Mixing In The Box / Outside The Box

  • Thread starter Thread starter Sonic Surgeon
  • Start date Start date
Well, again, digital cameras do not really saturate. When the sensor photon wells are full, no more light has any further effect. The result is all digital "1"s and at that point you get blown highlights - which appears as solid white blobs with absolutely no detail. In other words, you've reached the point of digital clipping.

Digital audio converters will capture any analogue signal whether it be clean or crunchy, so long as the amplitude of the analogue signal does not exceed the full scale input voltage for the converter. The full scale voltage for a particular product will be stated in the product specifications.

If you want to push analogue gear (say, something with output transformers) to get a character sound, you could end up with a very hot signal which does exceed full scale input to the converter. To faithfully capture this without digital clipping, you need to reduce that hot signal before it hits the inputs at the converter. You could do that with a pad, or with a variable attenuator pot (like Classic API sells).

In this way, you can capture your character sound, but completely prevent digital clipping and do so with plenty of headroom.
 
Jay, like I have been saying all along, I never said record too hot ! I guess you did not really comprehend what I said.
If you did in fact say "record as close to 0dbfs as possible without going over", you were saying "record too hot". That seems to be the part that you don't understand.

Do you understand how digital audio converters work ? All digital audio converters are made to record at their highest possible audio resolution when audio levels are closest to digital 0 db. Do you understand this fact ? Or do you know something that is not in the engineer manuals ?
You are correct about how the digital side of things work, but you are failing to realize that every bit of circuitry leading up to the actual converter is analog. This includes the signal path inside the converter box leading up to the actual part that does the conversion. The reason you should not try to record as close as possible to full scale is because you are running your preamps and everything else in the analog path too hot in an attempt to get the digital meter up that high. You aren't getting digital distortion, you are getting analog distortion. Analog distortion that gets converted to digital. Once it is digitized, it can't be fixed.


A very good analogy is: the more light that a digital camera CCD can pick up, the brighter the sunlight or artificial light, the better the picture resolution becomes. Better picture quality, brighter better colors, higher contrast ratio and more digital pixel recorded information. And the same factual analogy applies to digital audio sampling converters when exposed to higher recorded audio signal levels, gives it more audio information and thus much higher audio quality.
Any analogy between audio and photography falls apart for a number of reasons that I'm sure someone else will identify.


Getting Closer to digital converters 0 db recording audio level is always better recording quality, and it still applies today, not just in the 1990's, and or whatever time period you think it only applies. This is measurable, provable and you can perform your own test to hear this.
No, it isn't. Recording at line level yields the best recordings, it's measurable and has been proved. Again, now that the nominal line level equivalent has been moved from -9dbfs (giving you only 9db of digital headroom) to -18dbfs (giving you 18db of digital headroom) You are free to capture more dynamics because you have the headroom to do so. That's why it's better. But if you use up all that headroom with sounds that have no transients, you will be pushing your mic preamps too hard and they will start to sound like crap. Crap that will be faithfully digitized. Crap that will be blamed on 'the harshness of digital' instead of bad recording technique.


All I am saying is why record at lower signal levels just because you have a full 24 bit recording ? You then are not getting the full resolution or dynamics of a digital recording as you possibly can get, that your digital converters are capable of. Utilize them to the full and why not ? What are you scared of ?
First off, you can't record more dynamics than the thing you are recording actually has. If the dynamic range of the instrument you are recording has a dynamic range of 96db, at 24 bit, you can record that signal with the peaks at -40dbfs and still not lose any of the dynamic range. Not that I'm advocating stupidly low recording volumes, but that is the scale of resolution available.

Now, that would have to assume that the converters were calibrated with the line level equivalent sufficiently low so that the analog recording level would still have been around line level. Otherwise the signal would be lost in the noise floor of the preamp.

And Why would you be satisfied with getting only 16 bits of audio resolution out of a 24 bit recording ?
It was an attempt to show you the magnitude of the resolution you get from 24 bit. You seem to have missed the point.


And I never said using compressor/limiters improves sound quality immensely, but sometimes it may improve the sound if you have a good quality compressor/limiter and use it sparingly to get audio levels more consistent and hence louder. And yes it can make the sound much fuller, have you ever recorded vocals or bass using a good compressor/limiter ?
Yes, but you were advocating compressing the signal to get higher recording levels, which will give you a more accurate recording. Think about it, how can changing the nature of a signal ever give you more accuracy? If you have compressed it, you have already thrown accuracy out the window before you even get to the converters.

Make up your mind: Are you arguing about accuracy or sounds better?


Recording with your digital audio gear whether it is 16 bit CDR, DAT, ADAT and or 24 bit DAW or whatever, line level is not the level you record at because Line level can be any where on the audio level signal meter, its relative so line level has nothing to do with getting the loudest possible recorded audio level.
You aren't supposed to attempt to get the highest level possible. That's the point. Line level is the level that all the equipment was designed to run the best at.

And what is line level really ? Since it can be -10 db with consumer gear and at other times it can be +4 db levels with pro gear. Line level is not usually the recorded level anyways. And you may need to raise the line level on your gear using its volume control to get a louder recording. That is why you need to trust your digital audio recording meters and then also listen back to your digital recordings to make sure it isn't audio clipping, going over 0 db, the digital limit. And then you can also adjust your analog mixer with a little headroom safety margin accordingly to get your safe maximum digital audio recording level.
If you decide to string together gear with different line level standards, it is up to you to balance those levels across the system. It's called gain staging.

1. Mic plugs into the preamp to bring the mic level signal up to line level. (which ever line level the next thing in the chain expects)
2. Preamp (may) plugs into compressor, eq, etc... the signal is processed and the makeup gain or output level control is used to send a line level signal out to the next thing in the chain.
3. The last thing in the processing chain is set to feed a line level signal into the recorder. Which ever line level the recorder expects to see.

This is recording 101. It's the first thing I teach my interns and the best path toward a good recording.


How you maintane your headroom in your digital mixer or analog mixer applies to how you set your audio levels at playback. So that it does not overload your master stereo buss on your mixer, usually engineers like to keep everything playing back at 0 db on the stereo master buss meter, and always keep the working mix at that level even on the highest audio peaks. That will keep you well within all your audio gear headroom limit.
You do realize that 0db on the analog mixer is not the same as 0db in the digital mixer, don't you? It's two completely different scales. 0dbVU (analog meter) IS line level (whichever line level that device is calibrated to) and generally have 18-24db of headroom for transient peaks. VU meters are very slow and are meant to give you a good idea of the average level, it ignores the peaks and transients.
0dbFS is the end of headroom, there is no more level that can be recorded above that. Digital meters are very fast and only measure the peaks. Line level obviously has to be lower than 0dbfs, because line level is a measurement of average power, so there has to be peaks above that level. Since you can't have signal above 0dbfs, the average power of the signal has to be much less than full scale.

The two types of meters are telling you two different things and are referencing completely different things.


All digital and analog audio gear meters need to be universally reading the same audio signal levels together,
No. They don't because on the analog side of things, the average level is what is important. Preamps and the other circuitry in analog equipment can handle short peaks way above it's nominal operation level (line level), but can't cope well with continuous level much above that level So the average level is the important thing to watch with analog equipment.

Digital is the opposite. It will be completely linear from the noise floor to just under clipping, then completely fall apart as soon as it crosses that threshold. So it is most important that you keep track of the peaks and transients, since those are the things that will ruin the recording.

but not all do so thus again you need to do your own testing/caliberating from your digital recording levels then adjust your other analog audio gear to your digital gear. Usually your computer DAW is accurate on its audio level meters, so you can use it to adjust all audio gear.
Great advice, if you actually understand how it is all meant to work together.


On analog mixers you always have more headroom than digital 0 db would give you, so I have never encountered headroom issues when mxiing with digital audio gear, again as long as I keep all audio levels well within 0 db even on audio level peaks, using a analog and or DAW mixer stereo buss meters.
Again, I think you are confusing your db scales.


But the only time I have run out of headroom on mixing is when using reel to reel analog audio tape, you have to be careful it can exceed 120 db audio levels very quickly !
120db what? The term Db by itself is meaningless. You must refer to the scale you are using.

DbVU= Volume Units. This is the scale that is referenced to line level. 0dbVU = line level
DbFS= Full Scale. This is the scale that is referenced to the end of headroom. 0dbFS= can't go any higher
DbV= volt. This is the scale used to measure voltage referenced to 1 volt. Consumer line level is -10dbV (0.316 volts)
DbU= unloaded. This is the other scale used to measure voltage. It is referenced to 0.775 volts. Pro audio line level is +4dbU (1.228 volts)
DbSPL = sound pressure level This is the scale used to measure acoustic volume. It is referenced to the threshold of hearing.

Good 2" analog reel tape can blow away most digital audio converters when it comes to higher dynamics, higher recording levels and higher frequency response. Because Analog tape is so dynamic it can damage audio gear fairly quickly. I never had digital audio damage gear, although it can happen.
I can't remember off the top of my head, but I'm pretty sure you struggle to get 85db of dynamic range out of 2 inch tape, even with noise reduction. 24 bit recording can record 144db, much greater than any analog circuit could possibly feed it. 16 bit will record 96db of dynamic range.

But by getting all your recordings as hot as possible when you mix on playback you then have the option of raising the volume a bit on the mixer track without adding a huge amount of background hiss and or noise, and without losing any audio quality. Or by lowering the tracks audio level you lose less of its dynamics the hotter the original audio recording is. This applies to 24 bit audio as well. It's a provable audible fact.
It's a complete misunderstanding of gain staging and the science behind all of this.

Get your money's worth out of your gear', that is all I am saying here.
And if you don't like the sound of using a compressor/limiter to get a little more fuller sound and more punch from your audio tracks, then don't.
I use compressors all the time when recording, but that's for the sound of the compression, not to get higher recording levels.


The choice is yours, as always, I am just suggesting some provable ways of getting better audio quality in recording your digital audio tracks. If you don't want to try it then that is Ok too.
Unfortunately, you are just furthering outdated information, old wives-tales and misunderstandings.


Like I said before not every person works with only new gear and 24 bit audio, some use higher and lower bit depths and sampling rates, and or they use older digital audio technologies, so yes getting recordings closer to the digital 0 db audio level whether using new or old digital gear, these concepts very much apply today and tomorrow.
Not really. What applies universally is knowing your equipment, how it's calibrated and how it needs to be interfaced together in order to use it to the best of its ability. Making blanket statements based on old gear and expecting that it is the best way to use the new stuff that was designed to a different standard is just silly.
 
Back
Top