Digital connections: Coaxial VS. Toslink Is one superior?

[MichaelD]

I'm no EE; that's for obvious. I've been doing some thinking (yes, it hurt) about the connection b/t my DVD player and my receiver. Currently, it's connected with a Monster Cable Digital Coaxial cable. The cable was not cheap...I believe in you get what you pay for, particularly with audio cabling.

I just hooked up my new CD player with a Toslink cable and was thinking that maybe a Toslink b/t the DVD and receiver would be better. Would it?

If you say yes, please explain why (back it up w/tech info, please) Thanks guys. I think I may buy a Star Wars-type flick on the way home...I'm in the mood for some space explosions and stuff.

 

[Superdoopercooper]

THEY ARE SAME. If your DSP gets a 1, its a 1. If it gets a 0, its a 0. That's the magic of digital.... 0's and 1's aren't corrupted by noise... then when they get to where they need to, they are converted into the analog signal. that is why Cell Phones went digital also. I should say that the ONLY real reason that Toslink may be better is that it will isolate the ground from you DVD and your Reciever... where as the RCA/Coax liklely wont (unless they designed the equipment well).

So, use the RCA cable for the Co-ax. If you buy anything more than a $5 cable (Radio Shack has some gold ended, very good RCA's for abotu 5-7 bucks), you ripped yourself off. Esp. if it's only one or two feet long. A longer run may need better cable, but still unlikely. think about it... It's not a very high frequency signal. IF it was, they'd have you use RF cables.

 

[Gunbuster]

It's digital, as long as the bit gets to the other side it should work fine, you could use a coat hanger to connect the Coaxial jacks


On the counter point insert ritch bastard audiophile mumbojumbo about $24,000 4 foot speaker cable: XXXX

 

[kami]

Oh man...this is a really touchy subject.

Most people just say...binary code is binary code, and it doesn't matter how it's sent. I somewhat believe in this, but I also believe in "you get what you pay for"....as for what's better, you'd need to have incredible ears and a badass setup to distinguish between the two, but most audiophiles I've seen prefer the sound from a coax cable. But when it comes down to it, the benefit of coax over toslink is that coax is a LOT more durable... it's extremely easy to either break a toslink cable or have it not connected properly.

I suggest you read through this, this, and this on the home theater forum.

 

[Superdoopercooper]

KAMI...

you are right in regards to cable that will be carrying ANALOG signals... such as speaker wire, etc. There are arguments that say the quality of cable has a distinguishable difference... and I'll agree with that somewhat (I did a Master's Thesis on just this subject of Analog RCA cabling). But, for Digital... A 1 is a 1, and a 0 is a 0.... If the bits are corrupted... then you are screwed either way... with TOS or with Coax. You can't make a digital signal that is represented as 1010110 sound better on Tos than Coax... if the signal from both is sent to the same DSP.

 

[MichaelD]

That is exactly what I thought, gents. Thanks much. My CD player only has the Optical Digital out, no coax digital out. So that's why I'm using the fiber.

The reason I spent the extra dough on a "good" digital coax is b/c I have a LOT of wires running behind my entertainment center (who doesn't, right?) and I didn't want to risk picking up any AC hum or interference from the speaker wires...after all, the Coax cable DOES have metal in it. To it's credit, the Monster Cable Coax is a great cable. To my ears, my DVD sound is first-rate. Thanks and have a nice day.

 

[mithrandir2001]

Unfortunately, I cannot explain the technical reasons, but allegedly coaxial is superior to toslink. I believe the reason has to do with jitter, which in high amounts makes digital sound bad. It's not simply 0's and 1's like data transferring from a hard disk. Digital audio data is extremely time-sensitive so the transfer process has to be flawless for optimal sound.

During the mid-1990s, audiophiles were obsessed with outboard DACs and jitter-reduction devices. I used to have a Sony 5-disc changer connected to an Audio Alchemy DTI*Plus which was then connected to an Audio Alchemy DDE v1.1 DAC. Recently, the move has been back to stand-alone CD players. The reason has to do with the digital connections between the transport and DAC. When you transfer digital audio over Toslink, it is streamed, but you really want to send two types of data simultaneously: the digital music and the digital timing data. With a coax/toslink, you have to break apart these two, send it over the line and then reassemble it. If you keep everything inside the CD player, you can have an internal clock maintain the timing...it is superior method. Some CD players have an output jack called I2S, which keeps the timing signal intact so there is no jitter issues; however, it is not a common format.

 

[Superdoopercooper]

Mith...

Ahhh... you are right. I think jitter is a spec that is very important. Maybe the optical drivers have some jitter problems due to light "on" and light "off" times varying with temperature and stuff?? Have no experinece with that... but a good point that you bring up!!!!

 

[mithrandir2001]

<< But, for Digital... A 1 is a 1, and a 0 is a 0.... If the bits are corrupted... then you are screwed either way... with TOS or with Coax. You can't make a digital signal that is represented as 1010110 sound better on Tos than Coax... if the signal from both is sent to the same DSP. >>


You would think this is true, but for digital audio it is not. Your argument works for data transfer inside a PC: as long as the bits are in the right order, your system doesn't care when it gets those bytes to the CPU. It will just idle until it gets it...the software won't break and the OS won't crash if it has to wait.

But digital audio requires perfect timing because you are playing back something in real-time. If the timing is off, the output will lose fidelity. With a stand-alone CD player, an internal clock keeps track of the timing of the data transfer and processing. But when you transfer digital across wire or fiber, that timing data is integrated with the 0's and 1's from the digital audio. Because the DAC/receiver is physically seperate from the CD transport, it cannot share the hardware clock...the timing is instead thrown into the bitstream. The DAC/receiver has to extract the timing data from the receiving bitstream and resynchronize the clock. In theory, all of this should work flawlessly. However, when you leave the design paper and enter the real-world of manufacturing tolerances, varying power supplies, and anything else that reduces fidelity from perfection, you'll realize that digital transfer across coax/toslink is not optimal.

It should be noted, however, that most average listeners really can't tell the difference between coax and toslink. It's generally something that becomes more readily identifiable to trained listeners in controlled listening environments. But the point here is to think that not all digital connections are equal and it does matters how there are deployed.

 

[Superdoopercooper]

I've done some more thinking, and was wondering:

Doesn't the chip on the receiving end really determine how the data is going to be decoded. IF your DSP has good anti-jitter circuitry.... or if the PCM bitstream out of the DVD player doesn't have a dependency on jitter.. then it shouldn't matter. As long as their is some sort of sync pulse/code in there somewhere, and the proper number of bits between sync, the receiving chip should be fine.

I think the jitter you are referring to may be somewhat different than what we're dealing with in the new DVD PCM datastreams. I honestly don't know though... it'd beinteresting to see the Tos/Coax output specs...

 

[MichaelD]

Well, these are some of the specs off the Sony site. The CD Player is a CDP-CX335.


Frequency Response2Hz - 20kHztolerance: +-0.5 dB
Harmonic Distortion0.0045%
Signal-to-Noise Ratio107 dB
Dynamic Range93 dB
Channel Separation100 dB
Line Output Voltage2 V
Line Output Impedance50 kW
Power Consumption13 W



See it here.


The instruction manual lists very little for the laser/optical part. It just says something like "Laser type-constant Wavelentgh 600nm" and that's it. No specs for jitter or whatever.

 

[mithrandir2001]

<< Doesn't the chip on the receiving end really determine how the data is going to be decoded. IF your DSP has good anti-jitter circuitry.... or if the PCM bitstream out of the DVD player doesn't have a dependency on jitter.. then it shouldn't matter. As long as their is some sort of sync pulse/code in there somewhere, and the proper number of bits between sync, the receiving chip should be fine. >>


I recall reading that syncs can be off even as low as 100 nanoseconds (1/10,000,000 of a second) and still have an perceivable, negative effect on audio output. If DVD uses a 96KHz sampling rate, the syncs must be even more accurate, so it seems to me DVD is more susceptible to jitter and tighter control of the signal is required. But I'm more or less guessing at this point.

Of course, if you are using your receiver to process the signal with additional DSP, you are better off feeding the receiver a digital signal via coax/toslink than feeding it multi-channel analog, which needs to be reconverted in digital, processed and then converted back to analog.

 

[Superdoopercooper]

<< Of course, if you are using your receiver to process the signal with additional DSP, you are better off feeding the receiver a digital signal via coax/toslink than feeding it multi-channel analog, which needs to be reconverted in digital, processed and then converted back to analog. >>



You sure...??? I thought the idea of multichannel analog outputs on DVD players was so you could feed it DIRECTLY to a 5 channel amplifier (or preamp).... you get the straight up signal how it was meant to be... without any silly DSP fields/effects added to it.

It wouldn't make any sense to redigitize a signal from a DVD player that has been already converted to Analog...