How to compare CD vs. vinyl on my system

I know this will probably open a can of worms, but am wondering if any sound engineers and/or veteran audiophiles (preferably those like both CDs AND vinyl) can weigh in on how I might go about comparing the CD vs. LP version of a given album given my equipment (see my profile page). I know the first step is making sure that the master is the same, but what other factors would I have to consider and attempt to control given the inherent difference between digital and analog playback? Basically I'd like to compare my digital playback source (Oppo BDP-93) with my vinyl rig (Pro-Ject Xpression III with Sumiko Oyster cart) to see which "performs better" for a given recording, but haven't a clue how to go about doing it accurately.

 

I think the first step would be trying to get similar sound levels .

 

Yep. Levels. That's it.

 

Hey, I like your user name. I could never do that very fast.

To compare sound between the two, I use rock records I purchased in the mid '70's and original mastered cd's made in the mid-late 80's. I do have to lower the volume of the cd's. Even some of these cd's may be different, such as the horns in Alice Cooper-Muscle Of Love- the song' Man With The Golden Gun' ( WPCP-3492 ) are louder in the cd than the lp.

I find with my old players is that the lp's bass is a tiny bit fuller and deeper, the highs are sometimes slightly extended and the mids are always more laid back, less harsh.

Let us know what your hear between the two. Have fun.

 

agreed the levels. The does the Oppo have a digital volume like my 105D? Outside of that quality needle drops might help.

 

Levels, same mastering....and the rest depends on your ears alone. I almost always like vinyl better.

 

Level match at low volume using the Oppo digital volume to match the vinyl's fixed volume (and you have to do this PER title as vinyl varies more than digital)....and to the quietest section of the first couple tunes on the side---THEN turn it up (meaning the output/stereo volume-leaving the Oppo matched) and compare. You should be able to sit with a remote and switch between than at that point--using the volume to turn them both up and down by the same amount.

The low volume is because you need to match it on the loudest bits--snare and vocal levels. If you match loudly--the digital, having more low end, will always seem vocal shy.

If during the climax of the song, you think one is way louder--it's actually the more dynamic of the two--do NOT adjust the Oppo's digital volume--that should be set to match the quietest parts.

Good question, BTW--I've been meaning to start a thread asking how people are gain matching for comparing--you can NOT do it with the output volume knob.

 

Also equally important proper alignment and loading for the cartridge. The thing about vinyl you can keep tweaking forever. It does keep getting better but at some point the laws of diminishing returns really start hitting you hard.

I was amazed how much better things like vocal clarity sound from precisely setting azimuth and VTA

 

It's not enough to have the same mastering. You need a real state of the art quality recording.

 

Good luck with level matching. If there isn't a proper test signal with the recordings this is always going to be dodgy.

 

Just out of interest why?

 

Without a steady simple signal or series of signals at different frequencies you will have a lot of misdirection. With the vinyl you will have surface noise adding to the level if you use a quiet passage to match levels. If you use a loud passage you will have added energy due to distortion from the vinyl. You will also have all kinds of cross talk and frequency response non linearities with the vinyl that could give dips and peaks depending on where you measure. A lot can go wrong in level matching when there are inherent mismatches in level between two sources.

 

Do you mean comparing by ear or comparing by measurement?

If the latter, one useful tool is Voxengon CurveEQ. It enables you to the compare the EQ of one source with the EQ of another source.

So, for example, you could rip a CD to your PC and, using Voxengo CurveEQ, you could get its EQ curve. This would then be your reference for the next two steps.

You could then play back the same CD through your CD player and record the output from the analogue, using your PC. You could then get an EQ curve for this recording using Voxengo CurveEQ. Still using Voxengo Curve EQ, you could then compare this EQ curve with the "reference" EQ curve you obtained in step 1. Any difference between the two would show that your CD player was deviating from a totally flat frequency response, but probably, it will be dead straight.

Next, let's suppose you have an LP version of the same album and you're absolutely certain that the mastering is identical. You can now play the LP on your record player and record the output to your PC. Using Voxengo CurveEQ you can then compare the EQ of your LP recording with the EQ of your "reference" rip. Again, any difference will show the extent to which your vinyl rig (mainly the cartridge) is deviating from a totally flat frequency response. Normally, you'd expect it to be wildly inaccurate.

Using these sorts of techniques you can measure the frequency response or EQ curve of almost anything in your audio setup. I use Voxengo CurveEQ all the time to compare the EQs of different masterings. Once you've found out what the frequency response of your cartridge is, you can easily and accurately compare LPs with CDs by simply subtracting your cartridge's frequency response from the LP's EQ curve.

 

This will certainly help to a large degree in determining frequency response differences but I'm not sure it gets you any further in trying to level match. How do you assign a level value to frequency response differences? We also still have the non frequency response distortions that affect the measured levels of vinyl.

 

Same mastering is a tough nut to crack. A month or two ago I did this comparison on my system and used the recent release of the Grateful Dead's 1972 show at Veneta, Oregon. The show is a 16 track recording that was mixed down using high-rez digital within the last couple of years, so had good, recent technology here. With portions of the show released as a video called "Sunshine Daydream", the full concert is available on LP and CD. The vinyl is a 4 record set and the CDs came with the video. I verified with Jeffrey Norman, who mixed and mastered the recording, that the same master was used for both (song order and flips were different with the two formats but that doesn't matter). I'm not sure I own another recording where I'm absolutely positive the same master was used for both the CD and the LP.

I didn't worry about getting too meticulous in level matching although I certainly tried to get the two formats at the same volume. Since I wasn't doing a straight A/B but instead over the course of an afternoon listened to a song on one format and then on the other, any small level differences would end up being meaningless. I listened to half a dozen songs this way, changing up the order. Sometimes I started with the LP version first, other times I played the CD initially. Any volume advantage would vary back and forth between the two formats. In other words, one format wouldn't always be louder. In any case, they were always pretty much the same volume.

This comparison took maybe two hours to do. The result? I was shocked how close the two sounded in my system (CD player: Sony SCD-XA5400ES /// Turntable: VPI Classic 2 platter/Lyra Delos cartridge/Parasound JC-3 phono stage). The CD and the LP sounded virtually identical. If you had one of the two playing and I walked into the room, I'm pretty sure I couldn't tell you which one it was. My two turntables sound much more different than the did the CD versus the LP.

I didn't expect this. In the past, with different hardware I've heard more of a difference between a LP and a CD. Not this day. I could easily see a casual listener hearing absolutely no difference at all.

 

As been said the mastering is everything. I recently posted a needle drop with Blood Sweat and tears (original US pressing) compared to a 24/96 capture of the same track from DSD and those that responded liked the vinyl better. That was even with increased noise dually noted.

With sample like that normalize and levels are matched. One could argue that both could sound better being heard directly.

 

You can never achieve a gain matched comparsion of two different sources, which is what vinyl and CD ALWAYS are....so, while yes it will be "dodgy"---it has to be done, and it has to be done on the source--NOT with the output knob. If you ignore this, you will think vinyl sounds magically better even when it's cut FROM the CD--because they're pre gain staged it for you.

 

OK how do you do it? oh wait, you said it can't be done. But it is necessary. hmmmm.

 

Which medium 'feels' emotionally better to your soul.

 

What you quoted is a response to another post---it can't be done on the level of accuracy that I quoted where someone was talking about test tones. Literally impossible because it's a moving target--test tones are about calibrating voltage outputs at specific levels and frequencies--no two different sources can be matched on that level--even though the outputs themselves can be calibrated, that's irrelevant when what you're feeding into them are two different things.

That doesn't mean you can't functionally gain match.

I do it all day more days than not in my work. I'm pretty OCD about it, actually....maybe my doing it regularly with meters to check at various points has trained me to do something Joe Listener would struggle to do. Admittedly, that might be a weak point of my insistance--maybe Joe L doesn't know what a few DB down sounds like. Maybe they aren't equipped to make the call that the snare or the vocal is the loudest element thus should be used as the aural match point...

 

Thanks for all the detailed replies. This would be more for fun and out of curiosity than anything. I know I can always just compare based on my ears alone but I think being a bit more scientific about it would be a good learning experience. What equipment would I need to measure levels? Output to the above-mentioned EQ software?

 

Sorry, hadn't read this before my previous post, you've laid it out pretty clearly. Thanks!

 

But if you use the loudest element as a match point between vinyl and CD you skew the match for everything below it. And matching by ear and by measurement will yield very different match points. So what is an actual match? I completely agree that levels are critical but level matching is very problematic and can accidentally tip the balance when picking a preference.

 

True. But, those aren't just "loudest" bits--they're midrange, heavy on the 1khz and solidly in the center channel (you know-north of 1970 at least)---basically the elements that will survive on vinyl best no matter how skewed the RIAA stage or cart's mid/side decoding is.....

The super low volume in important. Our ears hear lots of highs and lows as "louder"--turning the volume WAY down for the match is important. I realize I didn't make a big deal of that part, but it is. I have a "dim" switch on my console that cuts (by 20db?)--I always engage it to match two things.

 

I guess my point is that whatever methodology one uses to "level match" it will yield different levels from other methodologies since there is no true level match. It only takes a subtle difference in levels to sway a preference so the chosen methodology will likely impact the preference. It's problem with no perfect solution. If you are level matching the loud points by ear the fact that vinyl will be perceptually louder will skew the mid level and low level spls to make the CD louder in comparison. That will yield a false sense of greater detail and clarity. OTOH if you match the low level sound the vinyl will be perceived as more dynamic when the music gets loud. And since these level differences will actually be fairly subtle they will not be heard as distinct level differences. Without the source material offering some sort of test signals that we can use to measure level we're kind of screwed when it comes to level matching