Skip to main content
Topic: clipping may be good... (Read 3011 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

clipping may be good...

i was just thinking...

what happens when using replaygain or other such methods when the music is meant to be clipped.

i'm not sure if anyone here knows of any but there are songs that are meant to be clipped... for example "blast the speaker" by warp brothers...

doesnt this cuase a problem?

perhaps it is just a better idea to scale?

i'm interested to see what you all think.

clipping may be good...

Reply #1
Quote
the music is meant to be clipped


ReplayGain applies only to clipping due to the encoding process...

The music is not ment to be clipped. This happens for two reasons:

1) The guy who does the mastering is dumb enough not to aply dynamic compression to the peaks that go over 0.0 dB if he compresses the music. (most of them compress in the final stages of the mastering, because commercial music is ment to be loud).

2) (this is just my opinion, so I welcome any additions/corrections)

The lossy encoders don't know how to handle properly any compressed music (at least if it has a constant level close to 0dB), so the encoded stuff, when decoded, goes over 0dB.

So using replay gain on music like Chemical Brothers doesn't hurt. ReplayGain applies only to clipping due to the encoding process (the original "clipped" sound is due to heavy compression during the early stages of mastering, and a lot of new-age music would sound dull if it weren't for this distorted sound).

Generally speaking, any "modern" CD with levels over 90% would generate clipping for a few samples, but it's better not to use -scale because this alters the input (what's being encoded, and you may end with quality loss).

On the other hand, replaygain alters the coded stuff, which  contains information that may have not been coded if you used -scale.

Hopes this makes sense (I'm too sleepy), and if not I'd ask someone that knows this stuff better to correct me.

clipping may be good...

Reply #2
Spase,

I think you're right.

If the record producer intentionally clipped the digital audio signal when mastering for CD to give a distorted sound, you'll find that the "sound" of the distortion varies from DAC to DAC. This is because DACs aren't designed to cope with such signals, and the effect they will have is slightly unpredictable - each DAC will probably add its own (extra) distortion.

If you digitally reduce the volume, there's still a clipped waveform going in to the DAC. It's not at digital full scale, but it'll still look "clipped" with flat tops to the waveform. At this lower level the DAC should reproduce it just as it is, without adding any (extra) distortion of its own. It'll still sound very distorted, but there won't be extra distortion added within the DAC.

So, if you like the sound of a full scale distorted signal with additional distortion added within your DAC, and if that is what the mastering engineer really wanted you to hear, then reducing the level either using scale or replaygain/mp3gain will change the sound, and you might not like it as much.


However, most intentional distortion is not created by recording a clipped signal to the CD - in all the CDs I've looked at (of the type you suggest - and yes I do somtimes rip them just to see what the waveforms look like!) the distortion has been added upstream (using some FX box), and any clipping is due to careless mastering. The clipping isn't what's causing the audible distortion. Removing it isn't going to reduce the amount of distortion you hear, because the distortion that's been added is more than just clipping the peaks - nice rich distortion requires a lot more than this - that's why people use expensive distortion boxes, rather than just pushing the levels too high on the CD.


Finally, if you examine a whole song in a waveform editor, and it looks clipped throughout, the truth is it's probably just heavily compressed. You have to zoom to individual samples to see how many consecutive samples hit digital full scale. Usually you find that 1-3 samples in a row hit nearly digital full scale - but this isn't clipping, this is compression. This is what the hard limiter in Cool Edit (for example) will do, and it's not adding distortion - it's just making the sound as loud as possible. Most modern music is like this - but it doens't mean that it clips. It will clip after it's been encoded (that's where this debate started!), and that should be avoided.


Feel free to disagree, but I think Spase has a valid point.

Cheers,
David.

clipping may be good...

Reply #3
Quote
If the record producer intentionally clipped the digital audio signal when mastering for CD


Yes, but I won't trust any producer who wants audible distortion but leaves the distortion to the DAC. If he wants distortion, then he will apply it before finishing the master.

If any distortion occurs after the master is done, due to clipping (or encoding), it means that it is unintended distortion, thus it must be avoided.

However, the  question was related to encoding: I don't think that there's any sound engineer who says: "Hey, this will sound better when compressed to mp3 or mpc because in that particular passage I'll have some nice distortion."

Just my 2c...

clipping may be good...

Reply #4
Quote
I won't trust any producer who wants audible distortion but leaves the distortion to the DAC.


If he or she is producing music that you want to listen to, I don't see how you have a choice - or do you contact the record company and demand that they allow a producer of your choosing to go in and remix the album from the master tapes? ;-)

David.

clipping may be good...

Reply #5
hehe, you're right

 
SimplePortal 1.0.0 RC1 © 2008-2019