations for that.
No, it’s the fundamental question that’s BEFORE the concept of digital transmission. Before we can consider transmission, we obviously have to know what we are going to transmit. Once we know that, then we can design a transmission system that meets (or exceeds) that requirement. This is what Shannon’s theory is all about, what and how much data do we need in order to guarantee communication of information, EG. What “Channel Capacity” is required. Only then can
Absolutely, and that is down to the designers of the protocol used to transfer the data but again, this will be based on Shannon’s theory. Additionally, there must be some way of detecting errors in the first place, regardless of the error correction (or lack of it).
Fundamentally, there is no difference. With a CD player for example, the data is read by the reader/laser transmitted to a buffer where errors are detected (and corrected up to a defined tolerance of consecutive errors). Same principle with a synchronous stream or an asynchronous/packetised stream.
No, it’s the same. In theory the faster the transmission the more errors are likely to occur and therefore the more error correction required but in practice faster transmission protocols incorporate specifications to mitigate the higher likelihood of errors.
As far as I’m aware, this occurs as the receiver and transmitter “negotiate” an image quality based on connection speed. The transmitter/server initially supplying a low res image quality until a sufficient “Channel Capacity” is established/confirmed. In many cases this can be re-negotiated, for example the system may detect a fall in “channel capacity” and reduce image resolution. TBH, I’m not entirely certain of the techniques employed by various streaming services. It depends on whether they think their consumers would rather see a low res image or see nothing at all and wait.
That’s because DD+ uses an already highly optimised codec which can’t easily be reduced in size and doesn’t require much bandwidth/channel capacity compared to HD or UHD video any way.
A PC or laptop with a decent CD rom drive and good ripping software can recover all but the worst damaged CD’s ,
In practice, in regards to device that is receiving a digital transmission, its error correction is happening after it receives data. I was pointing out that transmission and error correction are different concepts. Granted, error correction schemes have the transmitting device send code for the the receiver to detect errors. This is part of the data, and not a component of the cable. There are also separate error correction systems during the reading stage of a CD player, or some computer memory chips.
Fundamentally, there are different error correction schemes, using different error correcting algorithms. The error correction in a CD player uses Cross-interleaved Reed-Solomon Code. HDMI is asynchronous like a network cable, and uses BCH error correction. Forward error correction being a standard error correction scheme.
You don't think h.264 and h.265 video streams are highly optimised codecs for video? When it comes to availability of video content based on specs, that's decided by the service's application. Netflix will make 4K Dolby Vision and Atmos content available on my Apple TV because my Apple TV is setup for it. 4K content used to not be available on internet browsers, but eventually was (when their HTML5 players began supporting HDR/Atmos codecs). The reason it takes longer to buffer video (and why the initial packets off the internet may have noticeable artifacts), is that the video stream is inherently larger than audio (and it might have noticeable artifacts as the buffer gets to full speed). My TV will still say it's starting HDR or Dolby Vision when the video starts: so it's not first setting video at SDR 1080P. I believe what you're describing in "handshaking" what size resolution the streamed video is what happens when you play a YouTube video with "auto" resolution. You can still overrride it if you want: there's a setting to go to a particular resolution. And if your internet speed isn't good enough, then you'll see artifacts. If your computer isn't fast enough at decoding and processing the video, then you'll have diminished frame rate.
If error correction is required then of course it must occur in the receiving device as the transmitting device obviously cannot know if there will be any downstream errors after it’s transmitted the data. The data that’s being transmitted obviously does have to be involved in the error correction before it leaves the transmitting device though, so that the receiving device can detect errors and where necessary, it also includes error code data to enable a receiver to carry out it’s own error correction.
And Shannon pointed out they weren’t, they’re integral.
Good example. RS code allows both error detection and correction in the receiving device but is embedded into the data before transmission. A CD carries audio data at the rate of 1,411,200bps (16 x 44,100 x 2) but if we design a transmitter with that “channel capacity” it will fail, because in addition to the audio data there needs to be “redundancy”, the Reed-Soloman code to detect and correct errors. The “Shannon Limit” dictates the minimum “channel capacity” depending on the level of redundancy required, EG. The expected worst case amount of noise/interference.
Very true but aren’t you arguing against yourself? Forward error correction is where code is embedded before transmission, such as your Reed-Soloman example. And even where FEC isn’t employed, code still has to be embedded (before transmission) for error detection, a parity bit and/or CRC for example. While I don’t know the protocols used in every device, say between the RAM and CPU cache, all the audio protocols I’m aware of include at least error detection code.
Of course but then a streaming service can easily switch between say a h.265 stream containing a 4K image res or a 720p res image, you can’t do that with DD+.
It seems you are taking me out of context here. There is no error correction happening in the cable itself. I don't know how many times I have to restate that a data stream is binary. With a serial connection, it's one binary stream, and parallel has more than one lane. Error correction is fundamentally different than data connection, as it deals with the actual data. They may both be used in telecommunication, but they are not always integral with each other and they are different (example of a data connection without error correction in areas of a data bus).
No, it seems you are wanting to argue semantics. I maintained that there are different error correction schemes and they happen in different places. The error correction happening in a CD player is for potential data loss on the optical disc. That's separate than what it does for sending data through a digital cable. It's also separate from the receiving device (in which itself is receiving data and forming a "bit perfect" data stream). If you want to continue to argue about error correction schemes also needing to have a "handshake" with the transmitting device...this is horse before cart that's pointless to argue.
Again, that's not what's happening in my example. h.265 is a recent codec created for 4K HDR (that requires quite a bit more data than a 1080P video). The signal that you're getting is 4K (as I indicated: the video first starts with HDR). With video protocols, there can be artifacts in a picture when data rate is not fully reached yet. It's not that the image is downgraded to 1080P, but as JPEG, the whole image is formed progressively (and if there's not a full image, pixels get blocked). The handshake that has already happened is saying "send the 4K video stream instead of the 1080P one". Again, the one main example I can think of where resolution changes is the "auto" setting on YouTube.
As I was attempting to explain Shannon’s theory, how it provides immunity from transmission noise/interference and as that’s what you quoted when you responded to me, I naturally assumed that was the context, have you changed the context?
Where did I say there was?
No idea why you would feel the need to do that, in fact I don’t know why you would state it even once to me, seeing as I’m the one referencing the “Mathematical Theory of Communication”! Maybe you haven’t read it, don’t understand it’s implications or don’t realise it’s the fundamental basis of digital communications?
Refer back to post #88: I explained what a digital signal is. It's a binary stream. You have continued to conflate it with error correction, which is an entirely separate topic that deals with the data that is in a binary stream. They may both be subjects used in telecommunications, but they are not the same thing!
You're right, this is going nowhere as you have falsely maintained that error correction is always integral with a digital connection. Refer back to my example of consumer RAM not having ECC and server RAM having ECC.
That makes no sense. Are you taking about a binary stream that’s just random zeros and ones which do not represent information, if so, what’s the point of such a stream? If that binary stream does represent information then it has to follow a transmission protocol, which in turn guarantees immunity from noise/interference as per Shannon’s theory. I don’t know all the protocols out there but certainly all the audio protocols require at least error detection.
It is going to go nowhere if you’re just going to make up falsehoods about what I “falsely maintained”! Clearly you haven’t read Shannon’s theory. There are cases, even with audio protocols, which do not need error correction, there might be cases that don’t even need error detection but I don’t know of any such protocols and all the audio protocols I’m aware of contain data in the bitstream to detect errors.
Why can't you understand that a digital connection is an electronic connection? It does not do any coding of digital information by itself. Go on and accuse me of not understanding Shannon's theories, conflating different topics, and do gross generalizations that may not be accurate (like saying decreased visual quality in streaming video are from reduced resolution of stream).
Is an optical connection “electronic”? Why can’t you understand that it doesn’t matter, as long as it complies with the specifications of the protocol?
Again, where did I say a cable/connection codes digital information?
Go on and falsely accuse me for a 3rd time of claiming cables/connectors code digital information or that the information being transmitted isn’t binary and failing to conflate the cables/connectors with the protocol.
Is an optical connection not a binary stream of data??
You keep conflating it with error correction schemes.
Did either of us say that digital information isn't binary? You have previously stated the binary stream and an error correction system are always integral. They are not, and why there's a fundamental difference between data transmission (which is either serial or parallel connection), vs the interpretation of data. That you refuse to acknowledge these differences, it is pointless to engage further.
Why ask? What protocols does a binary stream of digital audio data through an optical connection have to comply with?
You keep failing to conflate cables/connectors with the digital protocol specifications.
Then why did you just ask if optical is binary? What’s the obvious answer to this question?
No, I stated that the binary stream always contains error correction or detection code and therefore obviously it is “integral”. There is a condition where it may not be, Shannon’s “Noiseless Channel” but I’m unaware of any circumstances where that is applicable in practice and it’s certainly not applicable to digital audio data transmission.
How is it data transmission if it’s the wrong data? If the “interpretation of data” is incorrect then you have not transmitted the data/information! That’s the whole point of digital transmission in the first place.
Why ask? Because you continue to conflate data transmission (defined by serial, parallel, synchronous, and asynchronous data) with the actual data stream itself.
No, I'm not failing, it is you who are conflating the connection cable with the protocols of the system
No again, it's not necessarily in a binary data stream. It may be part of the audio protocols for a toslink cable, but not all digital connections employ an error correction. I previously listed ECC vs non ECC computer memory. USB has modes that don't use ECC.
