Red Hot Ruffled Chili Feathers

[andreasmaaan]

Here is a 1:15 long segment from both files. First, the CD version:

View attachment 18221

And the LP version:

View attachment 18224

Here are temporary links to the same segments.

https://tmpfiles.org/download/12810...ers-StadiumArcadiumCD-07-SlowCheetah.flac.txt

https://tmpfiles.org/download/12812...-StadiumArcadiumLP-01.07-SlowCheetah.flac.txt

Thanks

Downloading now. You've got very different window settings in those two screenshots (the CD goes up to about 24KHz and the LP up to about 100KHz).

 

[andreasmaaan]

Thanks

Downloading now. You've got very different window settings in those two screenshots (the CD goes up to about 24KHz and the LP up to about 100KHz).

Looking at them now. The LP definitely has much more dynamic range is the first observation. It's not necessarily better (that's subjective), but it's very different.

Going to look at the spectra now...

 

[andreasmaaan]

The LP seems to have 7 to 8 dB more dynamic range, and the peaks are about 1 - 2 dB lower in level.

Try lowering the volume of the CD by about 9.5 dB to make the RMS levels more similar and A/B'ing them. I've just done that and the apparent differences between the two masters are now far less obvious.

 

[andreasmaaan]

Other than that the spectra are very close to identical I'm sorry to say

I'm using the spectrum analyser in Ableton Live to determine this. I'm also listening as I do it. The spectral balance is basically the same. The CD master sounds more compressed and the vinyl master sounds like vinyl (some audible distortion and slightly rolled off top end). Other than that there is no real difference.

 

[Soniclife]

Have your seen...
https://www.audiosciencereview.com/...namic-range-database-caution.2216/#post-82834

I don't trust DR numbers much myself, they seem relevent for comparing different digital versions of the same album, but not for anything else. And are completely misleading across musical styles.

 

[hvbias]

MasVis is another option, it conveniently generates an animated gif.

Here is one I made of the original Led Zeppelin III CD vs the latest remaster

 

[andreasmaaan]

The vinyl spectrum:

And the CD spectrum:

I didn't go to excessive pains to get the levels exactly identical, but you can see that there's very little difference in the spectrum.

The main differences are the sub-20Hz rumble from the vinyl and the slightly rolled-off top couple of octaves relative to the CD.

There is a big difference in dynamic range though. The vinyl has about 7-8 dB more of it.

I'm estimating the dynamic range by getting the RMS levels to about the same levels and then looking at the peaks. The vinyl peaks about 7-8dB higher than the CD when the RMS levels are about the same.

 

[andreasmaaan]

I reckon the key here is to lower the level of the CD by about 9 dB when you're listening to them comparatively. That will get the RMS levels about the same. The CD may still sound better for various reasons, but in my opinion when you do this there isn't a massive difference in how they sound and it will come down to preference.

In my opinion, both have a lot of the flaws typical of (not necessarily inherent to) the two media. In the case of the vinyl, there's audible distortion (it sounds a bit papery to me) and the slightly rolled off top end. In the case of the CD, the mastering compression is aggressive and so things sound very thick and full but also a bit harsh.

 

[tomelex]

Nice work above guys, this is what this site is all about, thanks to OP and other posters, good information and technique.

 

[Don Hills]

Here is a 1:15 long segment from both files. First, the CD version:
...

And the LP version:
...

I see what you did wrong there.
(Edit) The tmpfiles appear to have expired, or I'd fix it for you.

 

[Jorj]

I didn't go to excessive pains to get the levels exactly identical, but you can see that there's very little difference in the spectrum.

The main differences are the sub-20Hz rumble from the vinyl and the slightly rolled-off top couple of octaves relative to the CD.

There is a big difference in dynamic range though. The vinyl has about 7-8 dB more of it.

I'm estimating the dynamic range by getting the RMS levels to about the same levels and then looking at the peaks. The vinyl peaks about 7-8dB higher than the CD when the RMS levels are about the same.

Nice work on those plots, @andreasmaaan, I appreciate your efforts. After peering at these plots for a while, it seems that the comparison is not just peak levels. The effective dynamic range between peaks at each frequency band are widely varied. Up in the higher registers, the CD has more range, and it varies a bit as you move down in frequency. In the 5K band, things are definitely squashed a bit, but I have seen much worse, and again, I think it sounds OK, definitely not the dumpster-fire that others portray it as.

It might be fun to find another track that is compressed to death and has a DR of 5 to see the difference. As @andreasmaaan said in post #24, despite having DR ratings of 12 for LP and 5 for CD, there is not much audible difference between these tracks. This is not always true, as I'm sure we can find other examples where the DR of 5 is just crushed to oblivion as compared to a well-engineered version with the same source material.

My takeaway remains fairly unchanged. Dynamic Range numbers, as applied to an entire track, are often a poor gauge of how a track was mastered, since methods for applying compression have actually become more sophisticated and really mess with the algorithms.

 

[andreasmaaan]

I agree partly @Jorj, but I don’t think CD master vs vinyl rip is a fair comparison, since the vinyl itself will degrade the quality so much.

To really decide if the CD master is overly compressed, I’d like to compare it to a digital version that hasn’t been so heavily compressed.

This doesn’t seem to be possible for this album, but can be done with a lot of heavily compressed remasters. In most of these cases IMO the heavy mastering compression does tend to be detrimental, although not to the extent often claimed.

 

[Jorj]

I agree partly @Jorj, but I don’t think CD master vs vinyl rip is a fair comparison, since the vinyl itself will degrade the quality so much.

To really decide if the CD master is overly compressed, I’d like to compare it to a digital version that hasn’t been so heavily compressed.

This doesn’t seem to be possible for this album, but can be done with a lot of heavily compressed remasters. In most of these cases IMO the heavy mastering compression does tend to be detrimental, although not to the extent often claimed.

Exactly, the vinyl is the only source I had for less-compressed material. If we could compare the recording to the final version, that would be interesting.

 

[andreasmaaan]

Exactly, the vinyl is the only source I had for less-compressed material. If we could compare the recording to the final version, that would be interesting.

Yeh

I’m pretty confident that after RMS levels were matched, for most recordings and styles of music, the less compressed version would be preferred.

 

[andreasmaaan]

After peering at these plots for a while, it seems that the comparison is not just peak levels. The effective dynamic range between peaks at each frequency band are widely varied.

Just reread this part of your post which I missed before.

This spectrum can’t tell us anything about dynamic range IMHO.

All it can tell us about is the spectral balance. I posted it only to demonstrate that the spectra in the OP were misleading (not intentionally so!).

 

[JJB70]

If you want to hear compression really kill an album then I recommend Twelve Songs by Neil Diamond. I really liked this album musically but find the CD almost unlistenable as it is compressed to the point of being physically uncomfortable to my ears.

 

[Jorj]

This spectrum can’t tell us anything about dynamic range IMHO.

So, when comparing the spectra, say at +\-500KHz from 5000KHz, if one version has higher highs and lower lows, would that not equate to more DR in that frequency band? Granted it's just a snapshot in time, but for that given instant, at that frequency, you can make an assessment. This is what I assumed the DR rating tools did, but I just don't see that they are getting it right. Maybe they are and I'm just an awful listener.

 

[andreasmaaan]

So, when comparing the spectra, say at +\-500KHz from 5000KHz, if one version has higher highs and lower lows, would that not equate to more DR in that frequency band? Granted it's just a snapshot in time, but for that given instant, at that frequency, you can make an assessment. This is what I assumed the DR rating tools did, but I just don't see that they are getting it right. Maybe they are and I'm just an awful listener.

Ah I see. No, this isn't a snapshot in time, it's averaged over about 10 seconds. And larger peaks and dips just indicate differences in level at different frequencies, but doesn't tell you anything about the dynamic range.

 

[Jorj]

Ah I see. No, this isn't a snapshot in time, it's averaged over about 10 seconds. And larger peaks and dips just indicate differences in level at different frequencies, but doesn't tell you anything about the dynamic range.

OK, did not know you'd done a 10 second average. That makes more sense then.

 

[dc655321]

@Jorj - there was a good thread back in June about compression in music production.
Well, at least I learned something... See here for a start (or just consume the whole thread?).