Attenuating CD Player Output for Optimal Playback

[btswire]

On this forum the issue of recording levels has come up. and the general consensus has been to keep the levels down to prevent analog gear from being pushed beyond it's optimal range. The posts from Massive Mastering, in particular, summed this up quite nicely.

Analog gear generally tends to perform best around 0 dbvu. If most CDs released these days are around -8 dbfs RMS, and a CD player is calibrated so that -18 dbfs = 0 dbvu, this means that the receiver is being hit with a + 10 dbvu signal, with peaks at +18 dbvu.

Is it likely that the sound quality would be better (more open and less squashed) by attenuating the output of the CD player? If so, what would be the best way to achieve this (perhaps something very simple that could be picked up at Radio Shack?)

[brew]

What is the CD player output connected to?

Here's an assumption: professional analog outboard or console. This gear usually will clip after +22 dBu, some after +26 dBu. Thus your CD, with a reference level of -18 dBFS = 0 VU = +4 dBu, will peak at +22 dBu. So it's skirting the threshold of linearity, or where THD increases substantially, 1% by common practice. Some gear still sounds decent in this zone and some falls flat on its ass. Just different design at work, doesn't mean one is good or bad in my opinion--headroom cannot go on forever.

In this case you don't need to attenuate the CD player, or maybe just a little. But if you were to sum another signal with the CD player, you'd be into the distortion zone of your +22 dBu gear. But this is still fine because consoles and gear have pads, input trims, and faders, so it's easy to avoid--you wouldn't need to add anything to the output of your CD.

If your gear is not happy at +22 dBu or you wish to reduce the level like you describe, here are two ways you could do it: A pro CD player may have output trim adjustment. You could tweak this, however, many do not. The purest way would be to take the digital out of your CD into a pro DAC, which will always have output trim.

This Rane note is good on I/O levels and distortion: http://www.rane.com/note145.html

[@?,*???&?]

This isn't exactly the case...

There are articles out there suggesting that it's actually the D-to-A converter that is distorting with higher disc levels.

Mastering engineers will generally try and keep the peak level below 0 dbFS for this very reason.

[btswire]

This isn't exactly the case...

There are articles out there suggesting that it's actually the D-to-A converter that is distorting with higher disc levels.

Mastering engineers will generally try and keep the peak level below 0 dbFS for this very reason.

Correct, the DA converter can distort at 0 dbfs, which is why the output ceiling should be set to around -.2 dbfs during mastering. This, however, is a separate issue from distortion that occurs from overdriving an input in the analog realm.

Brew, thanks for all the good info. To answer your question, I'm using a Nakamichi CD player with a late 70s Marantz 2385. My preference would be to try to find a very simple attenuater, which could be fixed at say -15 db.

Also, I was unaware of the difference between VU and dBu. Could you describe this?

[Waltz Mastering]

Here is a analog to digital conversion chart that diplays how dBVU relates to dBu relates to dBFS
http://www.sendspace.com/pro/dl/xcq6ms

AD and DA converters come pre-calibrated and some are able to be further calibrated by the user. You need to take into account where a converter is calibrated when referencing the chart. There is no standard for calibration so you might see anywhere from -22 to -14 dBFS = +4 dBu

[btswire]

Thanks Tom! I just did a quick google search and found this:

http://www.enjoythemusic.com/magazine/e ... thwell.htm

In this the author states "Many of us audiophiles complain about the harshness of CDs. This is true, but my experience is, after using these attenuators as well as using a Dynaco CDV-1 CD player which has an analogue volume control, that some or perhaps most of this irritating sound comes from sending too high a voltage from the CD player to the pre-amplifier. "

I'm skeptical about the need to shell out a lot of money for attenuators, but I do think something on the cheap might be worth a try.

[brew]

Mastering engineers will generally try and keep the peak level below 0 dbFS for this very reason.

Correct, the DA converter can distort at 0 dbfs, which is why the output ceiling should be set to around -.2 dbfs during mastering. This, however, is a separate issue from distortion that occurs from overdriving an input in the analog realm.

Yes, a separate issue, but the -0.2 ceiling you suggest is not adequate.

-2 or -3 dBFS is where you need to be to avoid overs at the DAC. TC Electronic has done much research into this: http://www.tcelectronic.com/media/niels ... bfs_le.pdf

A barrel attenuator like that shouldn't cost too much, maybe $20 to $40. You wouldn't get more than 6 dB attenuation from the method I described, so for 15 dB or more you'd want to use one of those: http://www.fullcompass.com/category/Att ... -Pads.html

[klangtone]

Is that A.Hollis as in Andrew Hollis from Univ.of Miami?!
If so, hey man!

I wanna say, yes! brew is definitely on to something about the -2 or -3 thing. I'm not necessarily suggesting you start mastering your recordings to only peak at -3dB, but... maybe not a terrible idea.

What would be nice on a CD player is not an analog volume control, but a digital one to turn down the recording before it hits the converter inside. If you go down 3dBFS it might help with the distortion at the top end of some converters.

I don't think anyone is suggesting that all converters behave in such a way. I think it's a you get what you pay for kind of thing. And guess what is most likely in your average consumer electronics gear? The cheap stuff.

I work as an Electrical Engineer at Qualcomm in San Diego. We make chips for cell phones and stuff. So I help design the parts of the chip that interface with the mics and speakers on the phone (thus, including the audio converters). So, I can tell you with some experience that things can get gnarly at the last dB or so. Some of it is the DAC itself and some of it is the little amplifier after the DAC that drives the load. See what happens is, the IC designers are up against a rock and a hard place. They have a requirement that their DAC output so many mW into such and such a load. But then they also are stuck with a certain voltage rail that is either as high as their silicon process allows (for reliability reasons), or as high as the architects of the entire system allotted for them for whatever reason.
So often they are pushing that hairy edge. And you have to understand too, the Silicon process is not very rigid. If you design a resistor in Silicon, it innately has about a +/-20% variation unless you do laser trimming which almost no one does because it's too expensive. So that means, a certain percentage of the parts shift to the "sucky" part of the bell curve. Obviously, every part gets tested and ones that are really bad are thrown out, but in order to keep the yields decent (and even a 1% yield loss is looked at with scorn) the specs for the part are basically manipulated to allow for such variation in process. So, a converter will be speced, for instance, for THD+N, but in the conditions, the input signal is -1dBFS, not 0dBFS. 0dBFS may be unspeced, or the spec may be severely relaxed.

Now, I'm not just dirting my employer, this is common practice among many companies that make similar products. Although I'm most familiar with converters designed for cell phone products, most of the companies are the same as those that supply for your typical consumer electronic or computer equipment (heck, even Pro-Audio gear...). And since low power, portable audio is quite the rage these days (pods, pads, etc.) chances are the converter of the cell phone is of similar quality to that of your mp3 player. I'm actually astounded at the level of quality we are able to achieve in our latest products at Qualcomm (especially stuff that's still in design or test). Somehow the designers are finding ways to eek out more and more performance even as they shrink the silicon area, reduce the power, and work with lower voltage rails.

But this issue at the top end of the converter will probably always be with us. And at the end of the day, it pretty much comes down to cost and marketing bs... like most things in the Corporate world... :-p

(oh, and yes, Engineers are marketed to in a very similar way as the end consumer. And it seems to work! Hard to believe, but true. Meaning, datasheets for components (like ADCs and DACs) are speced in a way that highlights them in the best possible light. So companies are always trying to out do each other in the way they spec their parts just so they can put in bold on the first page of the datasheet some amazing technical number that the part only really meets under one particular condition. And some engineers are either stupid, lazy, or are pressured by their marketing to select the part with best numbers, even if it's mostly bs.)

Sorry for the rant!!!

Roy

[Nick Sevilla]

Thanks Tom! I just did a quick google search and found this:

http://www.enjoythemusic.com/magazine/e ... thwell.htm

In this the author states "Many of us audiophiles complain about the harshness of CDs. This is true, but my experience is, after using these attenuators as well as using a Dynaco CDV-1 CD player which has an analogue volume control, that some or perhaps most of this irritating sound comes from sending too high a voltage from the CD player to the pre-amplifier. "

I'm skeptical about the need to shell out a lot of money for attenuators, but I do think something on the cheap might be worth a try.

You could build your own... it is basically a resistor network. Do use 1 or 2% tolerance resistors rated for the correct output voltage of the CD player.

Cheers

[btswire]

So I never got around to buying (or making) those attenuators.... then the idea occurred to me, why not dust off my old JVC tape deck, put it in record enable mode, and run my cd player through it? Obviously not the purest way to do it, but I figured it was worth a shot.

So I did, and set the input level so that the peaks are registering at about 3 dbu. I put in some really hot CDs, and suddenly they weren't so piercing!

I don't know how much psychology is coming into play, but I'm really finding that this makes a huge difference.

For those of you who still have tape decks around (or anything thing else for that matter that will allow you to attenuate a line level signal) I'd be curious to get your comments on you find the sound to have changed...