Meridian Unplugged - Powered by MrTechGuy.co.uk Forums Discussion Meridian - Systems Jitter review.

VK, my testing methodolgy, while not blind, reveals the minutest nuances of the kit.

What I do is rip samples of sound - eg a bell clanging, a piano being tinkled etc - samples which cover the gamut of requirements - bass, midrange, treble, transparency, presencs, dynamics etc - I then sit there and listen to a loop over and over and over until the sound is etched into my brain and THEN I switch live to the next product which is also running the same samples - I then listen to that loop for a while - and then switch back and forth. It's amazing how easy it becomes to distinguish differences with this method - there is no opportunity for faulty memories to cloud the tests. I take care to ensure things like volumes are identical and that all cables are identical. I'd do this double blind if I could get somebody to sit down and do the tests with me...

Mark

I agree nobody does double blind tests. I'll bet if someone popped in a SB3 in Crions DSP 7000 demo at that exact moment with the same music properly ripped it would be pretty impressive as well.

I test exactly the way Mark does. I find it's the only way to compare. For example I might focus on the position a violin is at hovering over the front left speaker. Then switch, is it stilll floating there. Yes or No. Now focus on the next instrument. Then focus on what ever reverb or echo you hear, is it in both? One nuance at time. I often try to set it up so it's a one button instant switch and often forget which was what (which is a good thing).

If I can't compare this way then it's useless. I have sworn up and down when not doing it this way that something was better by a mile. Then do this careful A/B only to find it's no better. It was my excitement hoping it would be better.

My point is simply that someone will listen to setup A at a shop. Then they'll go home and listen to setup B. It might sound better or worse. They then compound things by posting their comparative results. It annoys the hell out of me as any hint of objectivity is lost and it confuses anyone seeking objective comments.

And, dare I say it, someone who's spent a fortune on a high-end transport is unlikely to be happy that something far cheaper produces hifi of comparable or better quality. Subconsciously, this is going to affect the results.

Mark - your methodology (though not quite as elaborate) was how I came to the conclusion that my Squeezebox (1), Oppo 970 and Meridian 598 all sounded identical (to my ears) into my G68.

I simply ignore all opinions that aren't my own. It's the only way to go Listen to what people are saying, do research, get the kit for demo, test it, make a decision and then ignore the chatter.

Mark

I'd do this double blind if I could get somebody to sit down and do the tests with me...

Mark [/QB][/QUOTE]

Mark I am that man

Roger 01452 770 391

If I was in the market for a streaming transport to partner a Meridian processor/DAC, based on what I know now, for the best-bang-for-the-buck, I'd be inclined to go for a standard SB+ at �800 with a good set of cables and power conditioning.

If I had an analogue preamp I'd go for the upgraded SB+ at �1000 which is a step up again and with super-duper cables and power conditioning will give 800 and 808 owners something to think about.

If I wanted the last word in performance then I'm afraid I'm at odds with Mark on the best of the best, a modified Esoteric X-01 still rules the roost for me, from personal experience the Coherent modified X-01 with external word-clock is absolutely stunning! Jaw dropingly stunning! So stunning you just can't comprehend unless you've heard it!

Having said that I suspect even the Coherent X-01 is bettered by the Apl Hi Fi X-01.

Regards
Gary

Speaking of external word clocks. Wouldn't the best word clock have to come from the Meridian DAC itself?

I was wondering if the SPDIF outputs to DSP Speakers could be used to clock a source device.

This is how the EMU-0404 works. It sets up a SPDIF feedback loop. It can use an internal clock or an external clock coming from the device it's driving. This way it feeds it exactly at the rate it consumes.

You can have the worlds best independent clock but it's only as good as what ever is reading it. As far as peak to peak jitter the Meridian should more or less eliminate that with dejitter buffer. But the average clock drift can't be solved without feedback from what the Meridian wants.

If you think your clock is better than the Meridian then you might want to disable the dejitter FIFO. Because I think it then uses the source clock.

To VK's point, if you aren't doing some kind of direct comparison, you will never know. I've seen situations were perception of low level detail was enhanced simply because the "super resolution system" was in a room that was absolutely dead silent. Whereas, most of us live in environments where there is noise nearby (refrigerator, fan, traffic, tv in another room, etc.) our systems might be doing the same job and we'd never know it. When Absolute Sound reviewed that transport, don't you think it would be valuable (as a mandatory, objective function of the review) to double blind test it against other transports and media players?

PS: A tip to all who use media players - the HP Media Vault is a wonderful, low-cost solution.

I asked my mate who co edits 'Sound on Sound'
a few apposite questions on Jitter.

'What level of jitter, in picoseconds peak to peak,
is benchmark for digital recording?'

Wow! Now you're asking! This is a complicated subject indeed because
there are lots of different sources of jitter, and lots of different
ways of measuring jitter -- but jitter isn't always relevant to the
audio quality, and even where it is, there are different views on
what the audible results are.

Sample clock jitter is the most critical by far, obviously, but
interface jitter is the larger and more common... but also usually
less relevant.

You can't really measure sample clock jitter at an AES interface
(unless it is really, really bad!) because the band limiting applied
to most AES outputs (for EMC compliance) introduces jitter of its own
-- around 5ns typically.

The AES-3 interface is designed to accommodate something like 80ns
(peak) of interface jitter (at 48kHz sampling) which works out to
about 12ns as an rms value. Any half decent PLL clock system should
be able to get that down to below 120ps without too much trouble.

In any case, interface jitter isn't a problem within the digital
domain -- it only becomes relevant if, when decoding to analogue,
you're using the embedded interface clocks as the clock source. A
decent D-A will put a lot of effort into removing jitter in this
situation. An increasingly common solution -- adopted by Benchmark,
Cranesong and others -- is to use a sample rate converter to provide
complete isolation between source embedded clock and the D-A's very
accurate sample clock.

Sample clock jitter obviously plays a critical role in defining the
captured audio quality, and for a professional high quality 24 bit
system the received wisdom is that it should be below 100ps or so. It
Sadly seems that 300ps is not an unusual figure to find in reasonably
high quality domestic kit. And heaven knows what an iPod measures...
;-)

'What levels would you expect in a good recorder
and the finished product( ie CD.) ?'

Answered the first part above, but not sure what you mean by jitter
"in a finished product (ie CD)" I think the answer you need is both
'lots' and 'it shouldn't matter' The data stored on a CD is stored in
an analogue form defined by the pressed edges of the plastic, and
that will inherently be very 'jittery'. The data recovery and EFM
decoding stages are designed to remove this jitter effectively such
that the following circuitry can accurately transcode the digital
data. Jitter will only affect the analogue audio output if the D-A
stage can't deal with what's left effectively, as I've explained.

Hope that helps.


Roger

So if 100ps p-p is good for 24bit
should we be aiming for 133ps p-p for 16bit?

Another mate sends this

I'm a little late to this one, but a useful information document on the subject of jitter was published by AES Standards a few weeks ago. AES-12id, "AES Information Document for digital audio measurements - Jitter performance specifications" seeks to illuminate what jitter means and how the term is used differently in different circumstances.
It won't give you a hard answer in nanoseconds, but it will help to identify what sort of jitter the nanoseconds is describing!

Roger

"...reading it"? Not sure what you mean by that. My modified TacT/Squeezebox have nothing reading (sampling?) the clock - the clock feed is directly wired into the electronics replacing the internal clock sources so both devices tick tock identically. This is the whole point of a master clock. It doesn't matter which device is driving the clock as long as all devices are slaved to the same clock.

Mark

A link I often use to refresh myself on jitter is:

http://www.tnt-audio.com/clinica/jitter1_e.html

The previous section, 1.2, also contains some usuful info.

Mark