CD player replacement-any good DVD?

[Eugene Dvoskin]

I can put in an enthusiastic vote for the Pioneer DV-656A(aka DV-45A). It's a HUGELY reliable laser & mech, and this generation of Pioneers, from 656A up to the DV-79avi(which has wonderful OPA2134A output opamps stock), are shockingly great sounding cd players even in stock form, and can be upgraded fairly easily to be even better.
unfortunately, from DV-46AV, DV600, etc. forward, Pioneer went to a MUCH LESS reliable laser maker. So, until I see that change, anyone asking me for a cheap but terrific cd player gets the strongest recommendation for the DV-45A & it's brothers. There is not a new dvd or even blu-ray player from any maker on the market that I could recommend, reliability-wise(which is as important to me as sound quality).

How old are those, 7-10 years? It seems that it was a burst of attention to DVD audio 10 years ago. And now, nobody is really interested.

 

[stephensank]

The DV656/45/59/79 are only about 5 years old at most. They are 'universal' players, meaning everything but blu-ray. SACD.DVD-A,DAD,CD, etc. And, frankly, age doesn't matter much on them, since they are the most reliable dvd players I've yet seen. Been using a DV-45A literally 24/7 in my shop for at least 2 years now, and not a single problem.

 

[SoNic_real_one]

What ADI DSP has to do with clock jitter? Clock jitter is a hardware related item. To have a good jitter, it is necessary to start with a good crystal itself, then it requires good and low noise front end (input buffer/comparator), and proper design of direct or indirect (PLL) clock architecture. ADI DSP is used for signal processing.

Are you real? When the signal is processed in ANY DSP with its OWN clock (therefore async to the input), the samples are reclocked with that clock. All the incoming jitter is GONE. Remains only the jitter of the XTAL of that DSP. Sure, some FIR magic and dual port memory are involved... but as jitter goes that's the fact.

 

[gaetan8888]

Hello

Most low and mid prices DVD players use noisy switch mode power supply, it's pollute all the circuits and worsen the jitter.

I have two mid prices DVD, one Toshiba and one JVC, my Adcom cd player sound way better than those DVD players. OPPO and Pioneer Elite DVD player are sounding good but cost much more. But I still prefer my Adcom cd player.

Bye

Gaetan

 

[Guido Tent]

What ADI DSP has to do with clock jitter? Clock jitter is a hardware related item. To have a good jitter, it is necessary to start with a good crystal itself, then it requires good and low noise front end (input buffer/comparator), and proper design of direct or indirect (PLL) clock architecture. ADI DSP is used for signal processing. It may use it's clock to do generate reference clock, but why it is better than just hardware based clock solution? DSP needs clock for it's internal operation. Quality of these clocks is not really very high.

Usually the DSP clock is the same as used to convert the data. In that case the jitter should be low.

But you are right, for signal processing, one does not need low jitter as such.

best

 

[Eugene Dvoskin]

Are you real? When the signal is processed in ANY DSP with its OWN clock (therefore async to the input), the samples are reclocked with that clock. All the incoming jitter is GONE. Remains only the jitter of the XTAL of that DSP. Sure, some FIR magic and dual port memory are involved... but as jitter goes that's the fact.

So, what is the magic behind DSP clock and why it has better jitter than clock, generated anywhere else? And you have to sync your processing to the stream of data. Inside, DSP processing is async, all it has to do is to prepare the data before next sampling period. I can ask you the same question - Are you real? Or to be more exact, have you ever tried to design anything with DSP

 

[Guido Tent]

So, what is the magic behind DSP clock and why it has better jitter than clock, generated anywhere else?

You are right: No magic. Wrong implementation from Denon

The conversion clock should always sit close to the DAC, and be fed back to the DSP.

In addition, an oscillator on board with a million transistors switching on a DSP will never perform good in terms of jitter.......

happy new year !

 

[SoNic_real_one]

So, what is the magic behind DSP clock and why it has better jitter than clock, generated anywhere else?

The way it's working on any async DSP is simple -in all of them there is a buffer dual port or FIFO memory run by DMA, written with the incoming jittery signal and read with the stable XTAL clock. Much more stable than the jitter induced by an optical drive - CD or DVD.

And after that you have your FIR/upsampling/bass management DSP routines.

More: http://www.analog.com/static/imported-files/processor_manuals/ADSP-21160_HRM_Rev4.0.pdf

Shadow Write FIFO
Because the DSP’s internal memory must operate at high speeds, writes to the memory do not go directly into the internal memory. Instead, writes go to a two-deep FIFO called the shadow write FIFO.
When an internal memory write cycle occurs, the DSP loads the data in the FIFO from the previous write into memory, and the new data goes into the FIFO. This operation is transparent, because any reads of the last two locations written are intercepted and routed to the FIFO.
Because the ADSP-2191’s shadow write FIFO automatically pushes the write to internal memory as soon as the write does not compete with a read, this FIFO’s operation is completely transparent to programs, except in software reset/restart situations.

 

[Guido Tent]

The way it's working on any async DSP is simple -in all of them there is a buffer dual port or FIFO memory run by DMA, written with the incoming jittery signal and read with the stable XTAL clock. Much more stable than the jitter induced by an optical drive - CD or DVD.

but not good enough for decent audio playback.....(the groundbounce of such chips is in the order of several 100mV - this voltage is in series with the clock signal since the clock relates to external ground.....)

 

[Eugene Dvoskin]

The way it's working on any async DSP is simple -in all of them there is a buffer dual port or FIFO memory run by DMA, written with the incoming jittery signal and read with the stable XTAL clock. Much more stable than the jitter induced by an optical drive - CD or DVD.

And after that you have your FIR/upsampling/bass management DSP routines.

More: http://www.analog.com/static/imported-files/processor_manuals/ADSP-21160_HRM_Rev4.0.pdf

Sorry, I understand now what you meant. But the quality of the clock on DSP is comparable to the quality of the clock on SM8707HV, until some additional efforts are taken, starting with the quality of quarts crystal itself and the way, how generator is build. It may be possible to improve quality of both, by creating external generator. The other thing is FIFO. It is not ideal. You already have some jitter in the clock. FIFO is synced to this clock and while syncing, some jitter is added. It seems to be digital, but in reality it is all analog when it comes to jitter. The way how clock inputs are made should make big difference. It takes some time for the clock signal to switch it's state. During this time, any noise at the clock or at digital input will result in additional jitter.

In any event, they are on the order of magnitudes worse, compared to what Guido was describing. I'm not a big fun of audio extremism, it's just interesting how good is good enough. Guido's numbers surprised me a lot. Just wonder, how 1ps jitter was measured.

 

[Andersonix]

OPPO and Pioneer Elite DVD player are sounding good but cost much more.

My Oppo 980H didn't sound good until I swapped out the output op amp from TI NE5532 to LM4562. Nowadays one might find a used 980H for ~$120 and up.

 

[SoNic_real_one]

All of the players are plagued by cheap OpAmps. One more that others...

 

[Guido Tent]

Guido's numbers surprised me a lot. Just wonder, how 1ps jitter was measured.

We (Tentlabs, Grimmaudio) have build our own jitter analyser of which the noise floor is currently at about 10fs.

best

 

[Eugene Dvoskin]

We (Tentlabs, Grimmaudio) have build our own jitter analyser of which the noise floor is currently at about 10fs.

best

That's amazing! Can't imagine how this is possible

 

[Guido Tent]

That's amazing! Can't imagine how this is possible

well, it aint easy but it all comes down to low noise design......

 

[SoNic_real_one]

I am sorry I am not closer to you Guido, to test the "dirty" FIFO in Analog Devices DSP from Denon DVD players against a "clean" clock in a stock CD player.
I am sure that ANY optical pickup assy will have a higher jitter than one resulted from a Xtal driven DSP. Mechanical, I think the jitter is inherent in those optical drives... no matter how the clock is generated, there is the drive motor PLL.
Sure, you will say that your clock makes wonders regardless, but I cannot see the advantage unless is located in some kind of reclocking mechanism (like the ones in the async DSP's).
Below is the VCO from the NAD CD 512. How is the superclock improving the PLL located in the front end of the player? The low frequency mechanical jitter will propagate thru the system all the way to the DAC, unless is some kind of buffer memory in between. An async DSP provide that buffer.

An externally hosted image should be here but it was not working when we last tested it.

 

[Extreme_Boky]

I am sure that ANY optical pickup assy will have a higher jitter than one resulted from a Xtal driven DSP. Mechanical, I think the jitter is inherent in those optical drives... no matter how the clock is generated, there is the drive motor PLL.

Definitely. Current-starved spindle motor and CD mechanism encapsulated in high density plastics is solution… tackling DAC issues only is solving ½ of a problem.

Boky

 

[Eugene Dvoskin]

Definitely. Current-starved spindle motor and CD mechanism encapsulated in high density plastics is solution… tackling DAC issues only is solving ½ of a problem.

Boky

I misunderstand something here. You getting data and then recovering the clock.
Until you having unrecoverable error, it all depends on how you are recovering this clock. It has no direct relationship to the quality of the transport. It is how your PLL is designed. Your error may be added if you are using a complicated clock generating structure, just because you may have a couple of PLL's connected in series.

 

[SoNic_real_one]

The "recovered" clock has jitter. Directly product of the transport spindle motor. You have the correct data but it comes at the wrong time. Mecanically, the motor will have fluctuating RPM's and the PLL tries to adjust that constantly. That produces delays and advances in the recovered clock. And that low frequency jitter becomes noise in the DAC.
That's why I am advocating the DSP solution used by several manufacturers (Denon is an example based on what I have) that eliminates the incoming jitter due to use of FIFO buffer. Sure, adds the DSP jitter, but I think that is several orders under the optical pick-up one.

LE: Here is another aproach, withount use of the DSP memory, just plain 128MB of RAM:
http://esoteric.teac.com/dacs/d-70/

 

[nimo_jon]

Hi

I am new to this forum due to my quest of a good TDA1541A DAC. I am a believer that happiness can be bought by spending money very carefully.

I bought a Pioneer DV 320K in the far east (Made in Thailand). Its a thin & flimsy player. But it plays great music and video (no HDMI) in stock form. Less than US$80 in major stalls. I also bought the Pioneer DV 220K which has HDMI up sampling for video.

For the DV 320K, it uses the typical application of the Media teck MT1389EE DVD system chip. It is a VVLSI that does all the DVD control and outputs the SPDIF at pin 225. At pins 229 and 228. it needs a reference crystal of 27.000Mhz. There are many application notes and also repair manual on MT1389EE. Many cheap DVD players in the far east uses this chip.

Pin 225 goes thru series resistor and cap filter/blocking network to output to the 75ohms digital out and optical out. If you want it to play CD this is a what is needed.

I have removed the on board 27Mhz crystal and the caps/resistor oscillation circuit. I installed a dedicated small power transformer, LM317T (100ohms/300ohms divider with 10uf bypass) output at 5C dc to power a dedicated 27.000Mhz TCXO. The TCXO output goes thru two 74HC14 Schmitt trigger gates for square wave shapping; signal feeds thru an (optional) 33ohms resistor and into pin 229 directly. This new clock makes the player plays fine. Both video and audio significantly improved.

Over damping of the player can make the sound down and muffled. I found that as long as you put some weight (such as an amplifier) over the player it works like magic.

To use the DV 320k as a dedicated transport, I took the output of pin 225, goes thru two gates from another dedicated 74HC14, bypassed the resister networks, and feed straight directly into the Toshiba TORX optical out. I prefer the optical link because of the inherent limitations of using pulsed Tx. The sound from optical sounds cleaner, to my personal preference in sonic flavour.

The result is a nice sounding DVD transport that sounds great feeding a WM8805/TDA1541A NOS DAC. It is detailed and yet musical playing even the old Philips/EMI/Decca 1980 classical CDs. I had the opportunities to compare the transport to some very expensive players. I would say that the modified Pioneer is not the best, but not far behind either. For less than US$120 all in all, it does what I want. I can focus on spending time listening to music. Cheers.