Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[CWHumphrey]

Curve Dominant wrote on Fri, 04 November 2005 06:10

CWHumphrey wrote on Thu, 03 November 2005 22:44 electrical wrote on Thu, 03 November 2005 19:47 I propose that we use 0vu= -15dBfs as the transfer level, as that provides a nominal equivalent in headroom to any good analog system. Bob O.'s assertion was that this issue may be level dependant.  Wouldn't make sense to also test -18dBfs or -20dBfs?  And for the sake of accuracy, when Mixerman discovered this, where was his operating level at? Hmm...My understanding is with PT, input at around -16dBfs = optimal use of the kit's headroom. Not sure if -15dBfs would yield a noticable detriment in comparison.

That is the question, isn't it?  It's all about the details.  At Larrabee, we installed one of the first SSL 9K J's in LA.  After a few years, some engineers complained that they couldn't hit the mix buss as hard as they could the 4k's.  Mark Gruber and I concluded that since the console was direct coupled, and therefore more bottom end succesfully traveled the various signal paths, hitting this mix buss super hard (as in pegging the VU's) was a bad idea.  Go figure, SSL had optimized the mix buss for 0VU.

Most PT systems I've used have been setup for -18dBfs.  It's same old argument we used to have with the dat machines.  Do you push up the operating level to use up as many bits as possible?  Or do you set them up as a unity gain device?

Isn't this what has go us all into so much trouble with CDs?  Look, we can push the level louder than vinyl.  We need to push the level louder so we use up more bits...or do we?

This test needs to be attempted at different operating levels.  For that matter, I'd like to see this whole thing repeated at J.J.'s, since he offered.  Also, we'd have the chance to try the whole thing with his Apogee converters.

If there's an issue, what specifically is the cause?

Cheers,

Carter William Humphrey

[malice]

Curve Dominant wrote on Fri, 04 November 2005 07:10

CWHumphrey wrote on Thu, 03 November 2005 22:44 electrical wrote on Thu, 03 November 2005 19:47 I propose that we use 0vu= -15dBfs as the transfer level, as that provides a nominal equivalent in headroom to any good analog system. Bob O.'s assertion was that this issue may be level dependant.  Wouldn't make sense to also test -18dBfs or -20dBfs?  And for the sake of accuracy, when Mixerman discovered this, where was his operating level at? Hmm...My understanding is with PT, input at around -16dBfs = optimal use of the kit's headroom. Not sure if -15dBfs would yield a noticable detriment in comparison. It would be in hitting the convertors significantly harder than -16dBfs which might yield the "lacking balls" effect which Mixerman claims to have heard. No surprise there: Digital likes headroom. "Loudness Wars," anybody? What's the first thing to go when we quash the headroom out of our mixes...balls, right? What's one little dB, guys...how about you use 0vu= -16dBfs as the transfer level. No biggie. Glad to see Mr.'s Kashiwa, Albini, Fletcher, et al going to bat on this. Happy hunting!

Curve,

How can you determine an optimum input level to PT, when you don't know the output level from tape?

Do you understand what an operating level is ?

And even if you know the operating level of the tape, how would you know if the operator didn't chose to slam some of the tracks for some better sounding purposes.

For some tapes, -15dBFS might even result in clipping the ADs converters. Although I have no doubt Steve knows what kinda material he would bring.

You have to know with what tape you are dealing with. Otherwise, we should chose safer calibration such as -18 dBFS or even -20 dBFS.

malice

[Fletcher]

+15 was kinda known as the level at which 456 started to compress/get ugly... it was around +14 for Scotch 250, it's around +18 for 499. The thing was that 456 was forgiving, and 250 would bitchslap you and call you a fuckup in front of the entire room if you went too far... it was definitely sweeter sounding tape as long as you didn't go over the line... but sometimes you ended up going over the line a little [or at least I did on occasion].

Now this is 15db>O [a.k.a. 185nWb/m]... which in the case of a 3db>250nWb/m alignment leaves a little under 6db>0VU [0VU =+4; 3db>250nWb/m roughly = +5.2 though many call it a "+6" alignment... I'm not sure how the math adds up in the addition of +4 (is that dbm or dbv... I always get confused about that) to +5.2 on the tape].

I would be in favor of two things to be covered in this test, which is indeed going to be a "converter" shoot out in many regards [and yes people will say "interesting" and go back to doing things as they have always done them... otherwise we never would have moved forward from 1" 4-trk [or 2" 8-trk] but I digress.

We should cover two sample rates; 44.1 or 48k vs. 96k and two different headroom specifications.  I usually work at -20dbfs... no particular reason, it's just that I find I can be a good measure less careful at that level [as I was a good measure less careful with levels with 456 than I was with 250].

I would propose the following methodology:

Tape is put on the deck and aligned.  Rough mix of the material is generated.  If the desk in the CR in which this is happening has "TR" capability, a "TR" of the monitor mix is run.

The outputs of the tape machine are run directly to the inputs of PT.  Two passes are run with a headroom of -16dbfs [one at 44.1; one at 96k]; repeat with a headroom of -20dbfs. Once the material is recorded to PT, it is played back [and recorded; recording specification to follow]

Repeat 8 pass procedure [4 in the recording of each sample rate with each headroom allowance] with RADAR; repeat procedure with any other converter that is on hand.

At the conclusion of each pass a digital recording of each pass is made through Prism converters [which I believe are owned by CRC] to a storage medium [not Pro-Tools] at 24/96 for subsequent comparative purposes; and 16/44.1 for potential distribution.

This will bring an end to our broadcast day on this thread... I'm starting a new thread that will deal with "methodology" and methodology and methodology only.

Social banter will be unacceptable; I don't see this being a long/protracted debate but we'll see how it goes.

My goal is that by the time we get to November 18th that we will be able to follow a defined procedure and not have to make up a damn thing as we go along.