Remastering (Improving) your own Audio Albums as Big F.Y. to corporate Loudness War trend.

[LF78]

That isn't clipping. That is very hard limiting.

I'm not a mastering expert so I cannot reply with any insightful and detailed argument, but it really seems digital clipping to me. And unfortunately that album doesn't sound good too.

 

[digicidal]

And unfortunately that album doesn't sound good too.

Of course not, Hillel isn't on it... and if he's not on it - then it's crap IMO. Doesn't matter if the whole thing is smashed against the bumpstocks or has a DR of 70.

Not sure that is an issue in any way, as modern decent-bitrate lossy codecs are more than good enough for their use. Remember that people listen to vinyl and FM radio too...

Naturally, my point was simply that most people don't even care about file compression that definitely audibly degrades a recording (depending on bitrate of course) - so they're certainly not going to worry about signal compression that makes it sound better (or at least that's what most engineers of popular recordings seem to think). If half of their song is too quiet to hear on the bus or in the car - they definitely will care more about that than a few digital artifacts or a slight loss of fidelity.

 

[audimus]

If half of their song is too quiet to hear on the bus or in the car - they definitely will care more about that than a few digital artifacts or a slight loss of fidelity.

That is a key insight being missed on this thread.

There are two reasons for the compression trend. One, a high dynamic range is not good for listening anywhere except with better-than-earpods transducers in relatively quiet environments. Most recorded music these days is consumed outside of that - mobile devices with ear plugs, in noisy environments of commute, cafe, bar,etc. The latter prefer music that has a constant level over one with high dynamic range which can be a conversation breaker. Imagine a patron talking loudly over music suddenly sounding loud when the music dies down to a quiet passage! The streaming services that supply to this market primarily are not going to ask for less compression because it is a market requirement.

Second, is the related but different perceived loudness. Streaming services want this regulated because consumers get annoyed when perceived volume changes between tracks.

Because of these related but both needed requirements, the music will still be compressed and the gain brought down to whatever the requirement is - peak or average. The solution really is to have an additional channel for critical/quiet listening for those who prefer it. But that requires two different mixes. A single mix is not going to cut it for universal use. Whether this is commercially feasible is unclear.

 

[Julf]

Second, is the related but different perceived loudness. Streaming services want this regulated because consumers get annoyed when perceived volume changes between tracks.

Streaming services can (and do) implement their own average loudness matching/normalisation.

The solution really is to have an additional channel for critical/quiet listening for those who prefer it. But that requires two different mixes. A single mix is not going to cut it for universal use. Whether this is commercially feasible is unclear.

Actually you don't need multiple mixes (or even masters, as the compression is mostly done at the mastering stage, not mixing). You need one master with proper dynamic range, and then user-selectable compression and normalisation. Trivial to implement with modern digital technology.

 

[digicidal]

Second, is the related but different perceived loudness. Streaming services want this regulated because consumers get annoyed when perceived volume changes between tracks.

But that's the thing with digital audio - even with smartphone chips there's more than enough processing power to perform equalization and compression (as well as manipulate anything else) in real time. If it only takes me a matter of 10-15 minutes to normalize a couple thousand flac files while converting them to AAC, editing their ID tags, and scaling the artwork down in order to make a thumb drive for my car... I guarantee a phone could do it with a single song with at most a few second delay from you hitting the play button.

 

[Blumlein 88]

I agree with the idea of one mix and compression/EQ supplied as needed by the user. However, the problem is both EQ and compression is used to get certain styles of sound. It is the job of mastering. Compression isn't a simple compressing to fit a dynamic range window as done by the studios. Compression can have different attack, release, and ratios over different frequencies. Some tracks get more compression than others until mixed together. Such things and far more complex things can't be done the way mastering and mixing people want them done at the end user level. These things do in fact greatly alter how the track sounds.

 

[audimus]

But that's the thing with digital audio - even with smartphone chips there's more than enough processing power to perform equalization and compression (as well as manipulate anything else) in real time. If it only takes me a matter of 10-15 minutes to normalize a couple thousand flac files while converting them to AAC, editing their ID tags, and scaling the artwork down in order to make a thumb drive for my car... I guarantee a phone could do it with a single song with at most a few second delay from you hitting the play button.

True, as I wrote earlier, this processing can be moved, in theory, to the end user device. But inline real time processing can lead to some really bad and unintended results (whether the audience notices or cares is a different question).

Besides, other than Apple, none of the other services have any control over what it is listened on to embed this on every one of their access clients including the browser access. But even Apple has to worry about processor usage for battery consumption as well as music process being in the background of other multi-tasking on the device typically to do this in real time every time it is listened to. Waste of power unless there is a way to cache processed music to make it one-time and the uneven quality is not a problem.

A replay gain calculated once and tagged and used by the end device is much faster to process and easily done in real time. My entire stored library is tagged this way, much less artifacts than actually altering the bits. But this does not compress the range which might be a requirement.

 

[digicidal]

Such things and far more complex things can't be done the way mastering and mixing people want them done at the end user level. These things do in fact greatly alter how the track sounds.

That's the real crux of the issue - because with power and freedom comes a significantly increased potential for harm as well. I think that's the biggest obstacle to self-styled mixes... if you can make it sound the way you want, you can also make it sound like crap potentially. And if you have your friends listen to it that way - they may think the artist/label is crap too, which will hurt sales. But for enthusiasts... it would be delightful IMO.

 

[LF78]

Not sure that is an issue in any way, as modern decent-bitrate lossy codecs are more than good enough for their use. Remember that people listen to vinyl and FM radio too...

Ironically heavy compressed or clipped tracks may sound objectively better in MP3 than FLAC:

https://www.har-bal.com/1026/when-loudness-wars-visit-why-mp3s-sound-better-than-cds

 

[Tool]

I agree with the idea of one mix and compression/EQ supplied as needed by the user. However, the problem is both EQ and compression is used to get certain styles of sound. It is the job of mastering. Compression isn't a simple compressing to fit a dynamic range window as done by the studios. Compression can have different attack, release, and ratios over different frequencies. Some tracks get more compression than others until mixed together. Such things and far more complex things can't be done the way mastering and mixing people want them done at the end user level. These things do in fact greatly alter how the track sounds.

Nobody is denying compression as one of useful tools in mastering audio. It just has been overused during last two decades and it's devastating to the music and artists who apparently don't have much to say in this subject. Company's are selling damaged products and most of the people seem to be fine with it. When you buy food or anything these days you most of the cases are informed enough what's inside the box, with cd's, sacd, dvda and bluray audio it's completely not transparent and clear.

 

[PaulD]

I get the need of compression for the mass market, even if I don't like the result. But what about straight clipping? This is another story, there's no way it cannot be considered a gross mastering mistake.

What does the waveform relate to pleas? Song, Time etc? Just because it is your waveform does not mean it is anyone else's, posting it is meaningless without corroboration...

 

[LF78]

What does the waveform relate to pleas? Song, Time etc? Just because it is your waveform does not mean it is anyone else's, posting it is meaningless without corroboration...

I think I didn't understand you point, can you clarify? That waveform is a bit perfect CD rip from Red Hot Chili Peppers - Californication album (track: "Get On Top").

 

[PaulD]

Ok, RHCP Californication... Which song, which second? It does not surprise me so much if it's that album, which is very well known to be very poorly made and not very typical. It's a disaster. The vinyl is known to be far better because it cannot be so over compressed for tracking reasons.

I thought it may have been the Amy Winehouse album people were talking about as no specifics were provided. I was going to pull my version or it it into ProTools and check it. I imagine mine may be the same, for that particular album, as it it well known to be one of the worst available and not typical or most... If waveforms are posted the song and second need to be posted so others can verify the data..

 

[LF78]

Ok, RHCP Californication... Which song, which second?

It's visible from the screenshot. The song is "Get on Top", a first clipping occurs at 1:59.810, a second one at 1:59.822, etc...

 

[PaulD]

It's visible from the screenshot. The song is "Get on Top", a first clipping occurs at 1:59.810, a second one at 1:59.822, etc...

Thanks for the clarification. As I said, I'm not surprised, it's well known as one of the worst examples of released music and not typical...

 

[MRC01]

Is there any way to fight against loudness war just avoiding to buy these kind of records?

I don't buy them intentionally. And when I occasionally buy one accidentally (not knowing it was compressed until after buying it), I return it. HDTracks for example will refund your money if you contact support. This kind of mastering is garbage and I refuse to support it.

Fortunately, the kind of music I listen to most of the time tends to be minimally processed and is virtually never compressed. But I occasionally do like to rock out, though I've stopped buying that kind of music because it's virtually impossible to find any modern recordings that don't compress the life out of the music.

I advocate that rock/pop music produce clean recordings for CD and high res downloads. If they absolutely must destroy the music with all this processing & compression to make it sound as loud (and terrible) as possible, reserve that for radio and streaming.

 

[Bounce44.1]

This specific album is infamous within the loudness debate. Food for thought: the finished product is the culmination of everyone involved in the project. The people involved in this project chose to push the boundaries of "normal" or "accepted" for sure. One can like or dislike the result but to attack the expertise or competence of those involved may be a little harsh

 

[MRC01]

This is more common than just a single example track. Here's a modern recording from my collection:

Here's a modern natural uncompressed recording for comparison:

The irony is: the first is a 96-24 download, the second is from CD.

 

[tmtomh]

Loudness normalization is like application of ReplayGain. Its main purpose is to keep the levels close between tracks.

It can sometime also be done with end devices/clients (like the night mode in some AVRs).

It does not solve the industry practice of compressing the music to reduce the DR so that increasing the gain max staying within loudness mandates makes it “sound” louder (tv ads do the same thing to stay within the loudness mandates but sound louder to grab attention). This destroys the quality of the music for critical listening. Loudness normalization does not really discourage this.

Loudness normalization absolutely does push back against the excessive use of peak limiting, because it uses RMS (or a similar measure) to normalize tracks. So a track that uses a lot of peak limiting and then cranks up the RMS will simply be re-lowered in volume by loudness normalization, defeating the purpose of the peak limiting.

What loudness normalization does not deal with is more conventional forms of compression, where all soft sounds are boosted and all loud sounds attenuated. The kind of compression can be a real sonic problem of course, but it's been going on forever and unlike peak limiting it is a longstanding part not only of the mastering of rock/pop music, but also of the recording of individual instruments and of the mixing process.

 

[digitalfrost]

Thanks for the clarification. As I said, I'm not surprised, it's well known as one of the worst examples of released music and not typical...

What is interesting about Californication is that multiple different versions do exist.

1. The initial CD retail version, also seen on online digital music stores. 1 run through the first chorus. Organ beep at "it's the", organ continues in the second verse.
2. The rough mixes version. 2 runs through the first chorus. Organ only begins after the first chorus.
3. The Greatest Hits and Unmastered/Unsequenced version. Organ comes in with the drums in the first verse. 2 runs through the first chorus.
4. The 2012 vinyl version. Organ comes in with the drums in the first verse, but only 1 run through the first chorus.

https://www.rhcpsessions.com/californication-variants