K-System Metering

Page: < 1234 > Showing page 3 of 4
Author
Viktor
Max Output Level: -89 dBFS
  • Total Posts : 51
  • Joined: 2006/04/05 08:46:29
  • Status: offline
RE: K-System Metering 2006/12/04 20:00:15 (permalink)
the pictures adobe is showing are definitely wrong!
Regarding the first 9 showed samples, there is no way that the DAC will output something else than 0, simply because the DAC can not see (in the future) what the next samples will be.
For sure there are intersample overs, but not +6,5 dB.
It may look "nice" to some developer at adobe, but no DAC will output something like that.
#61
altima_boy_2001
Max Output Level: -55 dBFS
  • Total Posts : 2033
  • Joined: 2005/11/04 17:48:01
  • Location: Central Iowa
  • Status: offline
RE: K-System Metering 2006/12/04 20:01:52 (permalink)

ORIGINAL: UnderTow

ORIGINAL: altima_boy_2001
Somebody let me know if I'm wrong...

Isn't the reconstruction process itself determined by the DAC design, meaning no single standardized sample interpretation method? So if Sonar theoretically says your sample peak is -0.1 dB, but your intersample peak is +1.5 dB it could still be wrong. Someone's CD player will reconstruct the same samples to create a +2.0 dB intersample peak and another a +1.0 dB peak, etc.


Actually no. Sampling theory is about bandwidth limited signals so any half decent DAC should have the reconstruction filters limiting to the same bandwidth (half nyquist). The difference there is is that some DACs have more headroom in the analogue stages. A good DAC won't easily clip even with inter sample peaks while a lesser DAC will clip with the same signal.



That makes sense.

Thinking about this a little more, if different DACs converted the same set of samples into different analog forms (assuming no clipping/distortion) then a CD would sound slightly different depending on what player you put it in (if all other components remained the same). And in my experience that doesn't seem to happen.
#62
gnie
Max Output Level: -75 dBFS
  • Total Posts : 776
  • Joined: 2005/12/23 23:36:29
  • Status: offline
RE: K-System Metering 2006/12/04 20:38:30 (permalink)
Thinking about this a little more, if different DACs converted the same set of samples into different analog forms (assuming no clipping/distortion) then a CD would sound slightly different depending on what player you put it in (if all other components remained the same). And in my experience that doesn't seem to happen.

Really?
I've always noticed differences in the sound of CD players.
#63
DonM
Max Output Level: -34 dBFS
  • Total Posts : 4129
  • Joined: 2004/04/26 12:23:12
  • Location: Pittsburgh
  • Status: offline
RE: K-System Metering 2006/12/04 22:10:23 (permalink)
Yes indeed, peaks that would be greater than full scale--except we have to be careful about what we say here, because we're talking about analog voltage now, not sample values. But if we're talking about an analog voltage such that that 0dB "Analog" = 0 dBFS digital then yes, the analog voltage can be higher than 0. This is inherent in digital sampling and is exactly why the reconstructed analog waveform isn't stairstepped--in other words it's the result of a "good thing", we should just be aware of it.

T:
With all the respect you are obviously due.... Can truely calibrated 0VU equal DBFS - how could I be so confused on that? And since you've introduced the topic of voltage - I am now further confused - I thought we were talking about inter sample overs which is a purely digital issue - Oh... as I am typing I think I just got your point - if true 0 VU does in fact equal DBFS ...then... any voltage beyond 0VU would be over DBFS... okay I get that - but we then needn't consider the VU scale as the ballistics and raison d'etre of VU is totally different than digital meters right? I think from what I am gathering from you post - if I ignore the references to VU and just consider inter sample overs then things start to make sense. Now..... as for square wave - I think I just simply meant that an over would quantize as a square... but again as I type if it is truely 'intersample' the quantized waveform woudl'nt show a square obviously because it is 'over' - ahh now I see what you're saying. This is becoming increasing scary - So... all I want to know is if I increase my sample rate would I then increase square waves since I am reducing the time between samples and I would expect to reduce the intersample overs? As I begin to see the wisdom of your post - is it possible that what Audition is doing is showing overs based on a significantly higher sample rate than the PCM being written to file?

Very interesting!
-D

____________________________________
Check out my new Album  iTunesAmazonCD Baby and recent Filmwork, and Client Release
 
#64
John
Forum Host
  • Total Posts : 30467
  • Joined: 2003/11/06 11:53:17
  • Status: offline
RE: K-System Metering 2006/12/04 22:58:14 (permalink)
Very interesting yes. What good is it? All that is being said is that none of our great hardware and software is perfect. Wow I didn't know that. It will not stop me from recording and getting much better quality then I or anyone could have gotten 20 years ago. Over all the tools at our disposal now and those being purposed by the OP are so much better then anything I could have even dreamt of in the past.

There are so many problems with digital recording that if we were to list them in full none of the people that pioneered in this would ever have tried. (For example jitter)
(another is MIDI timing) You name it, one will find a problem.
But IT WORKS for the most part. No, none of it is perfect so what.

I am very happy that I can do what was unthinkable just a very short time ago. Plus, it keeps on getting better.

Heck I remember when wow and flutter was a real problem and having to crank that Victrola all the time. Not now!


Best
John

#65
Qwerty69
Max Output Level: -62 dBFS
  • Total Posts : 1435
  • Joined: 2004/02/19 17:44:10
  • Status: offline
RE: K-System Metering 2006/12/05 02:41:22 (permalink)
John -

I agree that the tools are lightyears ahead of yesteryear - particularly on the bang for buck scale. So said, this stuff IS fascinating and in gaining a greater understanding of these nuances, I am actively growing my knowledge and improving my results.

I don't doubt that there are large amounts of hair-splitting and 0.000x decimal places being calculated in the theory of this stuff, but ultimately I feel people get into these subjects because it does make them think more about the audio they are creating. Now if that person has the same amount of experience as you, then I guess this does seem a little pointless, but, if it is the vehicle by which people are learning, then where's the harm? Different strokes for different folks.

I'd just like some more customisations on the meter's GUIs.

Doesn't mean Sonar 6 sucks or I can't live without K-System meters, but the same essential urge in humanity which spawned digital recording as a solution to all the downsides of analog is what is motivating me to create this thread. I love my production tool of choice and would love it to get even better.

Ciao,

Q.
#66
John
Forum Host
  • Total Posts : 30467
  • Joined: 2003/11/06 11:53:17
  • Status: offline
RE: K-System Metering 2006/12/05 03:32:31 (permalink)
Ciao,
I was just venting. Pay me no mind. I do agree with you. I was just thinking that we are in a place where we "can't see the forest for the trees."
With forum members like yourself I don't really see this as a real problem.

And if you noticed I am not opposed to K- metering in Sonar at all. I was just a little concerned that some my think that what we already have is not very good after reading some of the posts here. Ultimately it is really a non issue.

Keep up the good work and all is well.

Best
John
post edited by John - 2006/12/05 03:51:14
#67
Qwerty69
Max Output Level: -62 dBFS
  • Total Posts : 1435
  • Joined: 2004/02/19 17:44:10
  • Status: offline
RE: K-System Metering 2006/12/05 06:05:59 (permalink)
Cool, cool... Nothing wrong with a good vent.

Q.
#68
Koed
Max Output Level: -90 dBFS
  • Total Posts : 20
  • Joined: 2006/10/09 14:39:54
  • Status: offline
RE: K-System Metering 2006/12/05 07:35:58 (permalink)
..geezzz .. this has turned into a complete blah blah thread..

Could we go back on topic and show the folks at Cakewalk the amount of support for implementing the K-sytem as a metering option amongst their customer base?
#69
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/05 10:07:10 (permalink)
ORIGINAL: Viktor
the pictures adobe is showing are definitely wrong!
Regarding the first 9 showed samples, there is no way that the DAC will output something else than 0, simply because the DAC can not see (in the future) what the next samples will be.
For sure there are intersample overs, but not +6,5 dB.
It may look "nice" to some developer at adobe, but no DAC will output something like that.

You're right, no DAC will output that exactly. However as I stated, those pictures are exactly what sampling theory dictates would happen in an ideal DAC. "Ideal" including infinite time in the future and past in order for all the lobes of the sinc function (time domain version of the low pass filter aka "smoothing filter") to add up to the final analog output.

I also admitted it's a contrived example to attempt to show just how far over 0 the analog voltage can go. If my contrived example were actually put through a real DAC, then no, it wouldn't be that much over. But it would be close, probably around +4 dB.
#70
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/05 11:39:24 (permalink)

ORIGINAL: DonM

Yes indeed, peaks that would be greater than full scale--except we have to be careful about what we say here, because we're talking about analog voltage now, not sample values. But if we're talking about an analog voltage such that that 0dB "Analog" = 0 dBFS digital then yes, the analog voltage can be higher than 0. This is inherent in digital sampling and is exactly why the reconstructed analog waveform isn't stairstepped--in other words it's the result of a "good thing", we should just be aware of it.

T:
With all the respect you are obviously due.... Can truely calibrated 0VU equal DBFS - how could I be so confused on that? And since you've introduced the topic of voltage - I am now further confused - I thought we were talking about inter sample overs which is a purely digital issue - Oh... as I am typing I think I just got your point - if true 0 VU does in fact equal DBFS ...then... any voltage beyond 0VU would be over DBFS... okay I get that


Yes, I never said anything about VU, and thats why I put "analog" in quotes. It was a "for example" case: For example, if 0dB in your analog signal = 0dB fullscale digital, then the analog voltage can go higher than 0 dB. It was to hypothetically set the 0 dB points to the same scale for digital and analog. Normally 0 VU is set to -20 dBFS for a sine wave (unless you set it to -18... or something else... but not 0 dBFS ) I have DACs here where a 0dBFS sine will output +24 dBu at the analog output.


- but we then needn't consider the VU scale as the ballistics and raison d'etre of VU is totally different than digital meters right? I think from what I am gathering from you post - if I ignore the references to VU and just consider inter sample overs then things start to make sense. Now..... as for square wave - I think I just simply meant that an over would quantize as a square... but again as I type if it is truely 'intersample' the quantized waveform woudl'nt show a square obviously because it is 'over' - ahh now I see what you're saying. This is becoming increasing scary - So... all I want to know is if I increase my sample rate would I then increase square waves since I am reducing the time between samples and I would expect to reduce the intersample overs? As I begin to see the wisdom of your post - is it possible that what Audition is doing is showing overs based on a significantly higher sample rate than the PCM being written to file?

Sample rate per se doesn't have anything to do with inter-sample overs, and Audition is not showing anything based on a significantly higher sample rate. (Unless you consider the visual pixels on the screen sampling at a higher rate... but that just confuses the issue so I shouldn't have even mentioned it). This is all based on sampling theory itself, and has to do with the final low-pass filtering stage.

A clipped quantized waveform does have a square top, and no increase in sampling rate will change that. And on the waveform display, no increase in sampling rate will change that either, because the waveform display (in Audition) is reconstructing the waveform (attempting to show what the analog waveform will be) from those clipped samples that are there in the digitized/quantized sample.

So if I understand what you're asking: If you're sampling at 48kHz (for example) and you have 2 full scale samples, there would be a certain amount of analog "over" when you reconstruct that waveform. Then if you're sampling at 96kHz, instead of having 2 full scale samples at that point you would have 4 full scale samples and thus less inter-sample over? Depends on the waveform that you're sampling. Sample rate really isn't related to this problem, although a higher sampling rate may make the clipping sound worse since there are twice as many samples at fullscale, therefore more "wiggle" on the reconstructed waveform. Just don't clip on your ADC stage.

Let's be perfectly clear here. The inter-sample overs that we are talking about are purely in the analog domain after the DAC (which is probably why hardly any digital meters/displays bother with it). And what we are talking about is a voltage that is greater than the voltage produced by one full scale digital sample--which doesn't necessarily mean distortion. If your DAC is of good quality and has the headroom (and we're talking about just a few dB here) then it's not a problem--it just outputs that voltage without being overdriven. You might want to worry about your consumers who may have equipment that might not handle it as well. That's why there's the suggestion of limiting your digital signal to -0.3 dB instead of just 0 dB full scale. That would take care of the problem right there.

And to the topic of this thread: Why is K-Metering important for this? While the original K-Meter spec didn't include provisions for inter-sample peaks, it was mentioned as a desirable feature (along with loudness weighting). K-Metering puts your working area well below the clipping point (K-20 and even K-14), and keeps your working area in the most accurate part of the meter.
#71
DonM
Max Output Level: -34 dBFS
  • Total Posts : 4129
  • Joined: 2004/04/26 12:23:12
  • Location: Pittsburgh
  • Status: offline
RE: K-System Metering 2006/12/05 18:43:49 (permalink)

Let's be perfectly clear here. The inter-sample overs that we are talking about are purely in the analog domain after the DAC (which is probably why hardly any digital meters/displays bother with it). And what we are talking about is a voltage that is greater than the voltage produced by one full scale digital sample--which doesn't necessarily mean distortion. If your DAC is of good quality and has the headroom (and we're talking about just a few dB here) then it's not a problem--it just outputs that voltage without being overdriven. You might want to worry about your consumers who may have equipment that might not handle it as well. That's why there's the suggestion of limiting your digital signal to -0.3 dB instead of just 0 dB full scale. That would take care of the problem right there.


Thanks for that clarification......up to.... this point: "...that we are talking about are purely in the analog domain after the DAC"

I am not sure what you mean about analog after conversion. Sorry to keep asking you to go back to the chalk board and school me again - Thanks in advance.

-D

____________________________________
Check out my new Album  iTunesAmazonCD Baby and recent Filmwork, and Client Release
 
#72
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/05 19:07:28 (permalink)
I am not sure what you mean about analog after conversion. Sorry to keep asking you to go back to the chalk board and school me again - Thanks in advance.


If I may answer this one (also to see if I'm understanding the subject here). I think what he means is that currently in most DAWs digital meters do not show inter-sample overs
because they occur after DAC (digital to analog convertion), meaning they happen after the wave form has been reconstructed and becomes analog again. If that's not the answer to your question then I quit from trying to understand all of this .
#73
John
Forum Host
  • Total Posts : 30467
  • Joined: 2003/11/06 11:53:17
  • Status: offline
RE: K-System Metering 2006/12/06 01:40:41 (permalink)
I think I'd check this meter out, too. Roger Nichols' over priced Inspector XL when he took it over, IMHO. Especially since the Nugen meter is now available. Nugen Audio Visualizer But I agree with mdw... leave the Sonar meter as is. Just like there's plug-in options for top-of-the-line effects, there's plug-in options for top-of-the-line meters, too. Cake doesn't need to make it their job to make the best of everything... just work on the DAW. And if there's anything sorely lacking in Sonar, it's individual pan controls for each side of a stereo track or bus instead one balance knob. It's ridiculous that this hasn't already been implemented. Please give us that before K System meters.

Javahut; Thanks for this heads up. I liked the Visualizer so much that I bought it. It has everything I was looking for in a analysis tool.

Thank you

Best
John
#74
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/06 10:50:02 (permalink)

ORIGINAL: DonM
Thanks for that clarification......up to.... this point: "...that we are talking about are purely in the analog domain after the DAC"

I am not sure what you mean about analog after conversion. Sorry to keep asking you to go back to the chalk board and school me again - Thanks in advance.

Jose7822 got it right. Inter-sample peaks by definition don't happen in the digital domain where only samples exist--nothing exists in between them. When you convert the digital samples back to analog, the reconstructed waveform is again continuous and can reach an analog value greater than the maximum digital sample value. Again taking into account we're comparing digital values to analog values.
#75
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/06 11:59:45 (permalink)
Jose7822 got it right.


Yesss! There's still hope for me . You just gotta love all this technical stuff---it's fun (man, I sound like a musical geek )
#76
DonM
Max Output Level: -34 dBFS
  • Total Posts : 4129
  • Joined: 2004/04/26 12:23:12
  • Location: Pittsburgh
  • Status: offline
RE: K-System Metering 2006/12/06 14:23:37 (permalink)
ORIGINAL: tarsier


ORIGINAL: DonM
Thanks for that clarification......up to.... this point: "...that we are talking about are purely in the analog domain after the DAC"

I am not sure what you mean about analog after conversion. Sorry to keep asking you to go back to the chalk board and school me again - Thanks in advance.

Jose7822 got it right. Inter-sample peaks by definition don't happen in the digital domain where only samples exist--nothing exists in between them. When you convert the digital samples back to analog, the reconstructed waveform is again continuous and can reach an analog value greater than the maximum digital sample value. Again taking into account we're comparing digital values to analog values.

Darn.... I thought my questions would end at that point..... So if I get this right, I digitize an analog wave the quantization in the digital domain represents the best it can the wave's form with DBFS at the top of the range... then I convert it back to analog and after that D2A conversion something appears that was there before digital quanitization. How or where were the overs stored...or are the overs not from the original analog waveform, they are just a result of the analog reconstruction? The latter makes more sense to me. Thanks again for your willingness to keep schooling me.

-D

____________________________________
Check out my new Album  iTunesAmazonCD Baby and recent Filmwork, and Client Release
 
#77
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/06 17:46:02 (permalink)

ORIGINAL: DonM
So if I get this right, I digitize an analog wave the quantization in the digital domain represents the best it can the wave's form with DBFS at the top of the range...

Yes. And the best of digital's ability includes all frequency components below half the sampling rate, and a dynamic range of ~6dB times the bits in the digital system. (more or less)

then I convert it back to analog and after that D2A conversion something appears that was there before digital quanitization.

No, considering what we're discussing. The only thing added is quantization noise, which isn't the same thing we're discussing--the inter-sample peaks. Remember, sampling theory states that exact reconstruction of a continuous-time baseband signal from its samples is possible if the signal is bandlimited and the sampling frequency is greater than twice the signal bandwidth.

I'll emphasize: exact reconstruction is possible. That's the beauty of digital sampling. Given the bandlimited signal, the only change is in the quantization distortion due to the bit depth, which isn't the inter-sample peaks.

How or where were the overs stored...or are the overs not from the original analog waveform, they are just a result of the analog reconstruction? The latter makes more sense to me.

The "overs" are stored in those sample values themselves. They aren't just a result of the reconstruction, they were there originally and put back in upon reconstruction. Remember that the original waveform in between all sample points is continuously varying, not a straight line. It's not wiggling between sample points, but it is a curve. (I really need a picture here...) And there are areas between some points where the peak of that curve is higher than the two sample points. It really is one of the beautiful aspects of sampling theory.

Unfortunately, the Wikipedia link I provided above is pretty technical. I'm still trying to find a good non-technical outline of the D-to-A process. At the bottom of the Wikipedia page there is a link to a Lavry Engineering PDF on sampling. Pages 23-26 give some good examples. The best book on the A-to-D process I've found is Nika Aldrich's Digital Audio Explained. I was hoping it would go more into D-to-A conversion, the sinc function, and waveform reconstruction but it kinda skims that part. However, overall that book is an incredible resource for anyone who wants to understand digital audio.
#78
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/06 17:51:01 (permalink)
ORIGINAL: DonM
So if I get this right, I digitize an analog wave the quantization in the digital domain represents the best it can the wave's form with DBFS at the top of the range...

Yes. And the best of digital's ability includes all frequency components below half the sampling rate, and a dynamic range of ~6dB times the bits in the digital system. (more or less)

then I convert it back to analog and after that D2A conversion something appears that was there before digital quanitization.

I'm not quite sure what you're saying here, but I think you're still not quite understanding considering what we're discussing. It's not like the digital samples aren't storing what was there in the original analog waveform, they are storing the whole thing. The only thing added is quantization noise, which isn't the same thing we're discussing--the inter-sample peaks. Remember, sampling theory states that exact reconstruction of a continuous-time baseband signal from its samples is possible if the signal is bandlimited and the sampling frequency is greater than twice the signal bandwidth.

I'll emphasize: exact reconstruction is possible. That's the beauty of digital sampling. Given the bandlimited signal, the only change is in the quantization distortion due to the bit depth, which isn't the inter-sample peaks.

How or where were the overs stored...or are the overs not from the original analog waveform, they are just a result of the analog reconstruction? The latter makes more sense to me.

The "overs" are stored in those sample values themselves. They aren't just a result of the reconstruction, they were there originally and put back in upon reconstruction. Remember that the original waveform in between all sample points is continuously varying, not a straight line. It's not wiggling between sample points, but it is a curve. (I really need a picture here...) And there are areas between some points where the peak of that curve is higher than the two sample points. It really is one of the beautiful aspects of sampling theory.

Unfortunately, the Wikipedia link I provided above is pretty technical. I'm still trying to find a good non-technical outline of the D-to-A process. At the bottom of the Wikipedia page there is a link to a Lavry Engineering PDF on sampling. Pages 23-26 give some good examples. The best book on the A-to-D process I've found is Nika Aldrich's Digital Audio Explained. I was hoping it would go more into D-to-A conversion, the sinc function, and waveform reconstruction but it kinda skims that part. However, overall that book is an incredible resource for anyone who wants to understand digital audio.

post edited by tarsier - 2006/12/06 18:09:58
#79
DonM
Max Output Level: -34 dBFS
  • Total Posts : 4129
  • Joined: 2004/04/26 12:23:12
  • Location: Pittsburgh
  • Status: offline
RE: K-System Metering 2006/12/06 19:10:00 (permalink)
T:
Thank you very much for your persistence. I apologize to those who feel we went off topic. I'm going to simmer on this for a while. With your permission I may open a new thread properly titled to carry on the topic as I think more about it. Thank you again!

-D

____________________________________
Check out my new Album  iTunesAmazonCD Baby and recent Filmwork, and Client Release
 
#80
gnie
Max Output Level: -75 dBFS
  • Total Posts : 776
  • Joined: 2005/12/23 23:36:29
  • Status: offline
RE: K-System Metering 2006/12/06 21:34:35 (permalink)
Yes, thanks to all for the great information.
It's been very interesting.
post edited by gnie - 2006/12/07 01:30:34
#81
Qwerty69
Max Output Level: -62 dBFS
  • Total Posts : 1435
  • Joined: 2004/02/19 17:44:10
  • Status: offline
RE: K-System Metering 2006/12/07 05:06:04 (permalink)

ORIGINAL: DonM

T:
Thank you very much for your persistence. I apologize to those who feel we went off topic. I'm going to simmer on this for a while. With your permission I may open a new thread properly titled to carry on the topic as I think more about it. Thank you again!

-D


ORIGINAL: gnie

Yes, thanks to all for the great information.
It's been very interesting.


...and I carry the motion. Thanks T. and all the others who contributed to this thread!

Q.
#82
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/07 21:44:53 (permalink)
I have a couple of questions about K-System calibration.

The way I know how to calibrate my monitors is by using a software program like "Room EQ Wizard" and play a pink noise set up to -20 RMS. Then I raise the levels of each of my monitors (one at a time) to were they show 83 dBFS on my SPL meter which is set up for C weighted, Slow response and pointed directly (in a 45 degree angle) to the speaker I'm calibrating at the mixing/listening position.

My question is, am I suppose to mix at -20 RMS or at 0 dBFS? Is the way I meantioned earlier the correct way of calibrating according to the K-System?

Thanks in advance.
#83
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/08 09:42:46 (permalink)

ORIGINAL: Jose7822

I have a couple of questions about K-System calibration.

The way I know how to calibrate my monitors is by using a software program like "Room EQ Wizard" and play a pink noise set up to -20 RMS. Then I raise the levels of each of my monitors (one at a time) to were they show 83 dBFS on my SPL meter which is set up for C weighted, Slow response and pointed directly (in a 45 degree angle) to the speaker I'm calibrating at the mixing/listening position.

Yes, that's basically it, as long as you know that Room EQ Wizard is playing out the noise properly. (I haven't tested it so I can't vouch for it. It has a good reputation, though) I would just add that you should have the faders on your hardware mixer (if you're using one) set to 0 and you adjust the signal level with the input trims (on input) and monitor control (on output). I couldn't tell if you were doing that or not. I use a digital I/O soundcard, and Yamaha 02R digital mixer so my flow goes Sonar Master Out Fader at 0->02R digital in->Line Faders 0->Master Fader 0->Monitors. I don't do any mixing on the 02R, it's just for tracking and monitoring purposes these days. I mix all in Sonar.


My question is, am I suppose to mix at -20 RMS or at 0 dBFS? Is the way I meantioned earlier the correct way of calibrating according to the K-System?

-20 dB RMS. But be aware that Sonar's RMS meters are calibrated -3 dB from the AES-17 (and thus the K-System) standard. So in this case, your Sonar RMS meters should be bouncing around -23 dB RMS during an average "nominal" section of the music. Like the first verse of a pop song. You might find that this gives you a lot of headroom before hitting 0 dBFS Peak. For most of the stuff I do, I've taken to mixing to K-14 instead. So in that case, I've turned down my monitors by 6 dB and turned up my mixes by the same amount. Thus my Sonar RMS meters are bouncing around -17 dB RMS when I mix.

Once you get used to how that sounds in terms of sound levels, try turning off the meters. You can get pretty good at adjusting levels and mixing without meters once you're comfortable with the calibrated monitor set up.
#84
gnie
Max Output Level: -75 dBFS
  • Total Posts : 776
  • Joined: 2005/12/23 23:36:29
  • Status: offline
RE: K-System Metering 2006/12/08 12:30:11 (permalink)
tarsier,

Just to be clear, you're speaking of final, master out levels here. Within the box, we still have the floating point engine's flexibility, right? Our main concerns are initial input levels [basically not exceeding the sound card's or converter's threshold] to output level. Because any signal, once in, will be properly represented numerically. Within the gates of the two conversions [that is, our individual tracks, synths, plug-ins running internally on the PC] do we need any targets?
#85
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/08 14:46:41 (permalink)
Transier,

Thanks alot for answering my questions, but I'm kinda confused about something here. I completely understand everything you explained except for the part where I have to mix at -20 dB RMS (-23 dB RMS in Sonar). To be more specific, after I calibrate my monitors to 83dBFS using pink noise at -20 dB RMS (or is it -23 dB RMS?) from "Room EQ Wizard", do I still have to have my mixes at a nominal level of -23 dB RMS in the Master Bus?...is this correct? I just thought that after calibrating the monitors that everything was done and all I had to do was to mix normally to 0 dBFS.

P.S. I don't use a hardware mixer, just Sonar. My chain goes from Instrument > ART StudioV3 pre-amp > M-Audio Delta 44 > IN to Sonar and OUT to Monitors. I also want you to know that I'm using K-20 as an example to understand all of them. Most likely I will end up using K-14. But for now I just want to understand K-20 and then make adjustments accordingly .
post edited by Jose7822 - 2006/12/08 15:07:35
#86
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/08 16:28:14 (permalink)

ORIGINAL: gnie

tarsier,

Just to be clear, you're speaking of final, master out levels here. Within the box, we still have the floating point engine's flexibility, right? Our main concerns are initial input levels [basically not exceeding the sound card's or converter's threshold] to output level. Because any signal, once in, will be properly represented numerically. Within the gates of the two conversions [that is, our individual tracks, synths, plug-ins running internally on the PC] do we need any targets?

Yes, I was speaking about the master out level. To your other question, I think you're talking about the gain staging issue? There have been a few good threads around here on that topic lately. I would say, generally, that for individual tracks yes there is a target level but it varies depending on the type of material in the track. Plus, you're going to play around with those levels anyway, while you're mixing.

Here are the recent threads on level setting:
Gain staging important in tracking too
What is trim for?

#87
tarsier
Max Output Level: -45 dBFS
  • Total Posts : 3029
  • Joined: 2003/11/07 11:51:35
  • Location: 6 feet under
  • Status: offline
RE: K-System Metering 2006/12/08 17:09:52 (permalink)

ORIGINAL: Jose7822
Thanks alot for answering my questions, but I'm kinda confused about something here. I completely understand everything you explained except for the part where I have to mix at -20 dB RMS (-23 dB RMS in Sonar). To be more specific, after I calibrate my monitors to 83dBFS using pink noise at -20 dB RMS (or is it -23 dB RMS?) from "Room EQ Wizard",

I don't have Room EQ Wizard, so I don't know exactly what it's putting out. Also, if you're using a mono pink noise source in Sonar, you have to verify your pan law settings and adjust accordingly. The best way to do it (IMO) is Download Bob Katz's pink noise file and play it back from Sonar. It's a stereo file so play it in a stereo track, out of a stereo bus with all trims and volumes at 0, and balances set to center. Verify that Sonar's meters are where you expect them to be while playing it back (in this case, the RMS meters should show -23) Then turn off one of your monitors and calibrate the other with your SPL meter, then switch to calibrate the other. That way you know you've got the correct levels set. Then if you calibrate again from Room EQ Wizard and the levels are the same, then you know everything is operating at the same level.

do I still have to have my mixes at a nominal level of -23 dB RMS in the Master Bus?...is this correct? I just thought that after calibrating the monitors that everything was done and all I had to do was to mix normally to 0 dBFS.

Yes, your nominal levels should be at -23 dB RMS on Sonar's master out meter if you're mixing to K-20. That's the point of the K-System: Focus on the message (the music) not the medium (the peak value of the CD, for example). Mixing to that level (which should be a comfortably loud level) lets you worry only about getting the sound right in your mix, not about where your meters are peaking. At K-20 your peak meters will probably never get up to 0. If your final output is a 24 bit mix, then you've got plenty of resolution to work with.

P.S. I don't use a hardware mixer, just Sonar. My chain goes from Instrument > ART StudioV3 pre-amp > M-Audio Delta 44 > IN to Sonar and OUT to Monitors. I also want you to know that I'm using K-20 as an example to understand all of them. Most likely I will end up using K-14. But for now I just want to understand K-20 and then make adjustments accordingly .

What do you use for volume control between Sonar and Monitors? Sonar itself? And yes, ultimately you do what works best for you.
#88
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/08 20:39:05 (permalink)
Hey tarsier!

Thank you for all the info. The K-system is a lil bit confusing at first but it definitely makes sense to be able to play different CD's without having to make extreme level changes between them when listening. Plus it would make everybody's life easier by having a common level. To answer your question about Room EQ Wizard, the program allows you to set the levels internally to whatever you want to. Then all you do is set the outside levels with your monitors, leaving all the faders, trims, etc intact in Sonar. I'm gonna calibrate my monitors using your method and then with Room EQ to see if I get the same results and I'll let you know about it. By the way, my pan laws are set to 0dB sides, -3dB center sine wave.

What do you use for volume control between Sonar and Monitors? Sonar itself?


I use the Delta 44 mixer. I basically set the monitors volume knob to full strength and control it from Delta's mixer. It's way easier than having to control the volume from the back of each monitor every time I need to turn the volume up/down. Again, thanks for your help.

P.S. My confusion came about because with Room EQ Wizard I don't have to touch anything in Sonar's Master Bus. In fact, you don't need Sonar at all...just open Room EQ, calibrate and your done. Then open Sonar and start mixing.

#89
Jose7822
Max Output Level: 0 dBFS
  • Total Posts : 10031
  • Joined: 2005/11/07 18:59:54
  • Location: United States
  • Status: offline
RE: K-System Metering 2006/12/08 21:00:01 (permalink)
Just wanted to add that Room EQ Wizard not only has a full range stereo pink noise, but also a narrow stereo pink noise for speaker calibration as well as a low range pink noise for sub calibration. Of course it has more test tones, but those are the pink noise ones. It has alot of cool features and its free, which is awesome. If youre interested in it you can download it from HERE. All you have to do is sign up to the forum and then you can download it.
#90
Page: < 1234 > Showing page 3 of 4
Jump to:
© 2022 APG vNext Commercial Version 5.1