So, what is the point of a bit, unless it is to show how little most people know about vinyl and CD's for starters.

Paul

Ahem, why are we even talking about vinyl?

*Gold-plated* relay? Well, at least the silence being inserted was audiophile-quality...



-brendan

I assume that means gold-plated contacts, which actually makes some sense.

Silence is golden.

Cheers, Slipstreem.

Because this bit thing should be dead by now.

Paul

The Tremeloes, 1964, their biggest "near hit".


Seriously, a couple of things got confuted here, but the summary I'd give is

1) It's very hard to get a 1-bit error reading a redbook CD. No or many bits, yes.

2) A one bit error can be audible, the result is purely signal-dependent.

3) Vinyl has error correction, you can lift the tonearm manually.

Yep, best listen to on a 100% jitter free AM pirate station with a Mexican dog howling at the back ground (Caroline, Radio London).


You must be as old as I am. Today there are better solution, I believe it is called scratching


Given 1, do I understand correctly that this is highly unlikely to happen when playing a CD?

To the best of my knowledge, it's approaching the realms of impossibility when playing back a conventional audio CD in a standalone CD player.

Cheers, Slipstreem.

Yep. It's very easy to get no errors. It is possible to get a LOT of errors. Not only the CD, but also the error-checking, has to be in a very unusual and hard to reach state to get exactly ONE bit in error. (Effectively, this would mean a lot more are "wrong" in the sense that they couldn't be recovered exactly, which is hard, and they happen to randomly be right, so to speak.)



Dude, I work with a guy who used to run the remains of Astatic!

Dang, I lose internet access for a week and I miss some of the BEST discussions.

I remember wanting to test my Magnavox 2040 (c. 1984) error correction when it was new; I stuck a piece of black electrical tape about 1 cm wide across the CD surface and played it. The player didn't even hesitate. Not bad for what was a 2nd generation (14-bit) player.

More of interest to the thread, I have a Robert Fripp CD that exhibited an actual click every time the CD rotated; so as the disc played the clicks came further and further apart. The noise did not seem to be from an output relay, but rather residual from the error correction. It sounded for all the world like a scratch on an LP, so I hung on to it for the amusement value. This happened on the abovementioned player as well as my Sony D-5 (the original portable player!). I need to run it on one of my newer players to see if their error correction will deal with it any better. I could never perceive any visual artifacts in the disc data layer, so I assume it was a flaw in the pit/land surface when it was pressed. It must have been huge, but I have no idea why it didn't mute the players. They will mute when I play my special lab test CDs on them.

I kept track of defective (as in audible defects) CDs for a few years, and I got about 2-5% of my purchases showing some kind of problem until about 1987. QC improved after that; I stopped using my oldest CD players about 1995 or so except for special instances so that's about the end of my data log. Today's players will play almost anything it seems, but I only have about 3 discs left that gave me errors to test them with.

So the time it takes to load something affects the quality of said something in a bad way. And here I was thinking hd/blu-ray and sd dvd was actually better than vhs. Stupid me.






(sarcasm)

If memory serves me, once the first block of data is found in the directory, it and not the directory points to the next block of data, so the speed of reading a file has nothing to do with the directory.

Paul

After corrupted bit, if the samples are as follows :

00000000-0000000/00000000-00001000/00000000-00000000/……………

--sample1---------/--corrupted sample--/---sample3----------/……………….

then it implies that there is presence oh a significantly higher frequency component, which is out of human audible range of 20KHz.

So I think we won't be able to hear that bit.

If sample1 and sample3 values are not forming any high frequency component when combined with sample2(corrupted one) then we will be able to hear that.

Cheers!
..Aniruddha

Of course this depends on samples 1 and 3, but an impulse of that kind will create a very wide-band sound, which is likely to be clearly audible. For example, if you have a section of digital silence with a one bit (say MSB) error in it, the difference will be spread over the entire audio band. Just because the sample is only 1/44100 seconds long doesn't mean that it only implies spectral information over 20kHz.

yes, thats true.

I created a .pcm file with all zeros except 1 sample having a non-zero value.
When I played the file, i could hear the 'khssk' sound.

That is a deliberate change, and not one that would happen when playing a CD.

Big difference.

Paul