Posted - 07/15/2010 : 19:55:53
So, why do some things work and others do not?
One thing you will learn in using NAT, unless you are remarkably lucky and never run into unusual technical issues, is that it is directly related to the science of digital data conversion. If you never had a desire to know more about the vibrating crystal slithers that convert sound waves into numbers, then what are you doing trying to use NAT? You have to have some crazy unusual desire to understand things that normal humans would rather take for granted.
Think of Lavry, who makes excellent converters in the almost affordable range and amazing converters in the hardly affordable range. His background is in the medical field. Clean transfer of minute data to an exhaustive degree is what NAT is going to uncover.
So, let's say you want to record something really simple like a preamp. What is a preamp? What happens with it when you set it up in a loop to sample through NAT?
Well, we've simplified the process down to a few steps. Giancarlo, actually, has created the ideal scenario baed upon current limitations. These limitations, I might add, which are ideal for preamps to the degree they are meant to be used.
A preamp is associated with two fields of related science: Hi Fidelity sound wave amplification, and recording arts equipment amplification. In the HiFi world, preamp is actually the stage that allows for interconnect between devices that make sound but have no amplification of their own, and the connected device that adds volume and sends it to the final output line or speakers at proper converted impedence. In recording, the preamp is primarily for mic, then for direct inject or instrument line level, and then balanced line pass-through. The preamp in the studio serves the purpose of giving an output level that is expected at the other end. The ohm rating is directly related to source, load, current, volume, and sonic character.
So, when you are sampling a preamp, you are measuring its characteristic when connected to what goes in and what comes out.
A LOT of talk has taken place in the past on the importance of converter quality and sampling, cable types, sample rate, etc. Believe me when I say this, that none of these things are nearly as audible as the result of a bad relationship between connected interface types. You do not want to improperly drive a mic or d.i. and try to retrieve the right sound in NAT. The quality difference between 44.1, 48, 88.2, and 96 is slim compared to the teeny analog volume difference and its impact on the rest of the chain.
So let's look at this once again:
We have converters, cables, sample rate lock, and a computer running NAT. What is the goal?
To sample something that sounds the same as when we use it in analog. Right on!
So, what are some reasons that things come out wrong?
Well, this will be revisited a few dozen times I am sure, but here are a few things I've run into, debugged, tried and tried again, and learned from.
1. You only THINK you are using a certain sample rate.
If you have an elaborate mixing, mastering, or production set-up, you may want to develop a plan for sampling NAT separate from all other tasks. You wouldn't believe how easy it is to misconfigure if too many things are going on. If you have a master clock, make sure it is in fact the master, that it is locked to the rate you want, and then test it in a DAW or editor and then visually to make sure you have that rate and not another.
How do you know you have the correct rate?
Using an FFT or any decent visual analysis, take a sample recording clocked to your system frequency and load it in visual analysis. If you are at 44.1, you will see the end of the wav data at 22,050Hz. At 48k you will see the ceiling at 24,000Hz, etc. up to 96k which ends at 48,000Hz.
Why mention all of this? Well, if you are sampling an eq and it comes out wrong, and you did everything right, but your highs sound midrangy or vice versa, you probably were monitoring one rate and sampling another, or you set NAT to the wrong rate and it convolved that rate.
2. You are sampling using a lot of distortion kernels and you haven't gotten good at sampling yet.
Distortion has a direct relation to frequency and dynamics. It is extremely tricky to truly analyze them separately, and you have to know what you are looking for to gauge a correct dynamic program. Make sure you are convinced of the quality of a clean preamp program first before trying to make a complex one. Check it with other samplers (I help edit programs when I can).
3. Don't try to slam levels to get lots of character!
Compression and its different forms take their own category in Nebula. They can be created, but I am personally certain it is not ready for the full compression sampling. But, this also means that the natural compression that occurs in analog must be eliminated from the concept of sampling things like preamps, tapes and line driving items, at least until the sampler understands the limitations and can make it work. Once you know the right relation of volume, converter, and hardware item to be sampled, you can eventually get good at adding more distortion kernels, and in some cases you can match the natural saturating overdrive that occurs in ultra-high end amps and not-so high end gear as well. Patience is a virtue. I didn't keep the first 300 hours or so of results in any programs. But I am thankful for what I learned in the process. Now I feel comfortable in sampling saturation that truly adds weight and volume just like real analog. It took 2 years to get it right.
3. Test your settings over and over again. It is easy to think you are doing everything right and then find that there is a whole series of eq settings or preamps not working correctly. Strange things can happen in alignment and I will go over them as they become relavent. Be prepared to do a lot of visual analysis. I personally have a system for fine tuning everything by ear and using music. It is much for informative to compare real versus memorex with music that constantly fluctuates rhythmically, in spectrum, and throws programs through the course.
You must be logged in to see this link.
Software, Mastering, Tutorials