More 'Audiophile' bullsh*t
Discussion
Seeker UK said:
That's true but it was a question about transports. A £12 transport will probably generate more jitter than a £1000 one.
If the tests in HiFi world are anything to go by, quite frequently expensive high end transports do generate a lot more jitter than expected.There are many good, cheap, low jitter transports out there - much of the Cambridge Audio stuff tends to yield low jitter figures.
DiyAudio has a thread that spans several hundred pages, describing how to use a cheap JVC boom box as a CD transport.
Recovering data off a CD reliably is so trivial these days, with modern machine tolerances.
Unlike 30 years ago or so, when tight tolerances were quite difficult to achieve.
When you think about Blue Ray - and the density of the data stored on a Blue Ray disk - it makes CD look quite silly really (in terms of data density).
And yet, even the cheapest of Blue Ray players manage to recover the data without significant* error.
- not significant enough to cause audio / picture dropout.
TonyRPH said:
If the tests in HiFi world are anything to go by, quite frequently expensive high end transports do generate a lot more jitter than expected.
There are many good, cheap, low jitter transports out there - much of the Cambridge Audio stuff tends to yield low jitter figures.
My newest bit of HiFi is about 15 years old so my knowledge stops about then There are many good, cheap, low jitter transports out there - much of the Cambridge Audio stuff tends to yield low jitter figures.
StuH said:
Globs said:
It's easy to criticise rather than explain but frankly it's extremely rude, although I can see the attraction of implying that you know more than I do.
Perhaps instead you tell us what you disagree with if you can? Maybe then we'll all learn something.
Apologies globs - I just read my post back and it came across a little more disparaging than intended. It's just that until today I've found your posts very good but today you seem to have come over a little holier than though on the high-end bashing front. Anyhow, I'm not going to get into a debate on the whole digital transport/dac/jitter/clocking subject as its apparent we have differing views. If the guys that work at the very cutting edge of digital audio like those at DcS don't have all the answers then I respectfully suggest that no one on this forum does either. I just go with what my ears tell me Perhaps instead you tell us what you disagree with if you can? Maybe then we'll all learn something.
Jitter of course will affect the sound because it moves 'the dots'
In my experience, it doesn't matter which forum you frequent, opinions always vary wildly, and many are expressed with great fanaticism.
Just in the same way that some people here on PH think that car 'X' is fast and others say car 'Y' is faster.
There are simply too many variables - in the case of the car, the driver / fuel / tyres etc. etc. - in the case of audio, the listening room, personal taste, source material (and the list goes on).
I have even heard several different pressings of the same CD sound grossly different (when sourced from the same master) in the same system.
I once owned two Marantz CD17 players - both the same model but not necessarily form the same batch - and even they sounded different to each other in the same system with the same cables etc.
Hence much of the argument / discussion / debate (call it what you want) on topics such as this tend to remain open ended.
I also find that many people find it difficult to be objective - particularly those who have spent £1000's on cables, mains filters etc.
I just like to sit back and enjoy the music - because after all, that's what its all about for me now.
Back in my 20's, I used to build and modify kit with some fanaticism - and then spend hours trying to hear the difference - I lost sight of the real purpose to listening to music. I woke up in my mid 30's (after losing much of my system in a house move getting it replaced bit by bit due to no insurance) - and I realised that even basic kit can sound good, because it's all about the synergy.
Just in the same way that some people here on PH think that car 'X' is fast and others say car 'Y' is faster.
There are simply too many variables - in the case of the car, the driver / fuel / tyres etc. etc. - in the case of audio, the listening room, personal taste, source material (and the list goes on).
I have even heard several different pressings of the same CD sound grossly different (when sourced from the same master) in the same system.
I once owned two Marantz CD17 players - both the same model but not necessarily form the same batch - and even they sounded different to each other in the same system with the same cables etc.
Hence much of the argument / discussion / debate (call it what you want) on topics such as this tend to remain open ended.
I also find that many people find it difficult to be objective - particularly those who have spent £1000's on cables, mains filters etc.
I just like to sit back and enjoy the music - because after all, that's what its all about for me now.
Back in my 20's, I used to build and modify kit with some fanaticism - and then spend hours trying to hear the difference - I lost sight of the real purpose to listening to music. I woke up in my mid 30's (after losing much of my system in a house move getting it replaced bit by bit due to no insurance) - and I realised that even basic kit can sound good, because it's all about the synergy.
budgie smuggler said:
Surely jitter can't be a problem in the real world though, it would be very audible as static/clicks when the clocking error affected the MSB or ones near it.
It's not that extreme to miss a bit, it's a more subtle jitter of getting the right bits but at slightly the wrong time.
A digital representation of an analogue signal is a compression because there is a loss of information. For CDs with a simple rate of 44.1kHz the highest theoretical waveform available is just below 22.05kHz (at exactly 22.05kHz the phase information disappears). This information may not be audible but nevertheless your speaker cannot replicate the steps of the digital waveform because that needs to/can only move smoothly back and forth (the green line in the graph).
There is therefore a stage where _something_ has to join the dots (see the dots in the graph?) - which represent the points in time and voltage that the waveform must pass through for a good sound. Jitter affects the exact timing, if those dots are shifting left and right even just a little it alters the shape of the waveform leaving the DAC.
So the clock needs to be rock solid, and ALSO follow the vagaries of the bit stream to remain synced without falling too far behind, so re-clocking is a clever process to match up to the bitrate and fix each dot in its proper time. My cheap pro-audio behringer allows me to clock the incoming signal to itself, to the DAC or to an external clock signal - because this is an everyday thing in pro-audio - to run all the digital parts from the same master clock that is stable and doesn't frequency modulate extra signals onto the audio.
CDs of course have the other problem of Low Density Bitrate at low volume levels - i.e. listen to something 64dB down and you are listening to a 4bit signal - which again presents issues with joining the dots. This happens because we hear sound logarithmically but CDs encode with a linear 16 bit number (each bit gives 6dB of change), so a loud signal is rendered with a high degree of accuracy, but a qulet signal looks more like a lego building set going into the DAC. Dithering shifts this problem into the frequency domain to an extent but doesn't solve it. This is another reason why DVD sound and SACD will sound better - with 24bit each of those clunking low level steps becomes 256 steps. (8 bits = 256 levels).
CDs can sound very good however, most of the crap from CDs is due to poor over-level mastering rather than the format limitation, no digital waveform was ever supposed to be masted with clipping, but substantial clipping is commonplace these days - which gives the DAC a whole (hole) new section to invent every few milliseconds.
Edited by Globs on Monday 7th January 10:04
custodian said:
budgie smuggler said:
Surely jitter can't be a problem in the real world though, it would be very audible as static/clicks when the clocking error affected the MSB or ones near it.
This is not true. Jitter will cause a loss of sound quality long before that. Don't get fooled by the old ones and zeros argument Edit: Globs, I am a programmer, I have written RT audio device drivers and know how PCM works.
Jitter absolutely is enough to flip a bit anywhere within the 16 bit value. It depends on what the previous bit was in the bitstream and a certain amount of luck. For example your wave form above could become this if the timing error was in the right place:
Edited by budgie smuggler on Monday 7th January 10:29
Edited by budgie smuggler on Monday 7th January 10:42
A few months ago I was talking to the chairman of the company I work for. He said there was more error correction in the original cheap CD decks than is built in to the most expensive cd decks these days.
As he used to own a recording studio before moving into the manufacture of mixing desks for other studios I think he may have some practical insight.
As he used to own a recording studio before moving into the manufacture of mixing desks for other studios I think he may have some practical insight.
FlossyThePig said:
A few months ago I was talking to the chairman of the company I work for. He said there was more error correction in the original cheap CD decks than is built in to the most expensive cd decks these days.
As he used to own a recording studio before moving into the manufacture of mixing desks for other studios I think he may have some practical insight.
As far I can remember from my rusty old CD player theory - the Reed–Solomon error correction used in CD players is not applied in a 'more' for player 'x' than player 'y' system.As he used to own a recording studio before moving into the manufacture of mixing desks for other studios I think he may have some practical insight.
It is a system that detects errors in playback, and provides correction.
I suspect your chairman may well be referring to oversampling, or possibly single bit vs. multi bit dacs.
I am from an IT background and know nothing about how a hi-fi CD transport may differ, so...
...I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
So, to my mind, it *must* be possible to extract the raw PCM data from a CD with zero bit rates. It must also be possible to send this data, lossless, to a DAC/processor, which can them assemble it into a nice big buffer and THEN that DAC/processor can do everything it needs to do to generate some kind of waveform, generating intermediate samples as required. Obviously at this final stage, with processing going on to upsample and create the analogue waveform, there is the potential for various quality thresholds in the process.
But I don't understand the need for all these locked clocks and so on. That would be necessary if the whole system was purely streaming end-to-end, needing on-the-fly processing, but why can't the DACs just have a suitably sized buffer from which they do all their processing? That would eliminate the need for all this clock locking and given that we KNOW that ANY CD transport can generate a perfect digital image of the bitstream, it would all come down to the DAC.
I know this isn't the case in the world of hi-fi, but can anyone tell me what I'm missing?
...I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
So, to my mind, it *must* be possible to extract the raw PCM data from a CD with zero bit rates. It must also be possible to send this data, lossless, to a DAC/processor, which can them assemble it into a nice big buffer and THEN that DAC/processor can do everything it needs to do to generate some kind of waveform, generating intermediate samples as required. Obviously at this final stage, with processing going on to upsample and create the analogue waveform, there is the potential for various quality thresholds in the process.
But I don't understand the need for all these locked clocks and so on. That would be necessary if the whole system was purely streaming end-to-end, needing on-the-fly processing, but why can't the DACs just have a suitably sized buffer from which they do all their processing? That would eliminate the need for all this clock locking and given that we KNOW that ANY CD transport can generate a perfect digital image of the bitstream, it would all come down to the DAC.
I know this isn't the case in the world of hi-fi, but can anyone tell me what I'm missing?
PhilboSE said:
I am from an IT background and know nothing about how a hi-fi CD transport may differ, so...
...I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
IIRC, due to the real time nature of reading a music CD, there is no time for re seeks....I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
PhilboSE said:
So, to my mind, it *must* be possible to extract the raw PCM data from a CD with zero bit rates. It must also be possible to send this data, lossless, to a DAC/processor, which can them assemble it into a nice big buffer and THEN that DAC/processor can do everything it needs to do to generate some kind of waveform, generating intermediate samples as required. Obviously at this final stage, with processing going on to upsample and create the analogue waveform, there is the potential for various quality thresholds in the process.
Part of the problem with CD players is that the information has to be retrieved in real time.A computer CDROM reading a data* disc* can simply retry multiple times until it either managed to recover the data, or fails.
If a CD player can't read the information first time, it simply fails, and bit(s) lost.
IIRC a few years back Meridian made a CD player that did actually cache the reads, giving it the opportunity to re-read multiple times (I'm not sure what the delay was - but think it was in the order of millseconds).
- of course an audio CD is 'just' data as well - but I used the term data (meaning computer data) for the purposes of the explanation.
PhilboSE said:
But I don't understand the need for all these locked clocks and so on. That would be necessary if the whole system was purely streaming end-to-end, needing on-the-fly processing, but why can't the DACs just have a suitably sized buffer from which they do all their processing? That would eliminate the need for all this clock locking and given that we KNOW that ANY CD transport can generate a perfect digital image of the bitstream, it would all come down to the DAC....<snip>
See the explanation above, re: Jitter.- Disclaimer: My CD player theory is really rusty (I did all this back in the mid 80's) so happy to be corrected.
Edited by TonyRPH on Monday 7th January 16:47
PhilboSE said:
So, to my mind, it *must* be possible to extract the raw PCM data from a CD with zero bit rates.
Apart from the zero time for re-reads issue, you also need to keep in mind the nature of CD is that unless you're very careful, the transport has to deal with all sort of stuff between the laser and the pits in the disc - greasy fingerprints, scratches, jam etc. There are test discs with increasingly sized manufactured scratches built into the disc. These can be used to establish how big an issue a given player can cope with. This isn't the same for all players due to various factors - stuff like how stable the platter is and therefore how much work the servo has to do to keep the laser tracking the "groove" in the first place, how well designed the servo tracking circuit is, how well regulated the power suppies are etc. etc.budgie smuggler said:
For example your wave form above could become this if the timing error was in the right place:
This however has never been observed in CD transports - diyaudio IIRC did a big test of bit accuracy and I don't recall it ever happening.I'm sure it's possible with ad hoc systems, but you'd need a jitter big enough to skip an entire bit/half bit which for even a cheap quartz crystal isn't going to happen.
The subtle timing of _when_ the value hits the DAC however will always be slightly out which is why a more stable clock will always improve sound.
clived said:
Apart from the zero time for re-reads issue, you also need to keep in mind the nature of CD is that unless you're very careful, the transport has to deal with all sort of stuff between the laser and the pits in the disc - greasy fingerprints, scratches, jam etc. There are test discs with increasingly sized manufactured scratches built into the disc. These can be used to establish how big an issue a given player can cope with. This isn't the same for all players due to various factors - stuff like how stable the platter is and therefore how much work the servo has to do to keep the laser tracking the "groove" in the first place, how well designed the servo tracking circuit is, how well regulated the power suppies are etc. etc.
In that case, lossless audio formats not stored on an optical disc are the future? Seems crazy to spend all the time, money and effort to make a bad format better. There are test discs with increasingly sized manufactured scratches built into the disc. These can be used to establish how big an issue a given player can cope with. This isn't the same for all players due to various factors - stuff like how stable the platter is and therefore how much work the servo has to do to keep the laser tracking the "groove" in the first place, how well designed the servo tracking circuit is, how well regulated the power suppies are etc. etc.TonyRPH said:
PhilboSE said:
I am from an IT background and know nothing about how a hi-fi CD transport may differ, so...
...I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
IIRC, due to the real time nature of reading a music CD, there is no time for re seeks....I simply don't understand how a music CD *should* be affected by re-seeks, error correction etc. If the same CD player wasn't capable of streaming the data for a computer executable with 100% reliability, it would not be fit for purpose. Even a single bit error would be enough to cause problems.
In fact many CD rippers have this facility and they rip a lot faster than the CD would take to listen to.
Many people are stuck in a world 3 decades old with thinking about CDs, a modern computer CD reader is simply light years ahead of anything available in the 1980s, in fact a £5 DVD reader would embarrass any 80s or 90s CD transport with speed and accuracy - comparing the two is almost ridiculous
Ripping a CD on a computer will get you as close to bit perfect as you are ever going to get from a CD, which most of the time is in fact bit perfect to the master. The days of paying over a tenner for a CD reader are long gone, raw/Flac CD files now sit on hard disks with total error correction to computing standards (i.e. _exact_). The real challenge is how to turn that data into decent analog, because the data is fairly sparse and there are challenges. The best way I have heard is to upsample and re-clock with a decent DSP/clock, and then feed into a decent DAC at that high (re)sampling rate.
CD transports do not and cannot alter the sound unless something is very wrong with the transport, and if there are any errors a PC reader will always be far better (assuming the reader is not faulty). Remember a PC CD drive can read in a DVD at high speed, a clunky old CD is a doddle. Paying money for an expensive CD transport is a total waste of cash, unless you want to donate to the factory that made it for other reasons.
Globs said:
For a 1x player that is true, for a 40x computer CD/DVD reader that is not true at all because the computer will already be well ahead of the time-point in the CD track (buffered in memory) so it has ample time to re-read parts as required.
In fact many CD rippers have this facility and they rip a lot faster than the CD would take to listen to.
But most computer drives wind down to 1x when playing CDs, so surely the same caveats apply as to a CD player?In fact many CD rippers have this facility and they rip a lot faster than the CD would take to listen to.
I know some can (and do) play back at high speed - but most software disables that.
In fact, if this was the case - then wouldn't the ultimate CD player have a drive that reads at 40x (or at least anything faster than 1x)?
Globs said:
Many people are stuck in a world 3 decades old with thinking about CDs, a modern computer CD reader is simply light years ahead of anything available in the 1980s, in fact a £5 DVD reader would embarrass any 80s or 90s CD transport with speed and accuracy - comparing the two is almost ridiculous
I hope that wasn't a dig at me lol!!! 
Globs said:
Ripping a CD on a computer will get you as close to bit perfect as you are ever going to get from a CD, which most of the time is in fact bit perfect to the master. The days of paying over a tenner for a CD reader are long gone, raw/Flac CD files now sit on hard disks with total error correction to computing standards (i.e. _exact_). The real challenge is how to turn that data into decent analog, because the data is fairly sparse and there are challenges. The best way I have heard is to upsample and re-clock with a decent DSP/clock, and then feed into a decent DAC at that high (re)sampling rate.
CD transports do not and cannot alter the sound unless something is very wrong with the transport, and if there are any errors a PC reader will always be far better (assuming the reader is not faulty). Remember a PC CD drive can read in a DVD at high speed, a clunky old CD is a doddle. Paying money for an expensive CD transport is a total waste of cash, unless you want to donate to the factory that made it for other reasons.
Ripping should always result in a bit perfect copy, but despite re-reads etc. somehow that still doesn't seem to be the case?CD transports do not and cannot alter the sound unless something is very wrong with the transport, and if there are any errors a PC reader will always be far better (assuming the reader is not faulty). Remember a PC CD drive can read in a DVD at high speed, a clunky old CD is a doddle. Paying money for an expensive CD transport is a total waste of cash, unless you want to donate to the factory that made it for other reasons.
At least not if you believe what some software tells you anyway.
Edited by TonyRPH on Monday 7th January 17:50
TonyRPH said:
I hope that wasn't a dig at me lol!!!
No! more of a dig at myself as I compared them !This is an interesting read:
http://www.diyaudio.com/forums/pc-based/208898-rip...
CDs do degrade over time too - become unplayable and unrippable.
Globs said:
budgie smuggler said:
For example your wave form above could become this if the timing error was in the right place:
This however has never been observed in CD transports - diyaudio IIRC did a big test of bit accuracy and I don't recall it ever happening.I'm sure it's possible with ad hoc systems, but you'd need a jitter big enough to skip an entire bit/half bit which for even a cheap quartz crystal isn't going to happen.
The subtle timing of _when_ the value hits the DAC however will always be slightly out which is why a more stable clock will always improve sound.
The DAC is receiving a series of high and low voltages, either it reads the correct for that particular bit or not.
Gassing Station | Home Cinema & Hi-Fi | Top of Page | What's New | My Stuff