Jitter the killer of digital audio

Systems

TECHNOKID wrote: »

Who knows, it could have acted as an effective dielectric?

While I don't know the specifics of Seafires power supply upgrade,my own personal expierience with such mod's would lead me to believe the improvements were the result of installing a better power supply/voltage regulator scheme.

ben62670

steveinaz wrote: »

Ben, there is white-paper on Benchmark Medias website that shows how to covert a Toslink output to coaxial. http://www.benchmarkmedia.com/appnotes-d/cdjeep.html

Thanks Steve. I wll take a peek at my sony changer sometime soon.

GV#27 wrote: »

Toslink is optical.The Tos comes from it being a Toshiba development.

Yeah I meant optical to COAX;)

Systems

ben62670 wrote: »

Yeah I meant optical to COAX;)

The signal is the same right up to that little black box,it does the conversion to optical.Your 1792 board has both coax and toslink,or do you have a transport with only Tos?

ben62670

Yeah my Sony CD changer is Tos only. I am thinking about ditching it. To mount an optical on the DAC would most likely be about the same amount of work as converting the Sony to COAX. I'll poke around later. You get the email on the boards?

Flash21

Heh, I have to laugh...I should have known that digital cables would have as many points of contention as anything else in this crazy, um, hobby...

As Dave Barry said: "It's a fine line between hobby and mental illness."

ben62670

Flash21 wrote: »

Heh, I have to laugh...I should have known that digital cables would have as many points of contention as anything else in this crazy, um, hobby...

As Dave Barry said: "It's a fine line between hobby and mental illness."

Glad to entertain:) The difference between computer digital, and audio digital is that timing is an issue. With computer type digital the packets are just reassembled at the other end. With audio timing has an effect on SQ.
Enjoy.
Ben

sucks2beme

I was about to buy a better quality CDP, when I switched gears.
I went to a NOS DAC via USB, using pc with lossless FLAC and
ASIO4ALL to bypass Windows crap. The result was pretty good.
I'm watching for that killer DAC deal that I know is out there.

I then went back and installed FRED diodes in the power supplies
of two of my CDP's. I also changed out some 5332 chips in the AMC.
That resulted in some good gains in the sound.
The fast recovery diodes helped bass, while removing muting transistors
on the AMC output helped increase top end sound.
While jitter may be a problem, there's a lot of other places to get good
gains in sound. Most CDP's have a lot of other weak points on the power
supply and analog side to contend with. Of course, a good external
DAC would fix most of that. It's a never ending game. There's always
a weakest link.

ben62670

What did you change the 5332 chips to? I am a big fan of the OPA134/2134 chips. For the price of $2 each it is a silly cheap mod to do on gear. Very dynamic, and great sound stage, but the mids are a little rough. I can't even be sure if that is attributed to other gear to be fair.
Ben

Systems

Flash21 wrote: »

..I should have known that digital cables would have as many points of contention as anything else in this crazy, um, hobby...

Atleast it didn't end up in a flame war.

As Dave Barry said: "It's a fine line between hobby and mental illness."

It may be a fine line but I stepped over it a long time ago.:D

seafire

sucks2beme wrote: »

I was about to buy a better quality CDP, when I switched gears.
I went to a NOS DAC via USB, using pc with lossless FLAC and
ASIO4ALL to bypass Windows crap. The result was pretty good.
I'm watching for that killer DAC deal that I know is out there.

I then went back and installed FRED diodes in the power supplies
of two of my CDP's. I also changed out some 5332 chips in the AMC.
That resulted in some good gains in the sound.
The fast recovery diodes helped bass, while removing muting transistors
on the AMC output helped increase top end sound.
While jitter may be a problem, there's a lot of other places to get good
gains in sound. Most CDP's have a lot of other weak points on the power
supply and analog side to contend with. Of course, a good external
DAC would fix most of that. It's a never ending game. There's always
a weakest link.

Glad to see someone else has had good results with ASIO4ALL Very good software and same as used by many studios;)

heiney9

I thought everybody knew about ASIO4ALL when running a PC rig. Bypassing all the Windows processing is mandatory to do it properly. I guess I assumed wrongly.

steveinaz

LuSh wrote: »

I'm not sure about Benchmark's claim.

I noticed the same when trying to use my Panasonic BluRay, and my Denon 2910 as transports vs my CEC. The CEC won easily. It was just far smoother sounding with a deeper/wider soundstange--treble was less stringent.

Fongolio

I have read that some extremely high end music servers use solid state storage (no hard drive but instead more like ram memory). The theory is the less moving parts the less chance of clock induced jitter.

Flash21

I'm still trying to wrap my head around the fundamental idea that variations in timing of a binary code can affect the sound...this only seems to make sense if the timing of the analog music notes is not contained in the code itself, but is instead based on the timing of the code's arrival at the D/A converter...but keep in mind that I am very computer challenged, I am probably missing something fundamental you computer geeks know.

Systems

When trying to recreate the original analog waveform with 1's and 0's timing errors result in the right data but at the wrong time.

megasat16

Thus creating phase shift and makes peaks instead of dibs?

Well, I guess one can reduce the clock jitter by multi-path DAC and ADC at both ends and transmit at higher synchronous clock rate and do some kind of hardware error checking. I mean we can absolutely transmit and receive data over single media at different frequency and encode and decode the data back to the original form at both ends. We can transit the data at least 4 times higher than sampling frequency and transmit it over the same media at 4 different channels using different frequencies (multiplex). We can also sneak in two extra channels for data correction (sort of like CRC) and master clock rate for synchronization and we can even piggyback the master clock signal for ultimate accuracy. We can receive all 4 different signal paths along with two extra channel at the other end and de-multiplex and compare against master clock rate. We can even introduce the data correct scheme in the transmitted data and checksum over the 4 different signal paths to make sure data integrity. The receiving end can do all the calculations to correct clock jitter and any momentary jitter caused by interference over the cable. I guess things like this cost a few bucks for computer (and data communication) hardware but no one even want to sell things like this for under a grand for audio. Well, it's already been implemented for all data communication equipments and digital transmission standards are set for Industry. Unfortunately for the Audio community, the standards are either rare or disputable due to everyone doing different things.

What the hell I am talking about?

Just use the Analog out from the CDP and forget all this crap I said above.

Flash21

GV#27 wrote: »

When trying to recreate the original analog waveform with 1's and 0's timing errors result in the right data but at the wrong time.

That's pretty much what I am saying...apparently the time axis of the analog waveform is contained not in the code itself, but in when the code arrives at the D/A converter.

Which is probably an efficient way to do it, saving bandwidth. Nowdays though, bandwidth isn't the problem it used to be.

seafire

megasat16 wrote: »

Thus creating phase shift and makes peaks instead of dibs?

Well, I guess one can reduce the clock jitter by multi-path DAC and ADC at both ends and transmit at higher synchronous clock rate and do some kind of hardware error checking. I mean we can absolutely transmit and receive data over single media at different frequency and encode and decode the data back to the original form at both ends. We can transit the data at least 4 times higher than sampling frequency and transmit it over the same media at 4 different channels using different frequencies (multiplex). We can also sneak in two extra channels for data correction (sort of like CRC) and master clock rate for synchronization and we can even piggyback the master clock signal for ultimate accuracy. We can receive all 4 different signal paths along with two extra channel at the other end and de-multiplex and compare against master clock rate. We can even introduce the data correct scheme in the transmitted data and checksum over the 4 different signal paths to make sure data integrity. The receiving end can do all the calculations to correct clock jitter and any momentary jitter caused by interference over the cable. I guess things like this cost a few bucks for computer (and data communication) hardware but no one even want to sell things like this for under a grand for audio. Well, it's already been implemented for all data communication equipments and digital transmission standards are set for Industry. Unfortunately for the Audio community, the standards are either rare or disputable due to everyone doing different things.

What the hell I am talking about?

Just use the Analog out from the CDP and forget all this crap I said above.

LOL that my friends about sums it up in a nutshell:D
That is why I am tempted with the German design as you can take any cd player and even a pc whith coaxial out and have near external dac quality.Problem is it's not that cheap and I think by now there must be some good external Dacs floating around with built in jitter correction:rolleyes:

megasat16

While Jitter is definitely a pain in the S and often over looked in the Audio Industry, this is a lot more of an issue for data comm equipments where switching 1 to 0 and 0 to 1 makes the whole data a complete garbage.

Anyhow, the problem with the Audio Industry is that there is very little set standards and protocols so a well design and good CDP with a good DAC section could cost 10K or 100 bucks. Likewise, any poorly designed CDP with so so DAC will sell for 100 bucks because they know they can sell it for that much and puts 90% profit in their pocket. It's much like what Cable to choose.

A well design CDP or any DVD player with a good DAC should take care of 99.99% of the synchronization problem. The other 0.01% is considered not enough to make a difference in sound quality at the other end. Of coz, one can always be extremely cautious about this and wanted to have something better. It's completely acceptable to me for someone to pursue the highest level of perfection one can obtain.

As always, a manufacturer can look everything under electron microscope and exaggerate any tiny correction they could make about it and overly built to accommodate such feature and charge 10 grands to those who don't mind paying for it. It's perfectly fine as long as there is the market and there is the manufacturer for those kinds of gear and I am happy they can afford these and not me.

heiney9

Here's a more down to earth explanation of jitter. I don't necessarily agree/disagree with what's said. But it is laid out in nice format and easy to understand conceptually. A good read if you are at all interested in understanding this phenomenon but don't have an EE degree

http://www.audiocraftersguild.com/AandE/cd-jitter.pdf

H9

P.s. I know it's dated but the first couple pages are good explanations

Flash21

Thanks, reading that link now...

What this means is that any reasonable transport will deliver an identical data stream from any reasonable disc, regardless of how much green ink is around.

megasat16

Thanks Brock! A Good Read For Any Audio Lover!

Systems

megasat16 wrote: »

Well, I guess one can reduce the clock jitter by multi-path ....

There does exist a format called I2S that allows for the transmission of the data and clock signals separately,not combined as in the SPDIF.Unfortunately only a few manufacturers adopted it so it never gained widespread appeal.

W WALDECKER

A dedicated clock power supply with an accurate clock is the way to go in a stand alone CDP

NJPOLKER

After a HARD week hunting and drinking, well not much hunting, I had the jitters and after a few drinks they were gone. A simple fix as far as I was concerned.