-

I don't think there's anything you can do... there are filters that will restore peaks apparently after you bring the gain down, but i'm not sure of specifics, and they do make mistakes (so you'd have to go over each track carefully). I also don't know if they actually sound better in practice.

There's a plugin called iZotope Vinyl for Winamp that reduces stereo separation, adds "analogue" distortion and ticks and pops to get that authentic vinyl sound.

There is nothing you can do to save a compressed recording. Either live with it or change music style and use eBay/amazon to get the old non-remastered versions of old music.

i, too, would recommend finding old, "non-remastered" recordings when available.

you can apply lossy gain to the tracks, but that only solves part of the problem - you will probably enjoy listening to it more.


later

"every single CD put out these days has tons of clipping and is usually over-compressed"

This over compression and clipping problem is mostly only the stuff from the big companies for which they hope to get radio DJ play. A great deal of music from many other labels is never destroyed that way.

take Jean Michel Jarre for example and compare his old records to the ones made after 1997... I don't believe JMJ is a "pop" thing to a regular radio station play like RHCP is ? These compressed records are a nightmare

To a certain extent, I respectfully disagree. For sure, the insane clipression we see on some CDs is limited to major label releases, but there is still too much compression on most CD releases. A few CDs that I've recently bought (which are certainly not major label releases and can hardly be accused of expecting any radio play) are: Rick Wakeman's "Retro" (album gain: -7.70dB); "Refugee" (album gain -8.61dB); and "Hatwise Choice" by Hatfield and the North (album gain -4.78dB). Even older CDs from way back (late 80's) typically exhibit album gains between -2dB and -5dB. Contrast this to my vinyl transfers, which after normalisation to -0.1dB typically have album gains between -0.5dB and +0.5dB.

But those low gains are not necessarily a sign of audio quality, they're a sign of the limitation of the medium. You simply cannot master very loud on standard 33 rpm LPs. Gain values alone say nothing, as you can master much louder on CD without adding any audible degradation.

There's a loudness race on vinyl too. The gain on today's DJ vinyl, 45 rpm 12"s with only one 7 minute track on each side, is easily 3 db higher than the same track on the 33 rpm LP version.

is it wise to obtain a LP instead of a CD given that they are differently mastered? Again with Depeche Mode (Playing the Angel) and Red Hot Chilli Peppers (Californication): Songs on CD are a complete square wave... will it sound nicer on Vinyl?

It's kind of funny that some people are actually waiting for an old album to be "remastered". Then they tell "ooh, this sounds crystal clear, this is superior to any old release". But, what's the bloody point being 98-bit remastered from original master tapes if it's compressed to ¤#&&#/¤%¤/? It's sick that the public opinion is that "remastered, oh it must be better."

OT: I feel the same way about people buying a widescreeen to watch 4:3 content.

I might have to disagree, or at least maybe I don't know how to master. All my mixes replaygain to around 0dB . Any further compression damages the sound, especially the transients; something that I hear all too often in the "mastered" sounds, especially the big studios. The modern sound is probably offset by BBE Maximizers and the like, which is wonderful technology, but the compression is just out of hand.
Try something like TransX.

I think some of you confuse loudness with compression. They are not related. One could make an ultra compressed song at a very low volume.

Compression in itself is NECESSARY. Without it, albums would sound dull.

The real problem is clipping. Now, THAT is a problem.

Regarding remasters... well... it sure sounds cool to say they suck (you are as cool as 99% of the people @ stevehoffman.tv), but I believe some people here exaggerate the matter. Some people bitch about the Fleetwood Macs Rumours Deluxe Edition which was remastered. Well... sorry... it sounds better to my ears than my old "ghost" remaster. Same thing with the 1997 reissue of Kind of Blue by Columbia/Legacy. The Pink Floyd remasters... the Led Zeppelin remasters (some pretend to 80s CDs are "flat transfers" and the 90s remasters done with Page aren't... blah)... come on....

A few years ago I bought an Avril Lavigne album. Like the songs, like the performance, hated the mastering. It wasn't simply overlimited and clipped, but also exhibited the odd bahaviour of chorus parts being less loud compared to the verse parts.

Thus I releveled the entire album to my liking with Vegas(nice and simple multitracker). The distortion of the original mastering is still there of course, but at least I now have quiet parts and loud parts, even if they don't sound all that good.

Bob Katz mastered a couple of Foo Fighter albums where heavy duty limiting is still present, but at least he has the skill to keep the loud parts louder than the more quiet ones. Most of the time anyway.

Btw, I mix for a living, and I'm not entirely shure anyone in the mixing business or mastering business actualy likes situation either. I hear a lot of engineers saying how producers bring their favorite loud record along for the mastering process. Bob Katz even mentioned doing three versions of albumbs, an open breathing, a tight one, and last and definitly least the competetivly compressed version. The bands usualy favour the open version, while the label and producers will often pick the competetivly compressed versions. That's why I hardly buy records any longer. C'est la vie. Maybe I'll get in to Jazz and classic music more from now on.

In many cases the vinyl releases of new records is just cut from the digital master. So although you can't press squared off clipping onto vinyl, it's still going to be as compressed as its CD counterpart.

However the vinyl release of the Chili Peppers latest album was made using all analog mixing and mastering. From what I've heard there is far less squashing than the CD.

Just to put this out there about re-mastered LP's and CD's: Mobile Fidelity has always done a great job of re-mastering without any of the pumped up sound and clipping with compression. I do not know how they are doing recently. Maybe someone else is current with them.

Remember that those replay gain values are computed *after* normalization. In other words, the LP transfers peak at 0dB (in fact, to -0.1dB, as that's the way I normalize my LP transfers). Given that, the replay gain value is a pretty accurate measure of the dynamic range. So this demonstrates the fact that LPs typically have *more* dynamic range than CDs. How ironic, given that the CD medium itself is capable of far more dynamic range than vinyl - it just doesn't get used.


I'm not surprised. There is a desire to make everything as loud as possible these days. Thankfully, 33rpm LPs put a mechanical limitation on how much can be done.

The race to make everything louder is stupid, in my opinion. Clearly, the public have managed to sink to the level where they can't use the volume dial when listening to their CDs! Radio stations even try to make themselves sound louder, as they believe that it will make them stand out as "better".

People who're actually aware of the pointless practises and subsequent waste of a format involved are left to whine about it in places much like this.

Not necessarily - you make two assumptions here:
- there's no significant distortion in the LP signal in the loud parts
- there's no significant noise in the quieter parts

One of these might be true, but both generally do not hold for vinyl: if the peaks are not very high, the signal is "clean" up to the peaks. But the quiet parts are then drowned in surface noise, rumble, and what have you. Normalizing just raises the noise floor with it - you get nice looking gain values but terrible S/N. If the peaks are high, then the music is distorted way before the actual peak value - you're then normalizing to a distorted signal.

I'm not making any such assumption. Indeed, I am the first to acknowledge that LPs have both significant distortion *and* significant noise.


What is the connection between a high noise floor and low ReplayGain values? Increasing the noise floor *decreases* dynamic range, so if anything it would result in slightly *bigger* ReplayGain values.


And a distorted signal sounds *louder* than a clean one. But as far as I am aware, ReplayGain is not able to take account of distortion when calculating the perceived loudness, so I don't see the relevance.

The bottom line is that ReplayGain values are a good indication of perceived loudness. If something has a larger negative ReplayGain value, it means that it sounds louder. So if an LP has a smaller negative ReplayGain value, that means it isn't as loud. And if that ReplayGain value was calculated *after* normalization, it means that the difference between the peaks and the quiet bits is bigger, ie. there is more dynamic range.

Look, I'm not here to argue that LPs sound better than CDs. I happen to think the opposite is generally true. I merely questioned the suggestion that it's only major label CDs that get additional compression applied to them. (I refer you back to the fairly high ReplayGain values I measured on two recent non-mainstream releases, ie. the Refugee and Rick Wakeman CDs I mentioned earlier).

You *might* argue that a modest amount of compression over and above that found on LPs is actually a good thing, and that if only it were possible to do it on an LP, it should be. That would be a fair point worth debating.

Louder does sound better; only if people did't have volume switches.

But the intense compression applied to much modern music is surely selling the format short. I suppose, if the mainstream doesn't complain, nothing will change. General listeners will probably not notice or care about the finer points of mastering and dynamics, so there isn't a lot that can be done. It's all about competition and sales . . .

But make sure you have not even *one* pop or click on the LP, which could skew the normalization , since the tick might well be the loudest thing on the record..that would lead to a falsely 'wide' dynamic range reading.

An interesting exercise for those using Foobar2k (and who have extensively-tagged files, showing for example the *CD* release date as well original release date) is to sort all your tracks by album replaygain values, and see which genres/years tend to correlate with highly positive, zero/nearzero, and highly negative values.

Perhaps I should not have mentioned radio play, although that may often be a guiding factor of the philosophy. I suspect the "loudness theorem" is applied to everything done by most labels that subscribe to the idea. Still, there are many that don't. Is anyone claiming that labels such as Chesky or Windham Hill, to pick just two from the quality oriented crowd, have succumbed to this practice?

Possibly I am out of date. I've been adding more than 100 CDs per year to my collection by dint of transfer from vinyl, so I haven't been buying many new commercial releases. This 2001 issue I just pulled off my shelf may be the latest I have. Things could have changed since 2001, but that seems unlikely to me.

EAC says the peak values of the disk's tracks range from 40.1% to 98.3%. That 98.3% peak is at -0.14dB. Relatively few peaks across the entire album are close. I suppose things could have been mastered into clipping, then backed off for the release copy, but it neither sounds that way nor seems rational.

The difference between minimum and maximum RMS values seems to run about 60dB across most tracks, with RMS minimum reaching about -72dB where there is anything at all beyond dither (and -83dB where there is nothing except dither). Quite possibly some compression was used to improve the recording, but it is certainly not like the majority of current pop releases.

By the way, no one who does vinyl transfers with any intent to improve the recording via declicking would have any click/pop peaks even remotely close to maximum music peaks. This is very easy to detect and control.