Rolling Stones London / Bowie RCA

[krabapple]

or is it that NR is bad in principle?

Yes.

Perhaps you'd better define whether you mean digital NR only, or also analog encode/decode schemes like Dolby. And whether digital NR includes things like declicking.

[lukpac]

Prosound Web has a dozen or so forums, all run by recording engineers and audio component engineers

http://recforums.prosoundweb.com/

Find me a link where someone claims that they can apply NR without any audible artifacts.

[Andreas]

or is it that NR is bad in principle?

Yes.

Perhaps you'd better define whether you mean digital NR only, or also analog encode/decode schemes like Dolby. And whether digital NR includes things like declicking.

Digital NR and declicking.

[lukpac]

Yes, unless you define 'degration' with something like 'changing the waveform'. I don;t believe lossy compression necessarily degrades the music either.

You don't consider audible artifacts degradation?

True, but Dolby NR is a method of perceptibly raising the S/N without necessarily degrading the music.

And I don't disagree with that. The point is that systems like Dolby change the signals to better "fit" on a tape, and then change them back later. With broadband digital noise reduction, though, there's nothing you can simply "change" - what's put down on tape is what the intended sound is. Like it or not, removing some of the sound while leaving the rest untouched is a far from perfect process. One side effect of that process is audible artifacts.

[lukpac]

And, for the record, I don't have any problems with declicking, since you're only touching very brief samples, vs. everything.

[krabapple]

Prosound Web has a dozen or so forums, all run by recording engineers and audio component engineers


http://recforums.prosoundweb.com/

Find me a link where someone claims that they can apply NR without any audible artifacts.

You mean artifacts other than the actual denoising?

[krabapple]

Yes, unless you define 'degration' with something like 'changing the waveform'. I don;t believe lossy compression necessarily degrades the music either.

You don't consider audible artifacts degradation?

Reduced hiss is an*audible artifact* of NR. A desirable one, as it happens.

True, but Dolby NR is a method of perceptibly raising the S/N without necessarily degrading the music.

And I don't disagree with that. The point is that systems like Dolby change the signals to better "fit" on a tape, and then change them back later. With broadband digital noise reduction, though, there's nothing you can simply "change" - what's put down on tape is what the intended sound is. Like it or not, removing some of the sound while leaving the rest untouched is a far from perfect process. One side effect of that process is audible artifacts.

What's put down on tape isn't what the intended sound is. It's the intended sound is *plus the unavoidable limitations of the medium*. It's unlikely artists or engineers would have included the hiss if they could have removed it.

Your assertion seems to be that no high-frequency wideband noise can be removed without also making the *signal* sound worse. I wonder how you'd prove that in anything other than a purely subjective sense.

[krabapple]

or is it that NR is bad in principle?

Yes.

Perhaps you'd better define whether you mean digital NR only, or also analog encode/decode schemes like Dolby. And whether digital NR includes things like declicking.

Digital NR and declicking.

[Crummy Old Label Avatar]

Yet another "I want to eat my cake and have it too, even though my position is untenable; by the way, I have no idea what I'm talking about" thread: analog NR = good (informal fallacy no. 1, appeal to authority: "Steve says so."); digital NR = bad (unless it's declicking on "small" samples [Luke], or not [Andreas])....

Krabapple has more patience and persistence than I do. So I'll just simply restate my earlier points:

1) To assume that you can ALWAYS tell that NR is being used is delusion.
2) Audible NR artifacts are a sign of excessive use.
3) Play all the semantic games you wish, but Dolby SR is (by name and by deed) Noise Reduction.
4) Subtle use of digital NR can be effective (as Luke unwittingly admits via his thumbs up to declicking) and undetectable.

I know this because I have used very subtle, but necessary, digital NR on select passages of my own recordings -- the most recent ones being soundtracks for large-scale gallery installations. Very carefully and patiently applied, the sound is not altered in any detrimental way, and I would defy anyone to point out NR artifacts. I myself cannot hear them, though I do know exactly where the treated samples are.

The truth of the matter is that it is easy, too easy to abuse digital NR. I literally spent two days applying it to the necessary samples. This is precisely the sort of detailed work you will more than likely not see on commercial recordings. No engineer is being paid enough to work so obsessively and meticulously on a single piece of music. It's far easier to run the entire piece through the NR filter (hence the artifacts and the damage) and move on to the next project.

That's all I have left to say about this.

[Andreas]

I am not familiar with how Dolby affects the sound. Also, if Dolby is used for the recording and/or mixing, we will never be able to compare it with a hypothetical dolby-free version. Can somebody give an example of a dolby recording or mix? Maybe Dark Side Of The Moon?

Digital NR is different, because it is applied to a finished product, i.e. the master tape. It will always affect the sound. I am 100% convinced (but I can't prove this scientifically) that any filtering/noise reduction affects the sound of every instrument or any sound in that part. Not necessarily by a huge amount, but still to some degree. Even if there are no artefacts (with which I mean audible digital noises), the vocals, drums, guitars, bass etc. will change their sonic signatures somewhat, and not in a natural way (as eq would). And that is what I oppose.

Lastly, I am tired of the argument "You only believe this because Steve says so". We are all capable of forming our own opinions. We all have learned a lot from him, but nobody (at least on this board) will unconditionally accept what he says.

And by declicking, I meant when transferring from a tape. If there is a pop on a master tape, I would rather hear it than having it edited out. Declicking when doing a needle drop is acceptable.

[Crummy Old Label Avatar]

I am not familiar with how Dolby affects the sound. Also, if Dolby is used for the recording and/or mixing, we will never be able to compare it with a hypothetical dolby-free version. Can somebody give an example of a dolby recording or mix? Maybe Dark Side Of The Moon?

Almost every professional analog recording from the 1970s onwards uses Dolby.

Digital NR is different, because it is applied to a finished product, i.e. the master tape. It will always affect the sound. I am 100% convinced (but I can't prove this scientifically) that any filtering/noise reduction affects the sound of every instrument or any sound in that part. Not necessarily by a huge amount, but still to some degree. Even if there are no artefacts (with which I mean audible digital noises), the vocals, drums, guitars, bass etc. will change their sonic signatures somewhat, and not in a natural way (as eq would). And that is what I oppose.

EQ is invariably "natural"? How so? This is a poor argument. Never mind that there is nothing about the recording process that is "natural" in any way, shape, or form.

Lastly, I am tired of the argument "You only believe this because Steve says so". We are all capable of forming our own opinions. We all have learned a lot from him, but nobody (at least on this board) will unconditionally accept what he says.

You are entitled to your opinion. But that is all it is. You say that you are "100% convinced (but I can't prove this scientifically) that any filtering/noise reduction affects the sound of every instrument or any sound in that part." If you cannot prove this, then why are you so convinced? What do you base that on? Experience, or mere "opinion"?

I did not say that "you only believe this because Steve says so." Please stop misquoting me. Rather, I pointed out that appeal to authority is a logical fallacy.

And by declicking, I meant when transferring from a tape. If there is a pop on a master tape, I would rather hear it than having it edited out. Declicking when doing a needle drop is acceptable.

That is you personal preference. Whatever.

[Andreas]

Almost every professional analog recording from the 1970s onwards uses Dolby.

In other words, we don't know how it affects the sound, because all we have are recordings with dolby from that era.

EQ is invariably "natural"? How so? This is a poor argument. Never mind that there is nothing about the recording process that is "natural" in any way, shape, or form.

Agreed, eq can be very unnatural. But would you agree that the sonic changes from equalization are of a different quality than the sonic changes from noise reduction?

If you cannot prove this, then why are you so convinced? What do you base that on? Experience, or mere "opinion"?

A mixture of both, plus common sense.

Please stop misquoting me.

Since it was not written as a quote, how can it be a misquote?

But I stand by my interpretation of what you wrote. Isn't writing about the "fallacy of appeal to authority" the same as saying "you only believe this because the authority says so"? If not, then what do I not understand?

That is you personal preference.

Sure.

[lukpac]

You mean artifacts other than the actual denoising?

Yes.

[Rspaight]

But would you agree that the sonic changes from equalization are of a different quality than the sonic changes from noise reduction?

Isn't NR just a very particular, specialized kind of EQ? I mean, what's the *practical* difference between digital EQ and digital NR?

Ryan

[lukpac]

Yet another "I want to eat my cake and have it too, even though my position is untenable; by the way, I have no idea what I'm talking about" thread: analog NR = good (informal fallacy no. 1, appeal to authority: "Steve says so."); digital NR = bad (unless it's declicking on "small" samples [Luke], or not [Andreas])....

It has nothing to do with analog vs. digital. It has to do with the way the two work. You could apply the same Dolby response curves in the digital domain and I doubt it would sound very different. It wouldn't surprise me if current Dolby units actually used digital processing (don't know if they do or not, though).

Systems like Dolby are essentially emphasis/deemphasis - emphasize certain frequencies upon recording, and then deemphasize them on playback. With something like NoNoise, though, there's nothing to deemphasize - the music itself is on the tape with the response characteristics it should have on playback. Filter the hiss, and you filter the high end of the music too.

That means you have to sample the hiss, and then try and subtract it from the music. The problem is, hiss isn't a sine wave - it isn't constant. It doesn't change *much*, which is why NR works as well as it does, but it does change. Which leads to artifacts...

And declicking isn't broadband noise reduction. It's transient removal. Heck, you could even do it in analog if you really wanted to.

1) To assume that you can ALWAYS tell that NR is being used is delusion.

Why?

Sure, I could use NR so sparingly that I probably wouldn't be able to hear any artifacts. But it wouldn't actually be removing any hiss, either. So what would be the point?

2) Audible NR artifacts are a sign of excessive use.

So you're saying that *every* known use of NR out there is excessive?

3) Play all the semantic games you wish, but Dolby SR is (by name and by deed) Noise Reduction.

See above.

4) Subtle use of digital NR can be effective (as Luke unwittingly admits via his thumbs up to declicking) and undetectable.

Again, declicking is something totally different. And how do you know NoNoise/Cedar/etc can be undetectable?

I know this because I have used very subtle, but necessary, digital NR on select passages of my own recordings -- the most recent ones being soundtracks for large-scale gallery installations. Very carefully and patiently applied, the sound is not altered in any detrimental way, and I would defy anyone to point out NR artifacts.

Send them this way then.