Does tracking hot improve presence??

[JPH]

Hey Slutz,

I've been reading a lot about tracking levels and I am curious if tracking at a lower level (-18dbfs) has an effect on a track's presence. This may seem like a stupid question and I probably am just confusing myself because of the tracks loudness compared to something tracked at -3 dbfs.

Let me know what you guys think.

[staudio]

Don't track hot. Peaks should be no higher than -6dBFS, usually about -10dBFS. Tracking hot increases noise and distortion depending on the system and equipment used. There is no benefit to recording individual tracks close to digital 0. When you leave a little room at the top you don't have to worry about clipping and the sound is better in my opinion. The only time someone should record close to 0 is when mastering, and even then not always.

Good luck.

[eddierodriguez]

tracking hot doesn´t improve presence and it´s no good idea, but a cranked preamp can improve it.

some preamps get their soul with much gain, used with enough pads you can definitely increase the presence.

[Lagerfeldt]

Originally Posted by JPH

I've been reading a lot about tracking levels and I am curious if tracking at a lower level (-18dbfs) has an effect on a track's presence. This may seem like a stupid question and I probably am just confusing myself because of the tracks loudness compared to something tracked at -3 dbfs..

If you're talking purely digital levels, then generally no. It's a fact that some A/D converters will start distorting when you're getting close to 0 dBFS. But most will be clean all the way up to at least -3 dBFS.

Same goes for the D/A, as many will start distorting on the way out too, especially if your signals have been exposed to hard digital limiting or clipping. As the waveform leaves the D/A it will naturally try to recreate the missing the peaks which leads to 0 dBFS+ signals, often distorting.

A rule of thumb is to record in 24 bits and stay below -3 dBFS all the way thru the digital chain.

Sometimes this A/D distortion will make things appear extra loud because distortion almost always has this effect. It can also make a signal appear wider if the distortion varies between the L/R channels.

However, this kind of distortion is not a very nice kind of distortion, and when processed later it will start sounding even more ugly.

As Eddierodriguez mentioned above, some analog gear will sound nice when pushed hard, both tubes and solid state equipment (two different sounds though). That's something else, and you should certainly experiment with that.

[brethes]

Originally Posted by JPH

Hey Slutz,

I've been reading a lot about tracking levels and I am curious if tracking at a lower level (-18dbfs) has an effect on a track's presence. This may seem like a stupid question and I probably am just confusing myself because of the tracks loudness compared to something tracked at -3 dbfs.

Let me know what you guys think.

In a digital Worstation working at 24bit, there is no reason to track hot as this will lead to distortion (of the nasty kind) in several ways. this may not be too noticeable on a single track but will be cumulative and will eventually lead to overloading/clipping effects and Master Buss, causing even more harm at the mixing stage.

This is the kind of problem that makes mastering engineers pull their hair in exasperation as while it is easy to make a clean (digital distortion free) track much louder and powerful, a squashed/distorted track will severely limit what can be done at the mastering stage. Adding the right kind of distortion to a mix to make sound "bigger" is no problem, but removing or even reducing digital distortion never works successfully.

[synthoid]

Originally Posted by Lagerfeldt

It's a fact that some A/D converters will start distorting when you're getting close to 0 dBFS.

Can you name some A/D converters that have this property? I don't necessarily advocate driving your A/D converter all the way to the 0dBFS edge, but I sure would be unhappy to buy a converter that distorted on the analog side at any signal below 0dBFS. And I can't understand why there would possibly be any need for distortion / a complete lack of headroom in the input stages of such a circuit.

-synthoid

[narcoman]

Originally Posted by synthoid

Can you name some A/D converters that have this property? I don't necessarily advocate driving your A/D converter all the way to the 0dBFS edge, but I sure would be unhappy to buy a converter that distorted on the analog side at any signal below 0dBFS. And I can't understand why there would possibly be any need for distortion / a complete lack of headroom in the input stages of such a circuit.

-synthoid

many of the cheaper ADs have very little headroom - and that's the problem.

[Lagerfeldt]

Originally Posted by synthoid

Can you name some A/D converters that have this property? I don't necessarily advocate driving your A/D converter all the way to the 0dBFS edge, but I sure would be unhappy to buy a converter that distorted on the analog side at any signal below 0dBFS. And I can't understand why there would possibly be any need for distortion / a complete lack of headroom in the input stages of such a circuit.

-synthoid

Several converters distort on the A/D side when getting close to 0 dBFS. Many brands even show a warning LED when exceeding - 3 dBFS, such as RME.

You can test this easily by sending material into a lower quality A/D and getting near to or clipping and listening to (and looking at) the resulting waveform.

A HQ A/D will have no jagged lines and will clip straight while a lower quality A/D will have the jaggies, even when not clipping.

It's very easy to hear (and visually confirm) if you're using a sine wave.

And most D/A's, especially in normal CD players will distort with squashed material due to 0 dBFS+ signals. That's more expected though.

[theblue1]

A practical rule of thumb: as long as you keep the noise floor of your analog input signal above the noise floor of your digital converters, you shouldn't be losing any sound quality.


The problem with pushing everything close to 0 dBfs is that it's possible to have such a thing as intersample peaks created on interpolation of near 0 dBfs signals. None of the samples (of course) exceed digital zero but due to the tech voodoo of how digital signals play back, interpolation of some signals can actually create 'overs' in the DA stage, resulting in (sometimes unnoticeable) digital distortion.

That's why it's good practice to either use an oversampling, intersample peak-aware meter plug in (like SSL's X-ISM intersample peak meter) or to simply give yourself some extra headroom. Some folks say 3 to as little as 0.3 dB should be sufficient with quality, modern converters. Others say something like 6-9 dB of extra headroom is necessary to be absolutely safe.


The best option is to stage your gain optimally in your critical analog front end. Since your (or any) analog front end has a lot less signal-to-noise ratio than that afforded by a good 24 bit AD, you've got a lot headroom to work inside. As long as the signal to noise ratio of the signal you're sending to your AD is less than the real world S/N ratio of your AD, you should not be losing anything.

An example: if you have an analog input signal with a S/N ratio of 80 dB, and your converter has, say, a practical S/N ratio of 110 (the theoretical 24 bit maximum is 144 but the best devices in use today afford about 115 dB), that means, for our purposes, that you could record with peaks as low as almost 30 dB below 0 dBfs without practical impact on the signal. (I'll leave it to others to drill down to the fine points.)

[James Meeker]

Personally I think tracking very "hot" in the digital world results in a pinched, shrill sound.

[theblue1]

Could happen... besides the intersample peak issue, there's also plug ins to consider.

Inside most modern DAW's using 32 bit (or better) floating point math, you have the basic ability to go over 0 dB internally -- as long as the signal you output to your DA is safely under 0 dBfs.


BUT... but... your plug ins (particularly dynamics-dependent virtual devices like compressor/limiters) may well have a 'sweet spot' range outside of which they may deliver less than optimal results.

[synthoid]

Originally Posted by Lagerfeldt

Several converters distort on the A/D side when getting close to 0 dBFS. Many brands even show a warning LED when exceeding - 3 dBFS, such as RME.

You can test this easily by sending material into a lower quality A/D and getting near to or clipping and listening to (and looking at) the resulting waveform.

Well, the warning LED has nothing to do with distortion of the waveform of course. It is simply another kind of peak level meter.

I don't see any evidence that the RME converters change their behavior as they approach 0dbFS. And it would make no sense to me that they would, because their input sensitivity can be calibrated. In other words, if they distorted near 0dBFS when calibrated to -10dBV = 0dBFS, then you might reasonably expect that they would go into wild distortion when calibrated to +4dBU = 0dBFS or +19dBU = 0dBFS, but none of these things happens.

The other converters I have handy are calibrated very high, so I haven't had the occasion to see how they behave when pushed to the calibration point. But they are all better quality converters than the RME, so I kind of doubt they would distort.

I'm beginning to think that this is largely urban legend.

I understand about intersample peaks, but that is a consideration that applies when the analog signal is regenerated by the D/A converter -- in other words, it applies when you send the signal back out, not when you convert it on the input side. And anyway, I'm talking about a case in which the analog input signal lies entirely and safely below the 0dBFS calibration point; it's in this case that I see no reason (and no evidence of) distortion in any of the converters I have access to.

-synthoid

[Lagerfeldt]

When you say "you don't see any evidence" I submit that you have not actually tested this.

Originally Posted by synthoid

Well, the warning LED has nothing to do with distortion of the waveform of course. It is simply another kind of peak level meter.

Actually, it does.

This is the warning sign that shows at which point the converters will start behaving less linear and eventually add distortion when getting very close to 0 dBFS on the RME. Same goes for several other converters (peak meter at -3 dFBS or not).

There's nothing wrong with stating an opinion but you're stating things like they are facts, which your statements are not.

You're simply adding to the confusion this way, so I suggest people looking for advice on levels read my first post in this thread.

[Chaellus]

tracking hot is something best left to analog tape for satuaration chacterisitcs from pushing hard, in digital there is no headroom so it just distorts and not in the pleasent way

[synthoid]

Originally Posted by Lagerfeldt

When you say "you don't see any evidence" I submit that you have not actually tested this.

But I have tested it, as best I am able to. Which is to say, I've experimented with a number of different tracking levels on the RME, and several different calibration settings. Ditto for other converters I have handy.

It's not very easy to test this in a rigorous way without high quality signal generators to give you waveforms at several different amplitudes that you are certain are otherwise identical. Let me think about how I can rig up a credible and repeatable test for this and I'll give it a go if I figure out how (suggestions welcome). If I had some inline pads, I could record a signal at one amplitude, then pad it down and record it at a lower amplitude, for example. That would work if we rule out distortion in the pad. But I don't have any such pads ...

Or, I could record a signal near to 0dBFS with the unit calibrated to -10dBV, then record the same signal with the unit calibrated to +4dBU. What would you think of that test? It could be argued to be apples to oranges, but it has the advantage of allowing an identical input signal to be used for the two readings.

Anyway we do it, it's going to be tough to determine whether two digital waveforms of different amplitude are identical (they will be out of phase of course).

If anyone from RME (or another converter manufacturer) is reading, I'd love to hear their thoughts about this. I don't see anything in their literature that indicates that the units distort on signals above -3dBFS.

-synthoid

[Lagerfeldt]

What RME converter did you test with?

I have done tests with RME ADI-8 DS and QS, MOTU 828 Mk2, Apogee Rosetta 800, Benchmark ADC-1, Lavry Blue, and DAD AX24.

It's quite evident that some converters handle hot signals better than others, same goes for clipping.

If I have the time Monday I can perhaps show or play the differences for you.

Originally Posted by synthoid

I don't see anything in their literature that indicates that the units distort on signals above -3dBFS.

-synthoid

Use your ears and eyes. It's audible and it's visible. RME would never write (=admit) that distortion is an issue above -3 dBFS, and to be fair, RME converters are not the only ones suffering from this problem. It has improved in the QS series though, did you test that one?

This is in no way a major criticism of RME, I think RME in general offers excellent sound cards and converters, especially if you consider the price. I'm just using them as an example of the near 0 dBFS problem in A/Ds.

[JPH]

thanks a lot for all the replies. i just did a session this weekend and tracked everything with peaks around -9. the guitars sound MUCH cleaner than when i pushed the pres in the past.

[Drlove]

I've found the best gauge is always your ears.