PG Music Home
Posted By: musiclover Digital audio clipping -a few questions? - 06/05/12 10:03 AM
Hello folks,

Thank you very much indeed for all the help and advice you have given me on this forum and on my latest thread on mastering.

I would be grateful if you would give me some pointers on clipping when working with audio.

Can I assume the following is correct?

1 Clipping (above 0db) should not be tolerated in any part of the audio process from start to finished mixdown or master?
This includes clipping in the input audio on all tracks and on all plug-ins on the track and on the master stereo out as well?

2 How do you personally deal with a short burst of clipping on a track? Do you lower the channel fader for the whole duration of the song, or if it’s not too noticeable just lower it for the duration of the clipping part?

3 When dealing with the Ozone mastering plug-in (placed in the master stereo out channel) I find that it may well be showing as clipping in the Ozone plug-in itself, but not as clipping in the Daw master stereo out channel, is this a fairly normal occurrence?

Thanks very much again for all your helpful advice.

Musiclover
For me, I just isolate the portion that clips and use a gain change to drop it 1 db at a time until it stops tripping the clip-o-meter. I know that sometimes there is a random measure somewhere that is at a higher level of attenuation than the rest of the track.
Posted By: Tommyc Re: Digital audio clipping -a few questions? - 06/05/12 12:32 PM
Your mix should not be more than -6 before mastering, that way you the have headroom to master.
Yes, yes and yes.

To avoid clipping, I record in 24-bit for more headroom.
Thank you very much for replies and advice.

If its ok I would like to seek your advice on another problem though really I should probably start a new thread, but I don't want to have a lot of different threads going as it may affect other posters.

I don't have a studio and just use my main computer as my DAW, I have 2 different pairs of speakers hooked up, one logitech 5.1 (though I do have it switched to 2 speakers) for listening to multimedia from the internet, playing cd's etc.

The other pair are an old goodman set I had that I hooked up to my m-audio usb interface and I use those for DAW work etc.(they are sort of behind the monitor due to lack of space, but that is not my main problem.

The problem is how do I calibrate them correctly for playback? I spent hours mixing and then exporting the mixdown but when I opened it in windows media player which plays it thoguh the logitech it was really really bassy, the goodman speakers have 3 controls volume, bass and treble and I have bass and treble set at approx 1 o' clock. Commerical recordings play fine thoguh the logitech, so I take it its the goodmans that are not setup properly, if indeed that is possible with them.

Do I just adjust the treble and bass more on the goodmans so that a commerical recording sounds sort of similar on both sets of speakers?

Do proper monitor speakers have the above controls volume bass and treble and would I get anything for less than say $200 if I do decide to buy?

Thanks very much again for your patience with me.

ps Just an afterhtought on this, since its even possible to edit the eq in the mixer in the computer itself for the logitech speakers this will probably even make them poor references.

Musiclover
Without proper acoustic treatment and a sound level meter, you probably will not be able to 'calibrate your speakers', in the sense you presented it. Mixing involves learning about the sound of your monitors and what the results are of mixing on them, then making the necessary adjustments.

Two quick mixing tricks:

1. Test your tentative mixes on everything you can find, including car speakers, boom boxes, cheap computer speakers, home stereo etc.

2. Try, when mixing, reducing the volume to just barely audible. Does any instrument stand out? That may be a problem (even for vocals). Can you hear every instrument? You should be able to.
Quote:

2. Try, when mixing, reducing the volume to just barely audible. Does any instrument stand out? That may be a problem (even for vocals). Can you hear every instrument? You should be able to.




I agree with the first part of this but not the second, Matt. I was just discussing this with someone over the weekend and we went for a drive in my car to test it. You do not hear every instrument at very low volumes. My Benz is a nice quiet car with the Bose sound system but it's still a noisy environment compared to a studio. We spent a good half hour driving around and listening to everything from country to jazz to heavy metal to classic hiphop. When we turned the radio down to where we could barely hear anything what's left is always the lead, either vocals or instrument, and drums. That's it 95% of the time (there's always a few exceptions) and of those two vocals or the lead instrument like your sax are always on top of everything. I've used this as a rough test of the mix for years. Of course this is just rough, it has little to do with panning, EQ, compression, all of that stuff. It's just that when I listen to songs people post here and elsewhere most of the time the lead is buried. The basic mix is way off and my quick down and dirty test will at least put someone on the right track.

My guess is amateurs have no confidence in their singing or playing so they don't mix themselves properly. My feeling is if you're going to do it, then do it warts and all. If you don't then the whole production sounds like amateur hour not just the singing. I'm occasionally guilty of that too btw. I have some live recordings where my piano solo had a few too many clams in it so I buried it a little. One I was able to cover by pasting in a short phrase I did earlier and that little repeat sounded much better than the totally awful chord I played. When I played it back you would never tell I did that, it just sounds like I repeated a lick. Whatever, at least I try to do it right and also improve my playing.

Bob
Quote:

To avoid clipping, I record in 24-bit for more headroom.




I'm not sure that 24 bit versus 16 bit recording will give you any more 'headroom.' What it WILL give you is more dynamic range. From Rane's Professional Reference Manual

Quote:

headroom A term related to dynamic range, used to express in dB, the level between the typical operating level and the maximum operating level (onset of clipping). For example, a nominal +4 dBu system that clips at +20 dBu has 16 dB of headroom. Because it is a pure ratio, there are no units or reference-level associated with headroom -- just "dB." Therefore (and a point of confusion for many) headroom expressed in dB accurately refers to both voltage and power. Which means our example has 16 dB of voltage headroom, as well as 16 dB of power headroom. It's not obvious, but it's true. (The math is left to the reader.)




If my understanding of this is correct (and I've been wrong before) if you record at -12dBr (referenced to 0dBFS), then you should have 12dB of headroom. If you're using 16 bits, then you have a dynamic range of 96dBr versus 24 bits which gives you 144dBr of dynamic range. However, if you're still recording at -12dBR, you still only have 12dB of headroom, 16 or 24 bit.

If anyone disagrees, please discuss civilly.

Gary
In response to the tweaking speakers, I don't know if this will work or not, but it's worth trying.

Create a series of individual test tones, between 20Hz and 20kHz. You should be make them all the same level. Position a GOOD microphone at your mixing position, and then record each tone via the microphone to an individual track. Make sure the speakers are set to neutral settings, i.e. no bass or treble gain or cut, and are correctly positioned for mixing.

Use a Real Time Analyzer plug in to play back each tone recording. Ideally, the RTA should indicate what the original test tone was sent out. If they aren't, then you can attempt to adjust the speakers bass and treble, although with only two controls this will be difficult.

Finally, mix the signals into a single file and then play it back. Record another track of all the signals being played back through your now tweaked speakers. You should be able to see a broad spectrum analysis of what your speakers are sending out.

From there, you should be able to use either a 10 band or parametric EQ to get closer to a flat curve. Save that EQ curve and use it in the master section to correct the overall tonal quality of your speakers. It won't be perfect, far from it, but it should bring you closer to a flat curve.

Gary
The main argument for 24 bit is more digital room for the audio, more dynamics, less need for compression while recording and more "headroom". Every article I have read about 24 bit vs. 16 bit recording uses the word "headroom" about every other sentence.

I guess you have to agree of a definition of "headroom". I think the general definition is "how much before clipping" -- 24 bit has more.
Kevin,
There is 144dB of Dynamic range in a 24 bit recording. There is 96dB of DR in a 16 bit recording.

So, what does this actually mean? With a 16 bit recording, if you could find an absolutely quiet space, you have 96dB of range before running out of range. A piano being played fortissimo is about 95dB, and is considered 'very noisy.' A full symphony orchestra is 110dB! Does that mean that you can't record a full symphony orchestra with a 16 bit recording? No, of course not. What it does mean though is that it's going to be about 15dB above dead silence. A rock band at a live concert is about 130dB, which means that the bottom of your dynamic range is going to be about 35-40dB, not dead silence.

A true 24 bit recording, and you're never going to have a TRUE 24 bit recording, because you run into electron flow noise down in the mid 130's, would allow you to go from dead quiet to well past the pain level. Imagine standing next to a 747 jet taking off, you'll understand 140dB!

This is the true difference between 16 and 24 bit. Headroom is measured from clipping down, dynamic range is measured from dead silence up.

Gary
Hi Gary. These are very complicated topics for a forum such as this, and I'm trying to keep my answers short and perhaps oversimplified for Musiclover, who appears to be someone just getting started in audio.

I agree you're not going to hear everything even in the quietest of cars, and good for you that you have one. I did not mean to mix for a car, but to compare the sound in all the places you can listen, and go back and make small adjustments in the mix. What the OP should be looking for is a glaring mismatch in levels. You can't fix everything, though; bass solos, for example, are going to disappear in a car (mine, anyway) no matter what.

The best most of us can do is to learn how our monitors and our listening environment for mixing are not perfect, and we adapt. Let's all help the OP to get close without burying him in details that he/she will pick up in time. To that end, I will be guided by the level of the question from the OP.

You made a great point that you might as well put it all out there, warts and all, since hiding in the mix doesn't work. There's a funny message circulating now on Facebook to the effect that Jimi Hendrix remarked that people were copying his playing, even his mistakes. Gotta love it.
Matt,
Understood.

Kevin, something else that I do, and works for me and my style of recording. I record synth tracks only, no vocals and no acoustic instruments. Hence, I can go back and adjust as necessary, and record the track as many times as I want with the same MIDI sameness.

First, I record each of my tracks separately, one at a time. I try to find the peak of the track, and have that peak somewhere between 0 and -3dBFS or -3dBr. That assumes that 0dBr is the same as full scale and clipping. I record each track so that each is in that same range. What this does for me is to give me the fullest dynamic range of the instrument, so when there is quiet, I'm not getting hiss from the lower levels and running into the noise floor. Synths are inherently noisy creatures, it seems.

After recording all the tracks, I adjust the relative volumes of the tracks to what I consider to be pleasing. I have an arranger keyboard, which means play a chord, it generates the backing sequence, you play the melody on top. Since each of the backing tracks are at a different volume, you need to play with your mixer to get that relative volume setting back. So, I go through and I set all the audio tracks to give me what I'm looking for.

If I do it right, my main mix output should be just below 0dBr. There is no headroom. Now, what I'll do is to group all the audio together as a single group, and then lower the peak output to -3 or -6, or even -12dBr. At this point, I have not applied any effects whatsoever.

Once I pull all the channels down by the amount I think I'm going to need, then I start applying the effects I want, EQ, reverb, compression or whatever I'm using. I keep an eye on the main outs to make sure that I'm not exceeding 0dBr and go into clipping. If I do that, I need to remove all the effects and lower all the sliders some more, to give me more headroom for mixing and effects.

In the end, I should get a mix that has very good dynamic range (provided I don't mash it with compression), all the instruments should be clear and distinct, and I should have the peaks at between 0dBr and -3dBr.

As to applying this to a live recording, I've never tried it, I'm not sure how it would work, but I'd be willing to bet Bob Harvey would have some good insight into this.

Gary
Gary: I only record in 16-bit -- but my next set of recordings will try to do so in 24-bit. Of course I only have 16-bit BIAB wma files -- but I am not "recording" those.

Quote:

...have that peak somewhere between 0 and -3dBFS or -3dBr... ... If I do it right, my main mix output should be just below 0dBr. There is no headroom. Now, what I'll do is to group all the audio together as a single group, and then lower the peak output to -3 or -6, or even -12dBr. At this point, I have not applied any effects whatsoever. ...



I do the same thing, but I have been reading (can't remember where) that there is some sonic advantage to recording the instruments so you don't have to group them all and reduce the gain by 6-10 db. Something like reducing the gain increases the noise-signal ratio. I could be wrong, but that seemed to be another one of the arguments for recording in 24-bit.
Kevin,
Reducing the gain increase the SNR?

Well, I can see how you might think that, and I could also see how you would get there, but I don't think that's true for most cases.

If you reduced all the tracks by 50dB, and then increased the main outs by 50dB to bring the signal up to 0dBFS, then yeah, you're probably decreasing the SNR. But, you aren't. You're decreasing the signal strength enough to allow you the headroom for your effects, which still should bring the individual tracks up to an aggregate output of 0dBFS. If you effects require so much gain that you need to reduce the track level by 50dB, that's going to sound really weird, and probably not good. At least, not for the kind of music I record.

Mac, or some of the other more knowledgeable people may chime in and correct me.

Gary
I feel like I need to step in here and correct some misinformation that is being passed along.

24 bit does not give you any more dynamic range than 16 bit. Dynamic range is determined by the analog side of the signal processing, not by the converted to digital format.

What 24 bit recording does do is give you a better signal to quanitization noise ratio than lower bit depths.

For every bit depth there is an additional 6 dB of SQNR. You can go off and google and wikipedia this to get the complex math but I'll try to keep it simple here.

Here's the way to think about this. Let's pretend you have a pure sine-wave at a fixed peak-to-peak voltage. Let's pretend that we take this signal and we amplify it so that the the peak-to-peak voltage is the same as the peak-to-peak voltage that the A/D converter can handle. You can quantize this into 16 bits, or 2^16 values, or one of 65,536 possible values. The sine wave will have very tiny little stairstepped voltage values into which the wave is encoded.

Now, let's take the same signal, and quantize it into a 24 bit A/D converter, which has one of 16,777,216 possible values. As a result, the stair-steps are WAY smaller with 24 bit A/D conversion than with 16 bit conversion.

The 'stair-steppy-ness' of the signal is the quantization noise. Yet another way to think about this is to think of the smallest possible signal that can be encoded, a signal so quiet that the A/D converter switches between the lowest possible value at zero and the next highest value. It switches back and forth between 0000000000000000 and 0000000000000001. When those values eventually get sent back to D/A conversion, that twitching between the two values tweaks the output D/A and generates an unintended analog output noise: Quantization noise. (For those of you that know binary formats, and the difference between signed and unsigned stuff - bear with me here, just trying to make a point)

For a 24 bit recording, if we adjust the signal down so that the switching back and forth occurs between 000000000000000000000000 and 000000000000000000000001, the lowest two values, and as it goes back to D/A the tweak is much smaller and hence the quantization noise is much smaller. In fact the quantization noise will be roughly 48 dB lower than the 16 bit quantization noise.

Note - no magic with the dynamic range occurs.

The big benefit to recording with 24 bit recording over 16 bit is that one doesn't really have to worry nearly as much about using the full dynamic range of the A/D converter in order to get a nice signal to quantization noise ratio, and the little stair steps that occur in the digitized data are peanuts in comparison to 16 bit.

In other words, you can be quite a bit less careful about it, and just get on with recording.

Bottom line, real dynamic range has nothing to do with bit depth.

Other bottom line, disk space is nearly free these days and switching from 16 to 24 bit doesn't have a hard and fast noticeable cost these days. It's not even computationally expensive any longer with software that takes advantage of multiple computing cores.

If your input A/D on your audio I/O device has 24 bit capability, switch over to it if you haven't already done so.

Back to the OP's original question....
Thanks very much indeed for all your very helpful adivce and tips. If I only take in a fraction of the advice that has been offered it has been well worth it.

I will save this thread on my computer and then refer back to it, when I need to.

Thanks again

Musiclover
Posted By: rharv Re: Digital audio clipping -a few questions? - 06/07/12 10:53 PM
I was gonna post similar to Scott, but didn't have the time last night.
To simplify it; 24 bit gives more steps between zero input (none) and the clipping point but it doesn't 'prevent' an already clipping signal... although it can be more forgiving when you push the barrier.
It's really 'larger' chunks of data, which can buffer the impact a bit (compared to a smaller chunk that spikes at the right spot).

Like Scott said, it is preferable to use it when available.

What really bamboozled me when I first made to trip from analogue recording to digital is the folks who insist on using analog audio terms to describe digital recording and processing. They just don't work.

There are no real analogies to bit depth and sample rate in analog audio. They need to be understood without some ancient tape jockey babbling about headroom and frequency response.

The way I look at it, packages of digital audio information are processed to make music. Greater bit depth means the packages are larger. A higher sample rate means the packages are being delivered more frequently.

Clipping and digital distortion are not products of either bit depth or sample rate. Keep your track peaks below -0.25 and they are not going to make the nasty noises - it's that easy.
Posted By: ROG Re: Digital audio clipping -a few questions? - 06/08/12 08:51 AM
Quote:

They need to be understood without some ancient tape jockey babbling about headroom and frequency response.




Hi, all. I think I just recognized myself, there!

Actually, I completely agree with the 16/24 argument, but I think that, sometimes, we assume that better equipment makes better recordings, without acknowledging the difference made by the skill of the operator.

In my view, the differences between 16/24 are hard to detect, whilst the skill of the operator makes a huge difference. Some of us who remember working with tape, when you had a noise floor you could measure on the VUs and nothing was tracked without a compressor to preserve the s/n ratio without switching in the Dolbys, are quite happy to stick with 16 bit.

Just sayin....

ROG.
Quote:

Quote:

They need to be understood without some ancient tape jockey babbling about headroom and frequency response.




Hi, all. I think I just recognized myself, there!

Actually, I completely agree with the 16/24 argument, but I think that, sometimes, we assume that better equipment makes better recordings, without acknowledging the difference made by the skill of the operator.

In my view, the differences between 16/24 are hard to detect, whilst the skill of the operator makes a huge difference. Some of us who remember working with tape, when you had a noise floor you could measure on the VUs and nothing was tracked without a compressor to preserve the s/n ratio without switching in the Dolbys, are quite happy to stick with 16 bit.

Just sayin....

ROG.




I've worked in both worlds, tape and digital - in many formats. I don't have measurement equipment at home that is precise enough to show the difference between 16 and 24 bit. What I do know is that since switching over to 24 bit, I've not had to wring my hands about trying to maximize the input range of my A/D for every track. I can be quite sloppy with it, and when I then go back and boost those low-level recorded signals, they are still doggoned clean sounding. This is not the case when doing the same practice with 16 bit. I can indeed start to hear hiss if I take a recording that I made perhaps 25 dB down in 16 bit, then boost it up 20 dB. Try that experiment for yourself.

-Scott
Posted By: 90 dB Re: Digital audio clipping -a few questions? - 06/08/12 01:39 PM
Quote:

Quote:

They need to be understood without some ancient tape jockey babbling about headroom and frequency response.




Hi, all. I think I just recognized myself, there!

Actually, I completely agree with the 16/24 argument, but I think that, sometimes, we assume that better equipment makes better recordings, without acknowledging the difference made by the skill of the operator.

In my view, the differences between 16/24 are hard to detect, whilst the skill of the operator makes a huge difference. Some of us who remember working with tape, when you had a noise floor you could measure on the VUs and nothing was tracked without a compressor to preserve the s/n ratio without switching in the Dolbys, are quite happy to stick with 16 bit.

Just sayin....

ROG.








I always find these 16/24 discussions amusing. I started out with a Tascam Portastudio 144 in 1979 (4-track/cassette). You want to talk about noise floor?


I agree with ROG - the gear will not make you a better operator. It will, however, remove a lot of the sonic obstacles that plagued the "good old days".



Regards,


Bob
And don't forget that 0 on your digital meters is about 18dB hotter then 0 on an analog desk.
Posted By: ROG Re: Digital audio clipping -a few questions? - 06/08/12 05:32 PM
Quote:

I can indeed start to hear hiss if I take a recording that I made perhaps 25 dB down in 16 bit, then boost it up 20 dB. Try that experiment for yourself.




Scott.

The first thing I ask myself is - why would someone with your obvious knowledge and experience want to record anything 25db down?

One thing you must remember is that you have a theoretical dynamic range of your digital recording medium, but an actual one based on the noise floor from the inherent noise in the signal chain. If, for example, you put a cheap mic in front of an old Marshall tube amp, your dynamic range is, I think the technical term is "shot-to-hell". Even just a good mic through an expensive pre-amp will have a noise floor higher than the digital medium. The experienced engineer will will know how to minimize these problems and get the best results out of the available gear.

You'll notice that I agreed with the main 16/24 argument in my earlier post. I was just making the point that we encourage people to up-grade more often than we tell them to get some specialized training.

ROG.
I remember my home studio. I would mix down to a DAT Recorder. This was around 1990? I'd send the audio synced with midi from my Atari ST right to the DAT machine. It took some time to get used to setting levels. Many of my first mixes were full of distortion. What an ugly sound it was. I remember it sounding like blown speakers. I started to use a compressor/limiter to help me keep it even.

Wayne,
Posted By: Mac Re: Digital audio clipping -a few questions? - 06/08/12 06:29 PM
Quote:

I remember my home studio. I would mix down to a DAT Recorder. This was around 1990? I'd send the audio synced with midi from my Atari ST right to the DAT machine. It took some time to get used to setting levels. Many of my first mixes were full of distortion. What an ugly sound it was. I remember it sounding like blown speakers. I started to use a compressor/limiter to help me keep it even.

Wayne,




That was the old 16 bit Audio Engine.

Today, all audio recording programs are using the 32 bit engine. No more digital "thwack" from overruns, which is another way of saying digital clipping.


--Mac
Quote:

... I think I just recognized myself, there!

Actually, I completely agree with the 16/24 argument, but I think that, sometimes, we assume that better equipment makes better recordings, without acknowledging the difference made by the skill of the operator.

In my view, the differences between 16/24 are hard to detect, whilst the skill of the operator makes a huge difference. Some of us who remember working with tape, when you had a noise floor you could measure on the VUs and nothing was tracked without a compressor to preserve the s/n ratio without switching in the Dolbys, are quite happy to stick with 16 bit.





Rog,

Hoping that the analog chops we worked so hard to acquire could be translated to digital audio processing is common to most of us oldsters.
______________________________________________________________

In Linux, audio information is transmitted from one application to another for special-purpose processing, and the bit depth of the transmitted files can be 32 bit, or 64 bit. Files can even be encoded as 32 or 64 bit floating point.
I had not grasped the significance of this until reading Scott's description of "quantization noise" :

"What 24 bit recording does do is give you a better signal to quanitization noise ratio than lower bit depths.

For every bit depth there is an additional 6 dB of SQNR. You can go off and google and wikipedia this to get the complex math but I'll try to keep it simple here.

Here's the way to think about this. Let's pretend you have a pure sine-wave at a fixed peak-to-peak voltage. Let's pretend that we take this signal and we amplify it so that the the peak-to-peak voltage is the same as the peak-to-peak voltage that the A/D converter can handle. You can quantize this into 16 bits, or 2^16 values, or one of 65,536 possible values. The sine wave will have very tiny little stairstepped voltage values into which the wave is encoded.

Now, let's take the same signal, and quantize it into a 24 bit A/D converter, which has one of 16,777,216 possible values. As a result, the stair-steps are WAY smaller with 24 bit A/D conversion than with 16 bit conversion.

The 'stair-steppy-ness' of the signal is the quantization noise. Yet another way to think about this is to think of the smallest possible signal that can be encoded, a signal so quiet that the A/D converter switches between the lowest possible value at zero and the next highest value. It switches back and forth between 0000000000000000 and 0000000000000001. When those values eventually get sent back to D/A conversion, that twitching between the two values tweaks the output D/A and generates an unintended analog output noise: Quantization noise. (For those of you that know binary formats, and the difference between signed and unsigned stuff - bear with me here, just trying to make a point)

For a 24 bit recording, if we adjust the signal down so that the switching back and forth occurs between 000000000000000000000000 and 000000000000000000000001, the lowest two values, and as it goes back to D/A the tweak is much smaller and hence the quantization noise is much smaller. In fact the quantization noise will be roughly 48 dB lower than the 16 bit quantization noise.

Note - no magic with the dynamic range occurs.

The big benefit to recording with 24 bit recording over 16 bit is that one doesn't really have to worry nearly as much about using the full dynamic range of the A/D converter in order to get a nice signal to quantization noise ratio, and the little stair steps that occur in the digitized data are peanuts in comparison to 16 bit.

In other words, you can be quite a bit less careful about it, and just get on with recording."


Thank you, Scott!
Posted By: ROG Re: Digital audio clipping -a few questions? - 06/08/12 06:49 PM
Quote:

In other words, you can be quite a bit less careful about it, and just get on with recording.




Oren.

You don't think that good practice once learnt, is something which becomes second nature? Once we start to tell people it's ok to be careless about one thing, we then have to specify HOW careless and to differentiate between that and the things that DO need attention.

Start of the slippery slope?

ROG.
Quote:

Quote:

I can indeed start to hear hiss if I take a recording that I made perhaps 25 dB down in 16 bit, then boost it up 20 dB. Try that experiment for yourself.




Scott.

The first thing I ask myself is - why would someone with your obvious knowledge and experience want to record anything 25db down?

One thing you must remember is that you have a theoretical dynamic range of your digital recording medium, but an actual one based on the noise floor from the inherent noise in the signal chain. If, for example, you put a cheap mic in front of an old Marshall tube amp, your dynamic range is, I think the technical term is "shot-to-hell". Even just a good mic through an expensive pre-amp will have a noise floor higher than the digital medium. The experienced engineer will will know how to minimize these problems and get the best results out of the available gear.

You'll notice that I agreed with the main 16/24 argument in my earlier post. I was just making the point that we encourage people to up-grade more often than we tell them to get some specialized training.

ROG.




Rog,

It's not that I make recording at low levels a matter of practice, but allowing plenty of room does make for less hassles during mixing, not having to apply attenuation in the box to each track, so that when I sum the tracks, I'm not summing to something that's overdriving.

It's just a matter of convenience, I suppose. When I think back to when I had only 16 bit available, I spent a lot of time worrying about maximizing the A/D for each recorded track, and then dialing down from there inside the box.

I do almost no worry of that now and just get on with recording the next track and on to mixing. It's been a very rare occurrence when my songs have summed to the point where I have to go in and adjust gains of individual tracks down - and my 2 channel mixdowns sound very clean.

That wasn't the case in my previous 16 bit life.

I will say that my mixdown 'process' is much less structured than most here, as I don't really have a mixer paradigm in the DAW that I use and I'm fine with it being that way.

Hopefully this is making sense.

-Scott
This has been a very informative thread.

Thanks guys.
Posted By: ROG Re: Digital audio clipping -a few questions? - 06/09/12 09:29 AM
Scott.

Forgive me if I appeared to be critical of your recording technique. In threads like this, I tend to court controversy, which is a character flaw I find difficult to throw off!

At the end of the day. it's the final mix which we're judged on, not the way we got there.

I learnt my recording in the 60s and 70s, when attention to detail such as levels throughout the signal chain was essential, bearing in mind the equipment we had. Old habits die hard. I have to say, though, there's no way I would want to go back to tape and enormous mixers with no automation.

BTW, if you like maths, which you seem to have a good grasp of, try factoring in the fact that quantization noise is frequency sensitive - it gets scary!

ROG.
Rog,

No harm no foul. Your question was valid. Please understand that I still care about gain staging and so forth and I do still use most of the A/D available range when I record - I just worry less about it now. It used to make a difference. Now, with the limited time I have for this hobby, I spend less time worrying about it and it's none for the worse as a result.

Regarding maths, I learned this stuff about 10-12 years ago in the EE-638 Digital Signal Processing class at Purdue University. I was the only mechanical engineer in the course. The professor was Michael Zoltowski - great guy who knows a shed load of information about DSP. His specialty is spread spectrum communications. I bugged him a great deal with my questions about applications to audio for music purposes.

My favorite moment in the class was the exercise we went through to show how sending this stair-stepped digitized signal through the D/A filter actually DOES result in a smoothed analog signal (up to a certain frequency) through the superposition of each bit's impulse response through the D/A filter stage.

That was an eye opener for me.
Quote:


You don't think that good practice once learnt, is something which becomes second nature? Once we start to tell people it's ok to be careless about one thing, we then have to specify HOW careless and to differentiate between that and the things that DO need attention.
Start of the slippery slope?







ROG,

Music as art has intrigued me, and the science of music is fascinating, but the craft of making music - from composition to performance to recording to post production - is what really puts the frost on my pumpkin.
In that context, the care with which a craftsperson approaches an audio project is paramount; the defining quality of the experience, for me. (not that you'ld automatically know that from some of the doo-doo I've generated - but we're all entitled to our mistakes... )

So "yes" to good practice as second nature, and "no" to slippery slopes, analog or digital!
© PG Music Forums