Adjusting Sound Level

[jaskier]

Hello,

I've noticed that different CDs have different volume level - this gets quite annoying when I'm using the 'shuffle' function.

Any idea how to automatically adjust volume to the same level? ITunes has a nice function in options for this, but nothing similar in Sony's software...

I would appreciate your comments, thanks

[Gokce]

I so do not expect Sony to have thought of something like that, considering so far they have missed lots of much more crucial stuff in their MD technologies.

Properly authored audio CD's should have a somehow standard sound level, since they are all in digital format. If you are having problems with a particular CD having too low or high levels you can rip the tracks to wave files and "normalize" the tracks with a wave editor before transferring them to MD by SonicStage.

[dex Otaku]

Properly authored audio CD's should have a somehow standard sound level, since they are all in digital format. If you are having problems with a particular CD having too low or high levels you can rip the tracks to wave files and "normalize" the tracks with a wave editor before transferring them to MD by SonicStage.

Unlike the film industry, the music industry has no set standards for what levels are to be used on any medium, including CD.

The general guideline is to "load the medium nominally," meaning several things:

* the dynamic range is compressed to fit between the noise floor and the peak levels of the medium [compression has basically always been used for all analogue tape formats, LP, &c. that had less dynamic range than CD - this raises the average level of the recording; one reason being that it's then far enough above the noise floor for the medium's weaknesses to not get annoying]

* Peak volume is kept below levels that cause obvious distortion

* EQ is applied to also make up for the deficiencies of the medium

* in certain cases [as with cassette tape] noise reduction [as with Dolby B, C, and S] or compansion [dbx I and II] is also applied

With CD EQ is generally not applied, though there are exceptions to every rule or guideline. Compression is applied depending on the type of recording and its intended use [radio mixes and singles are traditionally more compressed than albums, for instance]. Peak volume is, of course, kept under the brickwall limit of digital.

The fact that CD is digital does not force any kind of standard on the mastering processing at all. In the 1980s and part of the 90s, CDs were mastered so that the original dynamic range of the recording, without any compensation for the medium [since it didn't really need it], was kept. Basically, master tapes would be digitised and then made into CDs.

Since then the rage has been to use what we now call bit-pushing, which is really just dynamic range compression made a thousand times worse than it ever was in the age of analogue. The advantage to it is that it makes a CD seem a lot louder by making it so you can get more volume out of your player [turning the volume up seemed to be unacceptable for some reason]. The disadvantages are many, including severe listening fatigue, obvious distortion in the recording [if you rip and look at almost any recently pressed CD you'll find that most tracks will actually look like they're solidly clipping, i.e. distorting, from beginning to end], equipment damage [because the average level is so much higher, and people turn it up so it's loud, and damage their equipment, not to mention their hearing], and so on.

Standards sometimes exist by engineer, by studio, or even by record label.

The only standard that really applies, though, is to keep peak levels below digital maximum. The average levels of a recording will change depending on the total dynamic range of the recording, i.e. the range between it's loudest and quietest parts.

Since the dynamic range of basically every recording is different, the average volume is different.

The only way to make up for this is to normalise the recordings. This is best done by measuring the RMS [average] volume of a track and adjusting the overall volume of it, then compensating for any distortion by doing none other than dynamic range compression to it.

iTunes auto-levelling is rather dismal, if I may say so, but this is less a reflection on Apple than it is on the rather large difference in dynamic range from recording to recording. It is extremely difficult to make a normaliser and compressor/limiter that will work with everything you put through it without causing severe distortion to one extreme or the other [i.e. either quieter or louder material]. They chose a middling route and it happens to work with most recent recordings, but anything older [and IMO, properly mastered rather than totally mutilated as most things are now] will still end up being too quiet.

Conversely, the film industry has rather rigid standards that have developed since its inception to make sure that volume levels in a soundtrack are always within a consistent range, so that any playback system in any theatre can reproduce it properly and with relative consistency compared to other theatres. THX has helped this in many ways.

The audio industry has no such standards, and the medium, be it digital or analogue, imposes no standards other than to fit the recording onto it so it's playable.

[jaskier]

Wow, that's impressive - thank you for your time! I wish the guys in record companies would read this.

Actually I've noticed that some recently published CDs have these annoying noises. Now at least I know why.

[Gokce]

wow, for me too.. dex is right, I somehow did not consider the dynamic range and thought only the peak-average digital levels would make up the volume.

just to clarify, by compression i guess you mean amplitude compression, not data compression?

thanks for the valuable explanation.

[dex Otaku]

just to clarify, by compression i guess you mean amplitude compression, not data compression? 

Yes. More accurately, dynamic range compression.