Ok for some reason I can't post the scanned article but I have Micheal Stavrou's article which I will condense and hopfully I will be allowed to upload the scanned article latter.
Firstly a few of us have been having a disscussion about the merits of recording levels into Sonar. I have sugested as hot as we can before we clip, this has been shot down many times. Time to put up.
It also has nothing to do with 16 bit as it has also been suggested.
This information was taken from Audio Technology Magazine issue 13 pg68
This is the first two paragraphs:
I am by no means an expert on digital signal processing but would like to share my perspective on the differences of analouge and digital music recording. If I am right, we really have to change the we hit the digital medium.
You know how I am always saying "audio often behaves in the complete reverse of how you'd logically expect it?" Digital is no exception. For example in the analouge domain, reducing gain, always reduces distortion. Meanwhile in the digital domain, reducing gain increase's distotion.
Ok he goes on to give the analogy of a skyscraper and how analouge captures the middle of the skyscraper best and digital captures the tip or the peak of the skyscraper best.
Ok we will drop into what Stav has to say next:
The skyscraper is analogous to the volume level of the sounds we record. Analouge captures the most importent elements of the sound with pin-point accuracy (the middle of the skyscraper, (where the critical work is being carried out) while the top and the bottom where the activity is less critical (the peak and bottom of the skyscraper) is captured less shaprly.
We call the pin-sharp section or "operating level" or optium recording level because here it does not suffer from transient distortion or the noise and loss of detail from being under recorded.
Every medium has an optium recording level where it's specifications shine. Everything about this we call headroom.
Parapharsing: A digital veiw of the skyscraper is somewhat different we have to refocus our camera to the uppermost section of the building, only the tip of the Skyscraper is in focus now, while the rest of the building slowly loses focus as we move the camera down the building (I hope all are following the skyscraper analogy) Why? Stavs actual words:
The pinicle of the building is captured by the bulk of the recording bits and thus enjoying a higher resolutution.
If this analogy holds true, then the sad thing is that we finish our mix onto a digital medium- with the cleanest elements being the tip of the skyscraper and send it to mastering. What does the mastering engineer do? He or She surgicly removes te clearest most pin-sharp material to attain a higher overall level. So unlike analouge, where the worst elements removed, here in the world of digital the best most focused elements are removed, leaving a grainer version of the lower level components.
I take this to mean this is why digital can sound so harsh. Parahrasing: Stav then goes onto say most good analouge gear has at least 20 db of headroom, where as digital has zero, how do you work with device that has a zero amount of headroom? Stavs words:
What do I mean exactly?? Well when you look at digitals point of maxium clarity, minimum noise and minimum distortion, (0dbfs) there is 0db of headroom above this point -you simply can't go any louder without distortion cutting in. Maxium level is digital's optimum level. Therefore it is accurate to say there is no headroom in digital recoding devices.
Certinaly you can generate a false sense of headroom by targeting your recording level by aiming for say -18db
(which some around here still advocate) but you must be aware you are degrading the signal to some degree by not recording it at the optimum level-namley you will have fewer bits describing the critical elements of your recording.
The beauty of multi-track recorders is they appear to have no noise- you can run 32 tracks up at unity and hear nothing unwanted.
When recording to digital multi-tracks don't be decived as I was by the ultra-quiet noise floor to think that from here on down is useable dynamic range.
This false sense of security tempts you to play all the track back at unity gain and hit the digital recorder with just the right level to mix itsself on playback with all the faders at zero.
This is admittedly very convienent but one of the worse things you can do to the sound quality.
Paraphrasing: And this is what I have been advocating: It is often too much trouble to optimise the recorded level and then re-adjust the output level to create balance. You really need to massage the input signal until it reaches maxium level into the digital domain and then adjust the recorders channels output fader (trim, I prefer to use the input stage to readjust the balance then mix) to acheive the desired balance.
This folks iswhy I advocate recording into Sonar as hot as possible, the best part of digtal and the best operating level for digital is as close to 0dbfs.
Michael Stavrou is a world respected audio engineer who has worked for George Martin and Air Studios.
Next week I will explain my other crazy theory, optimum operating for your preamps and other outboard gear.
I will try and post the scanned (rather grainy) article for people to read it in more depth.
Peace Neb