While sorting through my huge RAID array, I came upon this old video, so I must share it with you now. This is a scope shot of the SPDIF output of a CD723.
Zippyshare.com - spdif_jitter.avi
Now, I'd like you to look at the video, please forget the terrible rising edge in the foreground, and focus on the other traces crossing in the background, which are other parts of the SPDIF signal.
Play the video. Notice how the traces wiggle and move around, as if the signal were a little bit drunk. Now, this is digital silence, none of that quirky audio sample data, just zeros, so it should be pretty steady on screen. But the jitter is visible to the naked eye, and you don't even have to squint.
CHALLENGE: Find what caused this.
Winner gets bragging rights.
Zippyshare.com - spdif_jitter.avi
Now, I'd like you to look at the video, please forget the terrible rising edge in the foreground, and focus on the other traces crossing in the background, which are other parts of the SPDIF signal.
Play the video. Notice how the traces wiggle and move around, as if the signal were a little bit drunk. Now, this is digital silence, none of that quirky audio sample data, just zeros, so it should be pretty steady on screen. But the jitter is visible to the naked eye, and you don't even have to squint.
CHALLENGE: Find what caused this.
Winner gets bragging rights.
Exactly.
Vibrations in the room cause the spinning disc to move and have to be compensated by the actuators which draw current. Voltage on the power supply reflected this. And the PSU design was bad enough that every supply inside the box (digital, analog, etc) had vibration-dependent ripple on it. Enough to affect output jitter (and not subtly).
Vibrations in the room cause the spinning disc to move and have to be compensated by the actuators which draw current. Voltage on the power supply reflected this. And the PSU design was bad enough that every supply inside the box (digital, analog, etc) had vibration-dependent ripple on it. Enough to affect output jitter (and not subtly).
I suspect any dampening would help.
But then what if the user turns the volume up: after all, CDs are for playing music so the room would be full of all kinds of vibrations anyway.
That's why I rip my CDs silently on a precision laptop drive (capable of reading recordable discs and DVDs at multiple speeds) and play them back later from vibration immune solid state devices.
The CD was a neat(ish) idea 33 years ago but today IMO even the best £multi thousand machine is simply beaten by a £5 computer drive and later playback due to simple physics. Just-In-Time 1x CD data extraction is neither warranted, wise or required these days - or in fact for the past decade
My personal opinion differs. Sound quality value aside, CDs are still a good idea. Part of that good idea is there is no constant upgrading or changing of format/tech in order to have the latest spec. A CD from the 80s works with a cd player of today and vice-versa. If listening to music is the goal, then it succeeds perfectly and has done for 33 years.
CD players from the 90s were already reading CD's at mulitple speeds and playing back from vibration immune solid state storage, albeit stored temporarily. Granted, you can do that cheaply with DIY these days if you want to spend a long time making it solidly built and nicely designed.
But I just wondered if the audiophile mucking about with seemingly unneccesary accesories like CD mats did actually have a basis in engineering terms to compensate for bad equipment design. Normally engineers etc dismiss their use as pointless but at the same time only appear to think in terms of perfectly engineered cd players etc.
CD players from the 90s were already reading CD's at mulitple speeds and playing back from vibration immune solid state storage, albeit stored temporarily. Granted, you can do that cheaply with DIY these days if you want to spend a long time making it solidly built and nicely designed.
But I just wondered if the audiophile mucking about with seemingly unneccesary accesories like CD mats did actually have a basis in engineering terms to compensate for bad equipment design. Normally engineers etc dismiss their use as pointless but at the same time only appear to think in terms of perfectly engineered cd players etc.
Probably worse - You don't want the disc to be steady; you want it to be steady relative to the CDP. And since the vibration is transferred from the floor to the CDP then to the CD... Best is a massless disc that is held on very tightly by the spindle and is very rigid, which I doubt the mat helps.
Not to mention the extra current needed by the motor to spin a heavier load on that weak power supply.
Yeah, me too, there's also stuff like vibration dampeners, differences between CDR brands while the bits stored are exactly the same, etc.
When output jitter is measured, the unit usually sits on a table with no vibration. I'm pretty sure using a shaker table would yield interesting results.
But it all comes down to perception: "This DAC is so revealing, it shows clear sonic differences between transports!" --- engineer translation: "This piece of junk faithfully reproduces source jitter and doesn't do any goddamn thing to filter it out like using a decent PLL, ASRC, or an ESS chip..."
This is the case of my Onkyo AV amp, CS8416 followed by an implementation of PCM1792 which I dearly hope the Burr-Brown engineers never saw, because they would probably want to curl up in foetal position and sob at the sight of this horror.
Or "everyone knows TOSLINK sounds like crap, you must use a $429 coax with precision machined RCA connectors lubricated with snake oil, nothing cheaper than Nordost Silver will do to connect to the impedance-mismatched, 10c mass produced gold-plated-over-nickel-over-steel RCA connector at the back of the player which leads to an impedance-mismatched trace on a single-layer board and/or an unshielded wiring harness with 1 GND conductor and the rest of the wires carrying various unrelated stuff (CD723 again)...
My favorite idiotic tweak is to put a cotton ball on the clock oscillator. This reduces air currents and stabilizes the oscillator's temperature. I haven't heard a difference in sound quality, but it costs nothing and it does yield a measurable improvement in low frequency phase noise.
Ah yes, sounds true. or relative to the whole mech assembly at least. I guess Stan Curtis had this in mind with his Cambridge Audio CD1 (~1987) with the spring mounted mech with heavy lead mass loading. It gave better noise measurements, I'm guessing at the raw data side from the laser.
I dunno, I used one from the bathroom and didn't look at the package, so I'm afraid the vital knowledge of the brand name has been lost forever...
That said, it really works, a kleenex squeezed into a ball and taped on the PCB works just as well, and jitter skirts on the FFT do get narrower. Stabilizing the temperature does reduce LF phase noise in an oscillator. It's basically a ghetto OCXO without the heater. All the benefits and none of the hassle, since frequency precision to the ppm doesn't matter.
Audible or not? I'd dare say the price to performance ratio sure beats a grounding box...
That said, it really works, a kleenex squeezed into a ball and taped on the PCB works just as well, and jitter skirts on the FFT do get narrower. Stabilizing the temperature does reduce LF phase noise in an oscillator. It's basically a ghetto OCXO without the heater. All the benefits and none of the hassle, since frequency precision to the ppm doesn't matter.
Audible or not? I'd dare say the price to performance ratio sure beats a grounding box...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.