Forums dot lukpac.org

lukpac wrote:

That's Eric Sardinas.

So if it it's not down to EQ, the sonic differences between the LP and CD may be due to:

1) Euphonic distortion introduced during the LP cutting process

or

2) The DA converter used for converting the master to digital

or

3) The sound signature of your playback cartridge, turtable, phono stage, etc.

or

4) The sound signature of your playback DA converter

or (most likely)

5) A confluence of 1-4

I have a Sony JA555ES MiniDisc deck. This dual power supply, 35 pound monster was the most elaborate, no compromises MD deck that Sony ever manufactured. It really excels at analog recordings, as Sony spared no expense at making the cleanest signal path they could manage, and the DA converters are second to none.

I've made many LP dubs on this deck, and the recordings faithfully capture and reproduce every nuance of the vinyl. So much so that I cannot tell the difference in an A/B test. (Yes, I've done that -- literally blindfolded, while a friend worked the machines.)

So, given a well-designed analog stage, if a lossy format like MD (ATRAC Type-R) can capture it, I don't see why uncompressed PCM wouldn't be able to do it as well.

lukpac wrote:Speaking of which...does anyone know a good way to match EQ from one version to another. Is there some type of spectral analysis tool that would come in handy for this?

Sometimes I try to find a passage of music that has just vocal and compare those. It seems that we all have a pretty good ear for comparing vocal tonalities (hearing voices all our lives, I guess).

I've never tried this, but I wonder if it would work. Match up a short interval of vocal only, invert the phase of one of them, sum to mono, and do a spectral analysis of what is left. That might pinpoint the differences better.

Dob wrote:Sometimes I try to find a passage of music that has just vocal and compare those. It seems that we all have a pretty good ear for comparing vocal tonalities (hearing voices all our lives, I guess).

I've actually been hearing some of the biggest differences in the drums/cymbals.

I don't trust any of you any farther than I can throw you, but here are 3 samples:

Sparrow (LP)
Sparrow (CD)
Sparrow (CD w/ EQ)

No, I didn't bother to match the levels. And yes, I do level match when comparing.

And here's the EQ (so far, anyway):

Crummy Old Label Avatar wrote:So if it it's not down to EQ, the sonic differences between the LP and CD may be due to...

So, given a well-designed analog stage, if a lossy format like MD (ATRAC Type-R) can capture it, I don't see why uncompressed PCM wouldn't be able to do it as well.

I performed a few bias-controlled A/B tests some time ago with some analog-loving audiophiles who utterly failed in picking the "warm" analog source from the digital clone. So I can attest to the fact that there is nothing inherent to the vinyl medium that is not captured by 4416 digital. A couple of these audiophiles got really angry and claimed my playback chain was not resolving properly, so I brought my DAW over to their place and performed the same tests again and the results were the same.

There is a thread here on ProSoundWeb that discusses it. Note the responses from Paul Gold and Bob Ohlson. The latter was mastering engineer for Motown during the 60s and onwards. This should lay certain audiophool claims to rest.

lukpac wrote:I've actually been hearing some of the biggest differences in the drums/cymbals.

Well, you could work with those, then. But I would start with correcting the midrange/low treble (2-3k) and then working your way out from there. Even if you think it sounds OK, spend a few minutes manipulating those frequencies just to see what it sounds like...I've found that problems in this frequency area are very common.

If there is a boost at 2-3k on the CD and you boost above that frequency to give some of that vinyl presence and shimmer to the percussion, it could end up sounding too harsh.

Did you listen to the clips?

I can't do it until I get home -- will try to remember.

Listening on headphones, I'm not sure why you *want* to match the LP. It sounds muffled to me --not 'sparkling'. I can see you've tried to replicate that with those cuts in in 10/12 kHz. Comparing the instrumetnal bits, it seems the LP still has fatter bass or mid than your EQ...maybe that's what you're missing in the drums.

krabapple wrote:Listening on headphones, I'm not sure why you *want* to match the LP. It sounds muffled to me --not 'sparkling'.

I'd say it's only "muffled" next to the CD. Before I heard the CD, I even thought the LP was *slightly* bright in places. Even if it's what the tape sounds like, the CD is way too bright IMO.

You're close. I hear a bit more bass on the CD w/EQ and more top end on the LP. The upper range of the vocals are a bit muted on the CD/EQ version. The LP is slightly more open in that area.

Yeah, my buddy Danny gave me something else to try. That's next...

That's a really disturbing avatar.

OK, now that I've finished dinking around with Raw Power, I gave the samples a listen.

You think the CD is way too bright? The cymbal sounds a bit overemphasized, but otherwise I think the tonality is not bad. The vocals have a trace of treble harshness but sound natural. My approach would be to start with a slight shelf cut (maybe a db to start) on everything above 8k or so. On your EQ I think you're taking out too much at 10k...your EQed version loses too many ambience cues.

Once you get a good tonality from the shelf cut, I would try boosting some frequencies in an effort to get some of that LP "air." The trick is not to undo the cut that tamed the cymbals...you may end up with a curve (for example) that boosts at 8k, cuts at 10k, boosts slightly at 12k, and has a slight shelf cut from there on up. In any event, from what I've heard, I don't think you're going to need anything more than a couple dbs boost/cut. The CD doesn't sound that out of whack to me. I'm also not sure about your cut at 2.6k...you might want to restore that while you're working on the upper treble, then add it back in later to see how it sounds.

You could try cutting some of the midbass frequencies, but I've found it easier to concentrate on boosing the treble instead, especially since you know the treble has problems and have to work with it anyway.

Luke,

dont' know if you're stillworking on 'sparrow' or not, but I finally got around to running Audition's frequency profiler on the mp3s from cd and lp that you posted -- the resulting difference stats are in an Excel worksheet of 3 columns (frequency, left, right) of >900 rows . Rather than post that, I've placed the worksheet

here

and included a graph to give you a rough visual idea of how the files differ -- you cna of course regraph the data yourself to zoom in on the <1kHz region, for example. There is one trace each for left and right channel, showing the differnce for each in dB between the CD and the LP (i.e., at frequencies where the trace is below the X axis, the CD is less loud than the LP; where above, it's louder than LP). As you see it's a fairly sinusoidal 'difference' curve...and the differences are substantial at almost all frequencies. To match the LP, it looks like reductions are called for from about 1200 Hz on up, while the region below that needs various amounts of boost.

I didn't normalize either file -- this is a comparison of the files as posted.

I haven't tried any automatic EQ-mapping software, so the only way I know so far to use this info, is to manually set up an EQ profile in any virtual equalizer you have handy, inputting the differences frequency by frequency (the max is 30 bands of EQ for Audition, while I've given you 900), and apply it to the CD file. Which I haven't tried either.