K&K Audio / Lundahl Transformers

Welcome! Need support, you got it. Or share youe ideas and experiences.

Return to K&K Audio / Lundahl Transformers

K&K Audio / Lundahl Transformers

Mo' Jitter (Thinking out loud)

68.167.248.222

[ Follow Ups ] [ Post Followup ] [ Thread: Display Email Next ] [ K&K Audio / Lundahl Transformers ]

With all this discussion of jitter, the thing that amazes me is nobody has had the wit to simply rip the Red Book CD into a GB of RAM (cheap these days), and read it out from there.
After all, much of the jitter is ultimately from the power-supply noise coming from the spindle motor, laser tracking servos, etc. etc. Dump the CD into RAM, shut off the drive, and a major noise source along with multiple tracking servos disappears. Everything gets real quiet with no moving parts any more.
Just think, the whole doggone 650-700MB can sit on a single SIMM circuit board, right next to a master clock for the DAC. The entire circuit board would be pretty small, with say a IEEE 1394/Firewire (or USB 2.0) interface between the external CD/DVD drive. Unfortunately, the wiseacres who design $15,000 audiophile players can't be bothered to learn about computer hardware ...

Follow Ups:

Don't even need that - Dan 14:40:23 06/26/04 (1)
just buffer it intelligently. Dan Lavry uses this technique (calls it CrystalLock I think), then regenerates the clock www.lavryengineering.com
Some papers there discuss it. Other interesting reading too. I haven't heard his stuff yet, but he's reputed to make the best A/D converters available.
- Re: Don't even need that - Cappy 16:00:52 06/26/04 (0)
  Speaking of intelligent buffer management, the following post reminded me of the cheap portable "jogger" players with a memory buffer:
  http://db.audioasylum.com/cgi/m.mpl?forum=general&n=47536&highlight=memory+buffer&r=&session=
  I also recall Alex Peychev mentioning to me he likes modifying Philips players because of the RAM buffer they have:
  http://db.audioasylum.com/cgi/m.mpl?forum=digital&n=81659&highlight=memory+buffer&session=
Re: Mo' Jitter (Thinking out loud) - Kuei Yang Wang 12:52:42 06/26/04 (0)
Konnichiwa,
> With all this discussion of jitter, the thing that amazes me is
> nobody has had the wit to simply rip the Red Book CD into a GB of
> RAM (cheap these days), and read it out from there.
Well, some early generation DVD Players (which sadly do not read recorded CD's) did it like that. The CD was read asyncronously, the data dumped into the 12MB on board SDRAM and read out fromn there. If you gave these suckers a good masterclock feed they where exceptionally competent transport and those with nice, top of the line pioneer legato Link DAC's also sounded very good as stand alone players.... I still have one around... ;-)
Sayonara
Re: Mo' Jitter (Thinking out loud) - Steve Eddy 10:51:13 06/26/04 (0)
After all, much of the jitter is ultimately from the power-supply noise coming from the spindle motor, laser tracking servos, etc. etc.
Since the only jitter that matters at the end of the day is the jitter in the DAC itself at the time of conversion, wouldn't you achieve the same results if you simply electrically isolate the DAC's power supply from the transport's power supply?
se
Re: Mo' Jitter (Thinking out loud) - Dave Davenport 11:22:53 06/25/04 (1)
Hi Lynn,
How long would it take to completely download a CD into RAM?
Dave
- Not long - Lynn Olson 00:32:22 06/26/04 (0)
  With 40X drives being the norm these days, a couple of minutes. With Mac OS X and iTunes, loading the whole thing doesn't take long at all. You hardly need a Unix operating system to transfer data off an external drive - some trivial off-the-shelf microcontroller programmed in BASIC would do fine.
Re: Mo' Jitter (Thinking out loud) - Cappy 10:26:50 06/25/04 (0)
Lynn,
I've been thinking along similar lines lately.
It does seem that the right combination of special purpose firmware and computer hardware could effect an excellent sounding solution.
Better than the current PC/USB or PC/firewire solutions with offboard DAC I don't know.
Once the CD is dumped into the RAM there needs to be intelligence to deal with the big RAM buffer. Per the discussion I had with Dave below - fast forward, skip, clear buffer, repeat, etc. This makes the problem somewhat trickier.
So now we are talking about a CPU and special purpose firmware. And a special purpose remote too. Doable, but it becomes a hardware/software collaboration effort. We will need a little operating system too.
Now the PC already exists, with available OSs and so on. It would be nice to leverage the PC and its plethora of hardware and software developments. How about an all solid state PC running some small OS? No hard drive noise, but we still have the power supply noise. And whatever else is on the motherboard noise. Still, it could sound good.
So this is why it hasn't been done (as far as I know) -- a special purpose, low noise computer is necessary.
Just thinking aloud here and rambling a bit.
Re: Mo' Jitter (Thinking out loud) - JamesD 01:06:37 06/25/04 (3)
Hi Lynn,
That is exactly what a lot of people are now doing. Ripping the CD to HDD and playing back via a huge RAM cache. Output to dac via USB2.
Its amazing how much difference there is through a good DAC. It just kills normal CD playback.
ciao
James
- Re: Mo' Jitter (Thinking out loud) - Lynn Olson 01:37:54 06/25/04 (2)
  Even a hard disk still generates a lot of power-supply noise from the motor and disk heads, and there's still a PLL inside the hard drive that syncs the disk rotation speed to the buffer.
  
  Granted, hard-disk rotation speeds and head positioners are much faster than CD/DVD drives, but there are still all the tedious issues of slaving a slow mechanical drive to a RAM buffer so there's no overflow - read data, send data over the bus to the buffer, buffer acknowledges readiness for more data, move disk head to the next track, start reading more data, etc.
  This creates lots of data-control logic going back and forth over control lines, spinning motors, head-seek data, and start-and-stop synchronization. All of this monkey motion is fine for reading and storing data, but it generates a lot of noise on the power-supply lines, and motor noise isn't easy to filter. Even a simple fan is not trivial to electrically filter and isolate from the rest of the power supply, much less a data drive, with its associated data and mechanical controllers.
  If the CD/DVD drive is spun down, power withdrawn, and the Firewire/USB data lines electrically isolated, noise in the RAM DAC then becomes a function of board layout, shielding in the right places, power-supply quality, and a heavily-filtered master clock with a minimum of phase noise. I suspect the sound quality would be a revelation with jitter down to really low levels, orders of magnitude better than any mechanical PLL-filtered and re-buffered data source. Ditch the mechanicals, go electronic!
  - Re: Mo' Jitter (Thinking out loud) - Pooge 06:21:34 06/25/04 (1)
    Sounds doable, but maybe not user friendly to wait for a download of whole disc into RAM before being able to listen. Also, maybe the present state of jitter reduction is reaching beyond audibility. One of the papers I posted earlier indicated the jitter level needed for 24bit resolution, and the current state of the art appears to meet this jitter level. It would be nice to hear a comparison, though, to see if fast access is worth giving up.
    - Re: Mo' Jitter (Thinking out loud) - akltam 07:51:00 06/25/04 (0)
      I am just trying brainstorm here (and I don't have much knowledge about how these device work).
      And maybe you can pre-download some of these disk into a CF, SD ...etc device so that you can minimise the load time. Since memory is so cheap these days, you can also have gigs of these ready so that this idea is more convenient.
      I really like this idea. I hope some one can come out with a product like this.
Re: Mo' Jitter (Thinking out loud) - James D. 01:06:26 06/25/04 (0)
Hi Lynn,
That is exactly what a lot of people are now doing. Ripping the CD to HDD and playing back via a huge RAM cache. Output to dac via USB2.
Its amazing how much difference there is through a good DAC. It just kills normal CD playback.
ciao
James
Part II (RAM DAC) - Lynn Olson 00:51:24 06/25/04 (0)
Jitter experts are aware that layout is critical to minimizing jitter, along with minimizing noise sources that impinge on power supplies. You want very compact layouts, with the master (non-PLL) clock adjacent to the DAC converter chip, so the cleanest possible clock signal (lowest phase noise) can be fed directly to the DAC.
By reading the data directly off a RAM chip that's physically close to the DAC logic, there's no need for any kind of PPL that's attempting to track a rotating disk, with its associated motor-control circuitry. No need to "slave" the CD player to the external DAC, either, since there isn't a CD player/transport at all, just an external CD/DVD computer peripheral with either Firewire or USB 2.0 interface to the DAC assembly. The logic control for the DAC assembly rips the data off the CD in a few minutes, and the external CD/DVD is powered down after the data is collected. You could even disconnect the Firewire/USB data lines with a relay if you wanted the last word in electrical isolation.
The physically small size of the RAM/data controller/DAC board should greatly simplify noise/jitter control for the 5V and 3.3V supply planes, as well as improving noise levels on ground returns. It's a lot easier to control noise in a board the size of an iPod compared to two large boxes with a S/PDIF data link between them.
If you want to play DAD or DVD-A, still not a major problem - even several GB of RAM isn't that expensive any more, so you just provide a few slots on the RAM DAC for future expansion.

This post is made possible by the generous support of people like you and our sponsors:

Parts Connexion

You can not post to an archived thread.