Power 1, 2, or 3

What's the difference between Power 1, 2, & 3 when "exporting" audio?

Actually, I'm not sure what any of the export options are about. I like Power 3---sounds are "clear," but a tad bit weak.

My stuff covers a lot of styles---every style has its own best setting?

From the online help file -

SONAR Producer offers five kinds of dithering:

• Rectangular--essentially white noise, no noise shaping. Advantages: least CPU-intensive, lowest signal-to-noise ratio, preferable to shaped dither when successive dithering can occur (e.g. bouncing, freezing). Disadvantages: suffers from intermodulation distortion, higher perceived loudness than Pow-r dither.
  
• Triangular--higher level than rectangular, no noise shaping. Advantages: low CPU-intensive dither, superior to Rectangular as it does not suffer from modulation noise effects. Preferable to shaped (Pow-r) dither when successive dithering can occur (e.g. bouncing, freezing). Disadvantages: higher perceived loudness than Pow-r dither.
  
• Pow-r 1--noise-shaped dither. Advantages: less CPU-intensive than Pow-r types 2 and 3, lower perceived loudness than Rectangular or Triangular. Disadvantages: less noise shaping than Pow-r types 2 and 3, not recommended for operations where dither will be applied successively (e.g. bounce and freeze).
  
• Pow-r 2--noise-shaped dither. Advantages: lowest perceived loudness, highest quality settings, recommended for audio export. Disadvantages: highest CPU-intensive settings, not recommended for operations where dither will be applied successively (e.g. bounce and freeze).
  
• Pow-r 3--same as Pow-r 2 except most CPU-intensive and transparent of all choices.
  

Q.

Thanks for the response.

It was my understanding that dithering reduces harmonic distortion and that it does not add this form of "noise."

Also, in the online guide it states that non-Power dithering has a "higher perceived loudness." In my opinion, Power dithering does seem to squish something out---whether it's just me "perceiving" something---I don't know. However, Power does sound better when applied to WAV and 256 MP3 exports. However, if there is a large orchestration going on, it seems to warble a bit on 128 MP3 exports.

I guess a lot of experimentation is in order.

Thanks again.

I think that means that the triangular dither itself is perceptually louder than pow-r, not that a mix using that dither will be louder.

Pow-r dither is less audible because it involves noise shaping, where it places noise in less sensitive areas of our hearing.

http://en.wikipedia.org/wiki/POW-R_Consortium

"Pow-r 3--most ... transparent of all choices."

That was all I needed to read. I set it to P3 when it was first rolled out, and haven't looked back.

They differ in the spectral distribution of the dither noise. Here is a graph showing the difference. According to the developers, 1 is for speech, 2 is for limited-dynamic material such as pop/rock, and 3 is for wide dynamics such as classical music.

As far as choosing one, throw a dart. It takes a highly trained ear to hear the difference, and if you're producing mainstream pop/rock/rap/metal with a typical 4 to 8db dynamic range (or for that matter even if you're at a conservative 14db range) you will not hear that difference.

ORIGINAL: bitflipper

They differ in the spectral distribution of the dither noise. Here is a graph showing the difference. According to the developers, 1 is for speech, 2 is for limited-dynamic material such as pop/rock, and 3 is for wide dynamics such as classical music.

As far as choosing one, throw a dart. It takes a highly trained ear to hear the difference, and if you're producing mainstream pop/rock/rap/metal with a typical 4 to 8db dynamic range (or for that matter even if you're at a conservative 14db range) you will not hear that difference.

Interesting post.

Since it's probably extremely difficult to hear the different dither types in those modern, undynamic styles, so you think it's necessarily better to dither only once, instead of multiple times during the mixing and mastering?

In otherwords, is it easier to hear dither if you dither to 24, and then dither to 16 bit, as opposed to dithering once straight to 16?

when I go to an 128 MP3 something goes awry

Not a dither issue. You're either hearing clipping distortion (see this thread) or aliasing. If the former, leave at least 2db of headroom for exports destined for MP3 encoding. If the latter, try encoding with VBR if you're currently using CBR.

ORIGINAL: bitflipper
They differ in the spectral distribution of the dither noise. Here is a graph showing the difference. According to the developers, 1 is for speech, 2 is for limited-dynamic material such as pop/rock, and 3 is for wide dynamics such as classical music.

I've seen conflicting information as to which of these is which, and now that I tried searching online I'm seeing even more conflicting information. There are posts in some user discussions stating that algorithm 1 is the one for highly compressed (low-dynamic) material, and I've seen other comments that algorithm 3 is the one for spoken-word. Lately the Wikipedia article seems to have the same rundown that bitflipper posted above. I wonder how we might be able to verify which is truly which.

I´m using OZONE4´s dithering, which I heard should be great and leave Sonar´s unchecked..

Looking at the pdf from Weiss, you can see that the type 3 dither is spread out over a wider frequency range and that it varies according to the sample rate.

Type 3 at 44.1KHz has more dither energy in the low and upper-midrange frequencies, which would seem to make it less suitable for spoken word, whereas Type 1 puts most of the energy much higher, near the Nyquist frequency. Types 2 and 3 are similar at 96KHz, but Type 2 would seem to have an advantage at 44.1.

I still maintain that for restricted-dynamic genres it makes so little difference that it's not worth agonizing over. Even with Type 1, most of the energy is at -130db (!), easily 60db below even the quiet passages in your song. A 60db difference is usually considered virtual silence (e.g. the RT60 standard).

ORIGINAL: bitflipper

when I go to an 128 MP3 something goes awry

Not a dither issue. You're either hearing clipping distortion (see this thread) or aliasing. If the former, leave at least 2db of headroom for exports destined for MP3 encoding. If the latter, try encoding with VBR if you're currently using CBR.

Heh, I didn't even know there were different types of MP3 encoding. I'll have to look into that. Thanks!

ORIGINAL: Melodeath

ORIGINAL: John

I've been wondering, if the Ozone output is routed to the master bus, then doesnt the fader ont he master bus bring the bit depth back up to 32 (or 64 if using double precision engine)? In which case, shouldn't you dither from Sonar on export down to 24/16?

This brings up a question. Where else would you have Ozone placed? Wouldn't be on the master bus?

Two things to keep in mind.

You will never hear a 32 bit or 64 bit audio file. No DA/AD converter is better then 24 bits in is output. Thus truncation must occur.

I do believe that Sonar also applies dither to the streaming sound going to your sound card when in playback.

Right. I've always assumed so since you can set which dither you want on playback in the Audio options.

And it's true that truncation must occur since D/As are 24 bit, but that's why we dither, to randomize the quantization error. That's why I was wondering abotu the inner workings of a dithering mastering plugin routed through a fader in the DAW. 24 bit out of a limiter into 64 bit processing by a fader back down to 16 bit for export is a bit convoluted.

This is the main reason why "mastering" software like Ozone and T-RackS have standalone versions as well - one of the offerings of such software (especially Ozone) is the specialized dither algorithm included in it, but you lose the benefit of that if you master from within SONAR. Mastering in SONAR using Ozone or T-RackS as plugins, you get the benefit of a familiar working environment (if that's what you're looking for), but you're limited to SONAR's own dither as the final dither. (Not to say that Pow-R is limiting; but Ozone's dither is no longer the CD-16-bit-conversion dither that you're using.) Mastering in standalone Ozone or T-Racks, you're getting their full signal chain including the final dither.

Now here's a follow-up question, since I'm not familiar enough with Ozone to know the answer: When using Ozone as a plugin, is it possible to tell it internally to convert to 16-bit, so that the audio coming out of the plugin is limited to 16-bit data? In this case, turning off SONAR's own dither would effectively make Ozone's dither still be the final dither, as the conversion would have already happened and any resulting data in bits 17 and above should be zeros anyhow, and therefore safe to truncate. Any thoughts?