Pan Law: Pan levels in PT? (link)

[Philthy]

Oh, Marc, I just noticed your last post, which snuck in as I was typing mine... so it is the same on the HD system!

Interesting, very interesting.

I'd love to know what the tech guys think.

Oh yeah, and Digi too!!!

[Barnaby Bristol]

Yes, the Digi pan pot attenuates 2.5dB of level when in the center position. Constant power pan pot. If you aren't using the hiRes mixer plugin, the pan pot hard left/right and volume at '0' will pass bit clean. Some mixers are designed to be unity and maybe bitclean in the center position and to add gain when panned outward.

I believe Digi had a very intelligent reason for selecting (close to) 2.5dB. They chose a dividing integer that results in an output of only 2 bits greater resolution. Example. Create a simple 16-bit session with one track of audio (16-bit of course) and a bit meter on the output. Panned hard left or right (with the fader at '0' and not using the HiRes mixer plugin), the bit meter will show 16-bit. Pan the track exactly up the center and the bit meter will show an 18-bit output. What are the three most likely pan settings? Hard left, hard right and up the center. While this still makes sense in today's higher bit systems, it was a smart move for PT-PCI systems with 18-bit DACs.

[ckevperry]

Some older SSL's are down 5db at center.

There are no serious flaws with the digi mixer. But I know that running direct outs to some analog mixer sounds better a lot of the time. This is the euphoric quality of non-linear analog systems. Our ears like the distortion and noise introduced by summing amps, tape and such. But there is no serious flaw in digi's mixer. (The 2.5db pan law is wrong, but not fatal.) Massenburg just stated on his forum that the problem with mixing in protools is not the sound but the ergonomics. But believe what you will...

I've tested and run the same tracks at the same levels thru Nuendo, Logic, Sonar, Protools and found that if you are careful you can get the mixes to cancel when you sum and invert.

And BTW, HD's mixer offers more headroom, but other than that there is no technical reason why it should sound better. The 48 bit chip spanning mixer is nice, but this can be done in Mix as well with a mix that stays on one mix. I have had 47 tracks stay on one chip.

[Henchman]

Originally posted by ckevperry:
(The 2.5db pan law is wrong, but not fatal.)

<font size="2" face="Verdana, Arial">Now let me explain to you why this is fatal and a serious flaw, as I did on the other board:

When I do an 8 track split stem for a broadcaster. I try and mix as close to the allowable level (-10dbfs) .
Now when that get's layed back by the transfer house to the Digi Beta's, various broadcasters have various requirements. One of them is as Follows: Tracks 1 and 2 Full stereo mix. Track 3 summed mono mix of music and SFX. Track 4 Mono mix of loaction sound, Interviews and Voice over.

Now when they happily combine my STEREO mixed SFX and Music, which now become 1db louder because of the ****ed up panning law, I get the master back because it goes over 1 db. That costs me money, and would make my client lose faith in our ability to deliver proper masters, resulting in lost revenue.

That is a pretty big problem IMO. That is the real world.

Also, I'm sure it's nice doing automated pans and finding that the instrument you're panning is luoder by 1 db whenever it is in the center, because of the -2.5 panning law.

And this is acceptable?

We're not talking audio quality or the summing bus, blah,blah, blah. We're talking a basic software issue. I can't see how the HD interface can fix a software issue.

[Noiz2]

So on an SSL your problem would be worse. Now how many times did your masters get returned before you figured out that you could lower the level 1dB and have them accepted? Know your tools, know their limitations and kinks...it's your job.

[Henchman]

Originally posted by Noiz2:
So on an SSL your problem would be worse. Now how many times did your masters get returned before you figured out that you could lower the level 1dB and have them accepted? Know your tools, know their limitations and kinks...it's your job.

<font size="2" face="Verdana, Arial">I didn't mix tv shows on an SSL. Stop making excuses. My job is to mix. not make up for poor design.

[stoneinapond]

So Henchie,

Since when have you been providing material to your clients mixed on ProTools? I thought you were a Fairlight shop. [img]images/icons/tongue.gif[/img]

Or is this just one more excuse.....

Peace

[Henchman]

FYI, this topis was started on another board, and moved here by a PT user.

[stoneinapond]

That I am aware of. Just not sure of the relevence.

But enough. It would be good to hear from Digidesign as to why they chose that figure.

Not holding my breath. [img]images/icons/cool.gif[/img]

[CO2]

Hello:

WOW!!!

Well, George Massenburg is correct in his explanation. Ideally, pan pots should be down 3 db at center.

What is interesing to me is that for years, analog consoles had various levelsets for pan center. Some were -3 db, some -4.5 db and some even -6 db! Part of the issues were the tapers of the pan pots themselves and others were the summing amplifier circuits.

So, I learned early to mix, and monitor through, stems recorded to tape and calibrated to 1 K tones. As long as my stems were calibrated to the tones, the pan center levels were far less significant.

To this day, I still mix, and monitor through, stems recorded in Pro Tools and calibrated to 1K tones.

Best Regards