Meridian Unplugged - Powered by MrTechGuy.co.uk Forums Discussion Meridian - Systems +3dB "LevelPerfect" Tweak Performance Improvement Poll

No way it does that. The way SPDIF is implemented you can't.

Like Steve said a bunch of posts back. You can kind of think of SPDIF as ALWAYS 24bit. It's just a matter if these VERY low level bits are used or not.

Thinks of CD's as really 16bit with least significant 8bits as 0 (WHEN TRANSMITTED over SPDIF). If you just stretched that you push real good audible bits down into the abyss you can't hear.

518/568 MAY internally stretch 16 it out to 72 bits. But once it gets back to SPDIF and it has to be more or less on the same volume scale as what came in. Which means the most significant 16bits HAVE to stay as the most significant 16bits.


Too much sunshine, yes you're slowy catching on See Carl's post.

That makes sense! Also you have treble and bass controls...or do you with DSP speakers? Your right I should keep quiet because I didn't read through all the posts.

You have me wondering about analog and at what volume is it full scale and analog kicks in.

Yes you have treble and bass. I never use it though.

All valid points.

Just think of it as like a 518 and it's up to the user where they want to dial it in. I choose to go with a true 0dB (to match what a direct wire would do). Based on comparing that direct wired sounded more detailed than processor did with it's -3dB.

I think Analog may always be full scale to DAC. But like you said you might be dialing in a sweet spot in the preamp or the "Hybrid" implemenation.

Sorry if I sound cranky this thread is exhausting. And so is my wife every time I sneak on to post or run another test. I want folks to challenge this to see if there are holes in it.

Meridian resolution enhancement in the 518/618 was used more for down converting 24 bit or higher bit depth music to a lower bit depth by minimizing loss via the use of dither and noise shaping to keep as much of the original 24 bit resolution as possible. It was designed for recording high bit digital and converting to 16 bit CD. In fact if you read the 518 manual for running 16 bit to higher bit depth they suggest you not use excessive noise shaping. I think they say no higher than A. I use flat dither in my setup. It sounds the best with all material. Noise shaping sounds too sterile to me. The 518 comes in handy with 20 - 24 bit DAC's using it as a digital volume control for 16 bit music.

It does not increase musical dynamic range as what you would get with stretching it out between 16 bit to 24 bit. You need a real 24 bit recording for that. It basically sets the correct dither levels for feeding lower bit rates to higher bit DAC's so they will sound better. If the dither is set incorrectly the DAC basically switches off. You need dither noise to keep the 24 bit DAC operating below 16 bits. I like it more than my 568's dither for 16 bit music. Maybe because I like flat dither more than the noise shaping used in the 568. maybe the PLL dejittering makes it sound better. I just like it!

Ron

I'd guess that it isn't because the processor would control any level differences (between loudspeakers) in the DSP, not at the analogue stage.

I'm not sure how the analogue levels are attenuated in the processors, a hybrid could take the form of reducing the volume in the digital domain from 0dB to -5.5dB in 0.5dB steps, once -6dB is reached (a whole bit) the digital volume reverts back to 0dB and -6dB is applied at the analogue stage.

This would maintain good resolution (for the DACs), good accuracy of volume steps and easy to implement -6dB steps in the analogue stage (-6dB is 50%).

Pure conjecture though.

Steve.

George: I turned my brain off and stopped typing in here and started to listen carefully and I commend you on this tweak. It is real. Thanks for your most excellent efforts.

(If you at some point in future have the time and inclination to figure out how to force set the Bass settings via MCONFIG such that they would be adjustable over different DSP presets that would be fantastic. And if it is only possible to set Bass across all settings at one fixed value that would still help.)

I thought of a way to test this! Using a 518 connected to the 568 digital output while in analog output mode. Use the 518 to monitor digital full scale (it has a clipping indicator). I have an RCA test disk with full scale test tones. I can monitor the digital output as I change volume on the 568 and see how the digital output behaves. I can use the gain on the 518 to see what the offset is as I change volume. I'll just turn my amplifiers off and test away.

Ron

Keep in mind the Digital Outputs work different in Analog vs Digital and it may not accurately represent what is going to the DACs. You would hope it doesn't because it's pure digital attenuation on digital outputs when in analog mode.

But in digital mode the 518 might help monitor things.

I'm not sure how the 518 knows something is clipping unless it clips it, itself.

It could tell you that you are full scale though and might be what it does because I don't believe material is normally full scale.

There are clipping indicators on DSP speakers too, only the modern DSP speakers with Matrix displays have them.

9 folks with DSP speakers say Improved 0 say no improvement.

Improvement will vary depending on how far negative you were calibrated.

By the way the reason why I suspected it had to be +3.0dB exactly is a I'm totally amazed with all the golden ears on this forum that can hear Axis, Unsampling, Differences between G68/861, and Apodizing never noticed the lower you calibrate the worse it sounds (including me for that matter). I thought the "golden" value was out of reach. But it's always been available. All I did is find more range and that +3db in MSL really means 0dB of Gain through the Preamp.

Ahem...

I've ignored 75dB calibration point for years and kept all the DSPs as high as possible by calibrating to the level of the lowest measured speaker at 0dB adjustment. I raised the -3dB processor gain issue many years ago, advocating direct DSP connection for CD transports to circumvent the veil induced by going via the processor. Many have heard this difference and commented on it, but chose to live with it as a tradeoff for Trifield and MRC.

Cheers

George

Sorry, I stand corrected. There have been many folks saying direct wired is louder and questioned it (including myself). And I know there are few folks that calibrated as close to 0dB as possible. And a bunch at -3dB (or -4dB), deaf dopes like me.

But I don't recall folks saying.

-11dB worse than -10dB worse than -9dB > .. -1dB worse than 0dB

If there was, it was certainly a whisper.

The only threads I recall were Ron's saying more Negative was better (which may be true for Analog and I had analog at the time). But I Believe there are folks with DSP speakers that moved in the negative direction.

And then there is the crowd that says but it's 48bit, 72bit precision and 24bit word so you won't hear it, even if 0dB is in theory better than -3dB it was claimed to be infinitesimal. The further negative the worse it gets and far from small.

I really would have expected if this was throwing bits of resolution away (which it IS) there would have been a much louder, steady stink about it and raised consistently at Q&A's. I'm just as guilty, but I'm deaf because I can't hear what so many other folks hear.

I don't mean to point fingers, but I don't think a lot of folks are as objective when listening as they should be.

+1

Mentioned this back in 2002 or so

I'm also Wired for Direct 2 channel and fluctuate backwards and forwards between Direct Connect 2-ch and Trifield thru 861 depending on my preference with that particular music. I do have a pretty good centre that does add a lot in Trifield for some Vocal material, so my 2-ch listening is about 50 to 60% using Processor, the rest being Direct to Speakers

Meridian have obviously chosen the wrong compromise here, but they have always said to play around with the Calibration, Distances and various other settings until you achieve the sound you prefer. I guess some have taken this further down thru the years than others. It was likely around 1997 since I last went with Meridian's Default Settings, Distances or Calibration recommendations

I firmly believe that the focus now should be on determining WHO should bring this to Meridian's attention and HOW to get a Prompt Response regarding their Rationale. They deserve their day in Court before we pronounce them "guilty"