The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
My Theory About Prosumer Audio Audio Interfaces
Old 11th June 2006
  #31
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by Dave Derr
AND in actuallity, if you let the peaks only hit -12dBfs, you ARE losing 2 of your best bits of resolution and dynamic range. Your most innaccurate bits are the last bits, and the best bits are the first bits. Algebraically, it only takes 1 bit to lose half your dynamic range (-6dB is one half the voltage amplitude).
If it were an analog recorder with 144dB of dynamic range, would you say that 6dB is half the dynamic range?

I think it's unforunate that A/D converters have an interface whereby they deliver 0dbFS at full scale input. Suppose they have 19 bits of usable output. Wouldn't it be better if they delivered some value (say 001111...11) at their rated peak voltage, but continued to give less accurate results for greater input voltages? In other words, give some headroom on the analog side of the converter. Actually, maybe they are already built this way; that is, maybe they are most accurate at something other than full-scale analog input, I don't know. In any case, it would be a lot better if we aimed for the sweet spot in their range rather than trying to push as close as possible to maximum input voltage, which is begging for clipping on the ouptut.

-synthoid
Old 11th June 2006
  #32
Lives for gear
 

Quote:
Originally Posted by synthoid
I think it's unforunate that A/D converters have an interface whereby they deliver 0dbFS at full scale input. Suppose they have 19 bits of usable output. Wouldn't it be better if they delivered some value (say 001111...11) at their rated peak voltage, but continued to give less accurate results for greater input voltages? In other words, give some headroom on the analog side of the converter. Actually, maybe they are already built this way; that is, maybe they are most accurate at something other than full-scale analog input, I don't know. In any case, it would be a lot better if we aimed for the sweet spot in their range rather than trying to push as close as possible to maximum input voltage, which is begging for clipping on the ouptut.

-synthoid
that is simply the nature of digital, you can not design a converter differently.
Old 11th June 2006
  #33
Gear Nut
 

Quote:
Originally Posted by Silver Sonya
I think I have come to the conclusion that very few people understand the concept of gain structure or headroom within DAW's. Furthermore, I think this is probably the number one problem happening in audio today.
TaDaaa! You are correct!

And, unfortunately, these novices exclaim "That's the effect I was going for". Bull.

I remember years ago a rental mic case with a picture of a baseball and a caption saying "Just because you know what it is doesn't mean you know how to use it."


David Brown
Old 11th June 2006
  #34
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by aussie_techie
that is simply the nature of digital, you can not design a converter differently.
?? Sure you can. Take a converter that has absolutely linear response with a signal that swings from 0 - 5V. Suppose that it puts out 20bits to represent this signal. Now pair it with a circuit that attenuates the signal by 12db and feed that attenuated signal to a second instance of the same converter. When the upper two bits of the second converter are on, use the output of the second converter; otherwise use the output of the first converter.

Now you have a converter with 22bits of output, but the "targeted" "sweet spot" input voltage is 0 - 5V. It can tolerate sigals up to 20V; it has 12db of headroom.

This is a very rough sketch, but you see the idea. There's no reason why converters have to be designed so that they are most accurate right at the edge of clipping.

-synthoid
Old 11th June 2006
  #35
Lives for gear
 
ArcCirDude's Avatar
 

I don't know if anyone here has mentioned it but Roger Nichols addresses this gain structure issue in this months Sound on Sound. He claims that so many have disregarded the traditional gain structure practices from analog days that many overload the summing buss, thereby creating the need for analog summing devices. He even gives a spectrum analysis of a properly structured ITB and OTB mix which are, as far as I can tell and which Roger claims are identical. Interesting.
Old 11th June 2006
  #36
Gear Addict
 

Quote:
Originally Posted by Silver Sonya
So what I want to know is this: is my DAW concept a bad idea? Or a stroke of genius. What if DAWs stopped the user from clipping?

Lemme know what you think.
It's a good idea as long as it is a selectable option. I also like your idea that raising the input could lower the other faders to achieve an automatic re-leveling. I find myself having to do this often enough.

Your idea is similar to automatic latency compensation. We know the problem, and we can manually adjust, but automatic delay compensation just makes life easier and reduces error.

The DAW is a tool that should perform mundane activities to allow the engineer to focus more on creativity. I learned math w/o a calculator, and this perhaps gives me some deeper knowledge and appreciation, but I have no shame using a calculator these days to add a string of numbers. There should be no shame in having our tools help us achieve better quality.

I still think engineers should learn about proper gain structure, analog/digital concepts, and everything else about audio and music regardless of how much the equipment automates or hides these concepts. Knowing how a tube or a transistor or a DAC works provides a deeper appreciation for the work at hand, and helps develop problem solving skills. I'm still learning, and I enjoy it, and I also enjoy automation.
Old 11th June 2006
  #37
Quote:
Originally Posted by Dave Derr
Although Im not a big fan of most digital metering myself, the -60 limit on metering seems fairly reasonable, considering if you don't go right up to Full Scale on the ADC, you are losing bits and resolution. This is by definition. If the ADC isnt working right, thats another story.

Its quite simple in some ways. Lets say the ADC measures 65000 steps (about what a 16 bit converter does), if your peak signal stops at -6dBFs (6dB below full scale), you are only measuring 32500 of the 65000 steps, and missing 32500 steps of resolution! You just lost the cleanest bit of your conversion and are now working at 15 bits at best.

AND in actuallity, if you let the peaks only hit -12dBfs, you ARE losing 2 of your best bits of resolution and dynamic range. Your most innaccurate bits are the last bits, and the best bits are the first bits. Algebraically, it only takes 1 bit to lose half your dynamic range (-6dB is one half the voltage amplitude).

So if you have useful music below -60dB it had better be the verrrrrry tail end of a decay or reverb tail, and not a note you want to pick out clearly. The metering is there to help you record right up to Full scale, and encourage you not to record so low as to fall off the meter. Remember that old analog VU meters only showed about 26dB of dynamic range usually, so 60dB should seem like a blessing.
Can you translate this into practical terms?

Supposome I'm using a 1073 to record a snare. I've got one setting where in a three minute song the snare track will clip 8-12 times during tracking. Or, I can turn down the 1073's gain (which may change the tone - let's ignore this aspect for now) and cut the signal by 5dB, have no clipping and I'm 1dB from "algebraically losing half of the dynamic range".

Generally, I go with the former. I think that 8 snare hits over the course of 3 minutes is insignificant. What's your thought on this scenario?
Old 12th June 2006
  #38
Motown legend
 
Bob Olhsson's Avatar
 

One can often only guess at what the level is going to be. My experience has been that in most cases it's best to error on the high side with analog levels but on the low side with digital.

Where digital clipping bites you is downstream rather than upon immediate playback. This is because a digital recording is only a recording of numbers. Digital clipping introduces meaningless values that will create a gross loss of resolution should they subsequently be multiplied or divided to alter the volume level or otherwise process the sound digitally.
Old 12th June 2006
  #39
Quote:
Originally Posted by Mike Caffrey
Can you translate this into practical terms?

Supposome I'm using a 1073 to record a snare. I've got one setting where in a three minute song the snare track will clip 8-12 times during tracking. Or, I can turn down the 1073's gain (which may change the tone - let's ignore this aspect for now) and cut the signal by 5dB, have no clipping and I'm 1dB from "algebraically losing half of the dynamic range".

Generally, I go with the former. I think that 8 snare hits over the course of 3 minutes is insignificant. What's your thought on this scenario?
Not Dave but my take...

Have a fader knob fitted to the output of the 1073

http://www.marquetteaudiolabs.com/se...g_archive.html

Then you have the chance not only to trim level so as not to clip your converters, but can mess around with 'driving' the input heavily and still keep the output level tame enough for the converter (not to clip)

I wouldnt re-take for the sake of a few overs, but if I see them I usually make a mental note (or ask an assistant to) to turn down the level to the converter on the next take. Seeing an over light on say a snare or tom track (the main culprits for me) pisses me off.. I want to see no overs at the end of a take.
Old 12th June 2006
  #40
Lives for gear
 

Digital metering has been screwed ever since DATs started being used.
Actually before that because SONY F1s and PCM 701s had meters that were difficult to translate.
I used to use a SONY 3324 DASH machine with a Neve 8128. When you set the gain structure up between the console and the tape deck you had to COMPLETELY ignore the meters on the deck. We could also switch to an MTR90ii via multiplugs, but it wouldn't match.
In 1990 we started using WaveFrame 1000s and although they were a cool machine and we made thousands upon thousands of radio/TV spots on the eight we eventually owned they had horrible gain staging! For a NINETYFIVE THOUSAND DOLLAR machine they should have had unity gain... they didn't and we just lived with it.

In the early days of digital everything was interfaced with a console that had METERS!
It was easier to calibrate the whole system.

We went to rooms without consoles and imedeately non of our engineers could get the levels consistent mixing to DAT (early '90s through '00)
I had to builld stereo meter sets out of old MCI JH-110 electronics and calibrate them so they'd have meters.

I also have a DAT "calibration tape" that one of the techs at SONY's California Pro Division made for me using a test rig and a "calibrated standard" DAT machine.
It was the only way to know where the meters were suppossed to be for a +4 signal!

The biggest problem is that no-one knows what gain staging is any longer!
If you hook up a bunch of stuff without meters, or with meters that mean nothing of course it'll be wrong! It's guess work!

Digital allows people to work in such a wide window of signal level before it is truely bad that a lot of people never had to learn or worry about gain structure.
In the full analog days there was a fairly narrow window of acceptable sound.
You had to know how to set up a deck and calibrate it with your console!
Add in noise reduction (dBx made it REALLY hard!) and it was even more difficult!
Old timers can recall that switching to Ampex 456 was abig deal and then 499 was even a bigger deal! It was ONLY a three db increment each time! That's how close it was to being "good" or "bad."

More on this later....maybe an entire thread?

Danny Brown
Old 12th June 2006
  #41
Quote:
Originally Posted by dbbubba
The biggest problem is that no-one knows what gain staging is any longer!
If you hook up a bunch of stuff without meters, or with meters that mean nothing of course it'll be wrong! It's guess work!
Chicken or the egg? You hit the nail on the head I believe. The problem as I see it is that meters don't matter any more because the meters themselves don't mean a damn thing as you said. I think people (younger guys) don't know anything about gain staging because many outboard units don't have meters anymore and DAW meters are almost useless IMHO.

Without true meters you have to rely on your ears which is a good thing... but you might not know that you can push the analog unit more to get different sounds or that you should back off the AD converter to get more headroom.



Quote:
More on this later....maybe an entire thread?

Danny Brown
Yes good thread topic.... I might even hit that one up myself (wonder if anyone will be interested).

heh
Old 12th June 2006
  #42
Quote:
Originally Posted by Jules
Not Dave but my take...

Have a fader knob fitted to the output of the 1073

http://www.marquetteaudiolabs.com/se...g_archive.html

Then you have the chance not only to trim level so as not to clip your converters, but can mess around with 'driving' the input heavily and still keep the output level tame enough for the converter (not to clip)

I wouldnt re-take for the sake of a few overs, but if I see them I usually make a mental note (or ask an assistant to) to turn down the level to the converter on the next take. Seeing an over light on say a snare or tom track (the main culprits for me) pisses me off.. I want to see no overs at the end of a take.
I was trying to create as real a hypothetical solution as possible. I have a tube circuit that I run the output through to reduce the gain.

I'm just curious how much risk he thinks its worth to get up to that best bit.
Old 12th June 2006
  #43
The Distressor's "daddy"
 
Dave Derr's Avatar
 

Occasional Transient Clipping

Quote:
Originally Posted by Mike Caffrey
Can you translate this into practical terms?

Supposo I'm using a 1073 to record a snare. I've got one setting where in a three minute song the snare track will clip 8-12 times during tracking. Or, I can turn down the 1073's gain (which may change the tone - let's ignore this aspect for now) and cut the signal by 5dB, have no clipping and I'm 1dB from "algebraically losing half of the dynamic range".

Generally, I go with the former. I think that 8 snare hits over the course of 3 minutes is insignificant. What's your thought on this scenario?
You're probably right, especially if your ears don't detect anything nasty. If you're a dB over clipping with a transient like a snare you will probably never be able to hear it, unless the ADC has a wierd nasty clipping problem. Bell Labs determined years ago that clipping that lasts less than 1 mS (900uS actually) can't be heard. I suspect your short snare transient clipping may be inaudible. It doesnt hurt to have the drummer pound his snare and kick simultaneously as loud as they can at sound check, and set the peak levels from that. This may give you the worst case levels with a simultaneous slam of the two loudest drums.

The place you can really pic out nasty digital overs is low frequency stuff... like vocals, bass, silky sine-wavy stuff.

Also, with stuff like snares that will be mixed in with other tracks, recording 6dB below clipping (and losing a bit of resolution) also wouldnt be a problem with modern converters. Ya still got plenty of dynamic range and Signal to noise. Thats the big advantage of the 20bit converters, if youre working under the gun and cant really tweak levels perfectly, playing it safe by recording just slightly lower wont affect quality as much as with a 16 bit converter. Still, if you can tweak the levels to use all the bits, go for Full scale. I often have certain setups memorized so I know that with the mic gain set at such and such with this mic on this source, I can kind of go for a preset that I know will be really close to perfect recording levels. When I'm using a compressor while recording, its even easier to do preset type level setting.

See you out at Tape Op!
Old 12th June 2006
  #44
Gear Guru
 
u b k's Avatar
 

i've said it before, i'll say it again: people have the right to make bad sounding recordings. art is not beholden to your narrow little ideas of what is good and what is not good, the muse has no need to be squozen into tidy little boxes built by the hands of an ego trapped in its own vortex of judgments.

let 'em clip, let 'em turn it to hash. make your stuff the way you want to make it, and give everyone else the same courtesy.

or not. i reckon it's only the volume of acid in your stomach that's at stake.


gregoire
del ubik
Old 12th June 2006
  #45
11413
Guest
Quote:
Originally Posted by u b i k
i've said it before, i'll say it again: people have the right to make bad sounding recordings.
"A fool who persists in his folly (eventually) becomes wise..." -- William Blake

"why be bleak when you can be Blake?" -- Coil
Old 12th June 2006
  #46
Quote:
Originally Posted by Mike Caffrey
I think that 8 snare hits over the course of 3 minutes is insignificant. What's your thought on this scenario?
What if those overs and general hot level overloads the plug ins you put on it later?

What if you have to lower the inputs to plug ins to prevent this?

What if you have to have the fader in your digital mixer very low because of the high recorded level?
Old 12th June 2006
  #47
Gear Maniac
 
JohnNy C's Avatar
 

Quote:
Originally Posted by not_so_new
I don't think we are talking about "clipping" here... let me put that another way, I don't think we are talking exclusively about clipping here. You can run your mix hot ITB without any clip lights showing up and still get a mix that will fold in on it's self.

I think we are talking more about headroom in the mix and the fact that is seems like (to me anyway) with a DAW you need to leave even more headroom than with analog. As we all know, well as many of us know anyway, many good analog units will let you push the top of the input without sounding harsh even when clipped but digital is not that way. Even without clipping the buss a digital mix can get small with too much data.

Back down the faders on the mix and things will clear up so yes I agree with the idea that maybe the programmers / developers of the software should put some more headroom into the buss. I would not think of it as the developers having more control over you I would think of it more as the developers realizing that there are limitations to the medium and then work inside these limitations.
Maybe YOU are talking about headroom in the mix, but the original post as stated by this quote sure the hell wasn't:

"with someone who brings me a DAW session that is just shredding the input or the output of some plug-in or channel or buss. I'm not talking subtle distortion here. Im talking just bone-stupid clipping and harsh, spitty audio."

I think that if you use your headroom wisely, there is no reason to complain. And if your mix doesn't come out as hot as you want, just boost it in the mastering stage.
Old 12th June 2006
  #48
Lives for gear
 

Personally, this is why I record at 24 bit even though the result will ultimately be 16 bit. I often don't like to compress acoustic instruments on the way in, so I set my preamp before the converter to the highest level where I don't have to worry about clipping it at all, and record at 24 bit. Then, I can afford to lose the most detailed 2 or 3 bits due to not running into the converter really hot, and, if necessary I can lose another bit or two by turning up the volume itb, and I still have better than 16 bit resolution when I mix it down.
Old 12th June 2006
  #49
Lives for gear
 

Quote:
Originally Posted by synthoid
?? Sure you can. Take a converter that has absolutely linear response with a signal that swings from 0 - 5V. Suppose that it puts out 20bits to represent this signal. Now pair it with a circuit that attenuates the signal by 12db and feed that attenuated signal to a second instance of the same converter. When the upper two bits of the second converter are on, use the output of the second converter; otherwise use the output of the first converter.

Now you have a converter with 22bits of output, but the "targeted" "sweet spot" input voltage is 0 - 5V. It can tolerate sigals up to 20V; it has 12db of headroom.

This is a very rough sketch, but you see the idea. There's no reason why converters have to be designed so that they are most accurate right at the edge of clipping.

-synthoid
The loudest signals is represented by the most data, that could not easily be changed if it is possible at all. it is simply the nature of the digital stream and not limited to the A/D converter. What you described is an interesting way to increase the dynamic range of a converter and increasing the dynamic range decreases the need to run such hot signals in the first place. Although i dont know how easy it would be to implement in reality. It would be interesting if converter designers did look into it (if they havnt already) but they may be better off working on making the analogue section of a converter quieter to increase the dynamic range.
Old 12th June 2006
  #50
Lives for gear
 
synthoid's Avatar
 

aussie_tech,

well, what I sketched doesn't allow you to reduce the signal level -- it has attenuation but no gain, and the sensitivity of the underlying converters is what it is.

in any case what I'm getting at is that in the old days, 0db was not a brick wall. there is no reason for it to be a brick wall now either, so long as the recorded representation has greater precision than the converter output. we have chosen to jam the converter output up to the high bits of our representation, but we don't have to make this choice. 32bit float and 48bit fixed summing busses are in effect sliding the significant bits down in the representation to make headroom. I'm arguing that the converters should do this in the first place -- and they should do it gracefullly by providing some headroom on the analog side of the converter as well as the digital side. Then we wouldn't have to get so spooked about the occasional digital over during tracking. There would be "soft overs" and "hard overs", but no one (well, no one except dummies) would ever get a hard over.

-synthoid
Old 12th June 2006
  #51
Quote:
Originally Posted by JohnNy C
Maybe YOU are talking about headroom in the mix, but the original post as stated by this quote sure the hell wasn't:

"with someone who brings me a DAW session that is just shredding the input or the output of some plug-in or channel or buss. I'm not talking subtle distortion here. Im talking just bone-stupid clipping and harsh, spitty audio."

I think that if you use your headroom wisely, there is no reason to complain. And if your mix doesn't come out as hot as you want, just boost it in the mastering stage.
Well you quoted me so did you bother reading what you quoted? Did you see the part where I said "I don't think we are talking about "clipping" here... let me put that another way, I don't think we are talking EXCLUSIVELY about clipping here."

Thanks... have a good one.
Old 12th June 2006
  #52
Lives for gear
 

It takes more than blinky lights to know what is going on!

Most all converters only have a few LEDs and God only knows how they are calibrated!
Even if there is an input metering window in a program how do we know what it represents other than a red segment coming on? Who calibrated it? I suspect that they are there to make us feel good!

Since there is no real standard who knows what gain structure a given device or plug-in really wants to see?

My biggest complaint from the first time that I tried to "calibrate" a DAW was that everything had been pre-determined by an engineer or engineering group. With analog gear I could meter each section and see what was going on. I could insert a meter at patchpoints or on the actual circuitboard. You surely can't do this with a DAW!

I use a phase checker that produces a pulse and I can look at any point in the entire signal chain to see if phase is revered. I can't do this with a DAW.

DON'T EVER THINK THAT RULES AND PREVIOUSLY AGREED UPON PRACTICES ARE NOT THROWN OUT THE DOOR TO MAKE DESIGNING AND DELIVERING A PRODUCT TO THE FINAL PURCHASER POSSIBLE! The mere fact that for years Apple didn't allow Macintosh computers to have an open code policy (operating system) meant that almost all programs worked right out of the box. Because anyone could design and bring a piece of Windows software to the marketplace meant that a lot of stuff designed for Windows machines was incompatible.

THERE ARE NO LONGER ANY STANDARDS WITH DIGITAL EQUIPMENT!

There used to be fairly good practices that allowed most pro analog gear to work together.
The differences were small, but you could nearly alway make things work together happily. YOU KNEW WHAT WAS GOING ON WHEN YOU INTERFACED DEVICES WITH EACH OTHER IF YOU CARED TO LOOK. You cannot do this with digital in most cases.

I for one do not trust the information that I get from manfacturers because I have been told incorrect stuff time and time again by help desk people. This is seldom the case with analog equipment. Ask anyone from say MOTU about anything that is not in the manual and they can't (and probebly are not allowed to) answer the question. There are areas where they are not supposed to start talking about.

The fact that we just don't know what is correct or waht is going on inside the box is the crux of the entire question that this thread asks!

We had to raise a bit of hell to get DigiDesign to publish the White Papers.
I have asked manufacturers a lot of specific questions and it wasn't until people on forums started asking that the answers came out.

Trust me, any company would rather that you accepted their published specs and marketing info instead of them having to take time to explain themselves. It's only common sense. I'd hope that no one wants to hide it if there is good design practice going on. Bending rules and doing things the "easy way" in order to bring a product to market at a target price range doesn't help anyone. To me it is the difference between a pro product and a toy.

Danny Brown
Old 12th June 2006
  #53
Quote:
Originally Posted by Jules
What if those overs and general hot level overloads the plug ins you put on it later?

What if you have to lower the inputs to plug ins to prevent this?

What if you have to have the fader in your digital mixer very low because of the high recorded level?
I pretty much never use plugins and I pretty much never move the protools faders.

If someone comes to me for tracking only, I do set levels, as well as other things, differently.

And it's not like I've got every track constantly at the threshold of clipping - I'm not unclear how to set levels. My questions was more philosophical about balancing risk and reward - is it worth the risk of clipping to go for that top bit?

Personally, I think it is. If you chosing between an A- signal for 3.5 mintes or a A signal for 3.5 minutes with 8 incidents of 1 ms of clipping. I think the latter is better.

Or put it this way. If you set your levels more agressively on a track and get it 1% better, is it really worth the risk? Probably not.

But when you consider the cumulative effect of making 24 tracks 1% better, or 48 or 96, that starts to show.

I'm not necessarily saying distort every track so that you can get more at optimal level. I'm not even necessarily saying push it to the point of distortion at all.

I probably should have phrased it as "What's the cumulative effect on 48 tracks of leavling several dB of headroom, droppling a bit and getting losing "half the algebraic dynamic range'?"

I beleive that Dave's point is that the optimal bit does make a difference and htat's it's not ok to say "Just leave a ton of head room, it doesn't make a difference."
Old 12th June 2006
  #54
Quote:
Originally Posted by Mike Caffrey
I pretty much never use plugins and I pretty much never move the protools faders.
Do you mix in PT? or patch multiple outputs to an analog console?

Quote:
Originally Posted by Mike Caffrey
I beleive that Dave's point is that the optimal bit does make a difference and htat's it's not ok to say "Just leave a ton of head room, it doesn't make a difference."
I think there are definitely two schools of thought.

One - Believes - it's important to make most of the bits and is dedicated to hot levels to converters.

Two - Believes - It's important to avoid ANY converter input overload if possible because the internal digital architecture (plug ins / digital mixer) etc all perform better not being overdriven. and is dedicated to having no overs.

I've tried both & prefer school two...
Old 13th June 2006
  #55
Quote:
Originally Posted by Jules
Do you mix in PT? or patch multiple outputs to an analog console?



I think there are definitely two schools of thought.

One - Believes - it's important to make most of the bits and is dedicated to hot levels to converters.

Two - Believes - It's important to avoid ANY converter input overload if possible because the internal digital architecture (plug ins / digital mixer) etc all perform better not being overdriven. and is dedicated to having no overs.

I've tried both & prefer school two...
Console.

I follow that. I was curious what Dave's take was after reading his post.


I guess the real question is how far below 0dBFS you go, or many dBs does it take before at the cumulative level, you've made an appreciable difference.
Old 13th June 2006
  #56
Gear Maniac
 
JohnNy C's Avatar
 

Quote:
Originally Posted by not_so_new
Well you quoted me so did you bother reading what you quoted? Did you see the part where I said "I don't think we are talking about "clipping" here... let me put that another way, I don't think we are talking EXCLUSIVELY about clipping here."

Thanks... have a good one.
Wait, you lost me here. The original post was speaking of clipping. I responded to that post. Then the other posts following, started talking about headroom. And that's when you wandered into the room. So where am I wrong here?
Old 13th June 2006
  #57
Quote:
Originally Posted by JohnNy C
Wait, you lost me here. The original post was speaking of clipping. I responded to that post. Then the other posts following, started talking about headroom. And that's when you wandered into the room. So where am I wrong here?
Nice talking with you John.....
Old 13th June 2006
  #58
The Distressor's "daddy"
 
Dave Derr's Avatar
 

Sweet spot of converters

Really modern converters are pretty simple and forgiving... EXCEPT FOR OVERLOADS. The sweet spot is somewhere just below clipping but extends down probably 20dB or more below full scale. The more resolution the converter (IE Higher number of bits as in 20 to 24 bits), the "huger" the sweet spot is. You really dont have to worry about getting right up to 0dBfs with these higher res converters because they are very linear over a 20 to 30 dB range below full scale. BUT ONCE AGAIN, go as close as you can to 0dBFs without going over.... just to eek the last bit of performance out. If your peaks are 10dB down and you have a perfect take going into a 20bit converter, I wouldnt sweat it. But if you're going to do another take, why not push the levels up and eek out a little more performance?

Im kind of surprised that people have so much concern over the levels since in the old days in analog tape, the sweet spot was probably only 5 - 12dB before you started getting into Distortion on the top end, and noise on the low end. Dolby SR was awful nice to work with because it really did extend the dynamic range quite a ways, but not as far as a modern 20bit converter does.

When you are getting ready to record, just have the player or singer nail a really loud note and try to get your meters up close to 0dBFs with that. If youre nervous still, back it down a dB or two and press record. A good engineer should be able to do one run thru of a song and have pretty good levels, even for several instruments. Sticking my neck out, most experienced engineers will compress lightly many sources such as Vocals, bass, acoustic guitars when tracking, sometimes just to control the peaks, but there are exceptions to this. Some engineers make huge commitments in the sound as they record, tracking with lots of eq, compression, and perhaps even effects. Guys who have done lots of mixing often go for a more finished sound when tracking. I DO NOT SUGGEST THIS FOR ANYONE WHO DOESNT FEEL TOTALLY CONFIDENT, WITH YEARS OF EXPERIENCE. Its safer to get the best sound you can from the source and mic, and not gouge the signal with EQ and compression etc.

Watch compressing drums too much when tracking, especially kick and snare since if you mess up, you will have bleed and have lost some attack. Often a top notch engineer will have a tracking path for instruments he kind of uses over and over, a setup which he knows will get him a good sound with the right amount of dynamics. This is why many of them carry their own racks with them, or perhaps demand a certain studio with a certain Neve console to track on, one that they already know the sweet spots on.

Again, not getting the last few dB of dynamic range out of your converter is NOTHING compared to having a great song, good arrangement, good performance, good raw sounds, good Eq, compression, and balances. I think people obsess over silly technical issues like converters, while in the meantime, some glaring musical or tonal flaw is bashing them in the face. GET THE BIG THINGS RIGHT and the little things won't hardly matter, when its said and done.
Old 13th June 2006
  #59
I think the big things and the little things matter....
Attached Thumbnails
My Theory About Prosumer Audio-3.jpg  
Old 13th June 2006
  #60
Or is that too much compression?
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
thejook / So much gear, so little time
15
Ziggy!! / So much gear, so little time
12
DeadPoet / So much gear, so little time
1

Forum Jump
Forum Jump