This is a thing i´ve been researching for a while, multiple articles describe different formulas. Veterans and Experienced engineers... how the convertion from analog level values is done when you add a digital equipment on the chain?
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Correct Analog to Digital levels
- Thread starter mpxbroadcast
- Start date
Any "formula" has to be based on at least some knowlege of the audio itself. The object in both analog and digital is to have audio between the maximum possible level and the noise floor, and if system signal to noise ratio is important, it should be as close to the maximum as possible without ever hitting it. The confusion everybody has is because typical analog audio distribution systems are monitored with a VU meter, but digital input levels are monitored on a peak meter calibrated to dBFS. There is no actual formula. There are general rules, standards, and opinions, but in all cases the audio should never hit analog clipping or 0dBFS.
The first problem then is knowing the actual audio level. There are different meters, different reference levels. The most common studio standard is to use a VU meter with the "0" reference = +4dBu (a real signal level on a connector or wire). Because the meter is not peak responding, normal program that moves the meter to the "0" point has peaks about 10dB higher, at +14dBu. Then there is headroom, or room for error. If there is anything live going on, there could be peaks much higher yet, so there should be another 10dB or more before the maximum of +24dBm (peak) is hit. If the audio is not live, or very well controlled, you might not need so much headroom.
Once you establish the above, you can select what reference level hits a specific point in the digital world of dBFS. 0dBFS is the digital maximum. Knowing absolutely nothing about the audio, you'd match 0dBFS to the analog clipping point, in the above case, +24dBu = 0dBFS (both peak of course). That will mean that analog program deflecting a VU meter to "0" VU will hit -12 to -10dBFS, with 10dB of headroom. If you know the maximum peak level of the analog audio, you can go to a higher dBFS point. But note that using a line-up tone will result in different numbers because the tone is steady and continuous, not varying program material, because the meter movement is slow and won't display a short peak value correctly. That's a big point of confusion, using a tone vs program, and using a VU meter vs peak dBFS meter. Tones move a VU meter higher than peak program hitting the same level. So a line-up tone would put "0"VU at -20 to -22 dBFS, assuming the levels above.
Any digital level you choose implies an assumption about the actual audio levels both of the program and the system capability. Not necessarily wrong, but you need more information.
A popular but incorrect assumption is to make all audio peaks come as close as possible to 0dBFS. Incorrect because that allows no room for error. It's achievable when using digital files only (using normalization), but not avisable for any sort of live audio.
And the last variable is the program distribution end point. What is it? A digital STL or transmitter? A stream? Mass content distribution like YouTube? They all have different reference levels and level standards, and measured with different means. YouTube, for example, will reduce a file audio level so it matches their -8 LUFS standard (but they won't turn low audio up). That level is HOT, and is why YT is a very loud audio source. LUFS is an entirely different means of audio level measurement that doesn't match VU or dBFS. So you can't make the final decision unless you know the final destination.
The first problem then is knowing the actual audio level. There are different meters, different reference levels. The most common studio standard is to use a VU meter with the "0" reference = +4dBu (a real signal level on a connector or wire). Because the meter is not peak responding, normal program that moves the meter to the "0" point has peaks about 10dB higher, at +14dBu. Then there is headroom, or room for error. If there is anything live going on, there could be peaks much higher yet, so there should be another 10dB or more before the maximum of +24dBm (peak) is hit. If the audio is not live, or very well controlled, you might not need so much headroom.
Once you establish the above, you can select what reference level hits a specific point in the digital world of dBFS. 0dBFS is the digital maximum. Knowing absolutely nothing about the audio, you'd match 0dBFS to the analog clipping point, in the above case, +24dBu = 0dBFS (both peak of course). That will mean that analog program deflecting a VU meter to "0" VU will hit -12 to -10dBFS, with 10dB of headroom. If you know the maximum peak level of the analog audio, you can go to a higher dBFS point. But note that using a line-up tone will result in different numbers because the tone is steady and continuous, not varying program material, because the meter movement is slow and won't display a short peak value correctly. That's a big point of confusion, using a tone vs program, and using a VU meter vs peak dBFS meter. Tones move a VU meter higher than peak program hitting the same level. So a line-up tone would put "0"VU at -20 to -22 dBFS, assuming the levels above.
Any digital level you choose implies an assumption about the actual audio levels both of the program and the system capability. Not necessarily wrong, but you need more information.
A popular but incorrect assumption is to make all audio peaks come as close as possible to 0dBFS. Incorrect because that allows no room for error. It's achievable when using digital files only (using normalization), but not avisable for any sort of live audio.
And the last variable is the program distribution end point. What is it? A digital STL or transmitter? A stream? Mass content distribution like YouTube? They all have different reference levels and level standards, and measured with different means. YouTube, for example, will reduce a file audio level so it matches their -8 LUFS standard (but they won't turn low audio up). That level is HOT, and is why YT is a very loud audio source. LUFS is an entirely different means of audio level measurement that doesn't match VU or dBFS. So you can't make the final decision unless you know the final destination.
Thanks very much for your help, This clarify a lot and confirms information i've found online. There's no correct formula. but you need to match the analog and digital dynamic range. Usually i like to run digital 0dbfs clipping at the analog max clipping level. 24bits sound cards and ADCs handle a huge dynamic range without introducing quantization noise. Another tip i've found online is matching the noisefloor. if your ADC noisefloor is around -120db. increase your analog gain until the noisefloor hits -119db. this means 1db of additional noise. but this should be done without introducing digital clipping. so you stay within the dynamic range limit of the analog equipment without introducing noise or clipping the ADC max eletrical limits.Any "formula" has to be based on at least some knowlege of the audio itself. The object in both analog and digital is to have audio between the maximum possible level and the noise floor, and if system signal to noise ratio is important, it should be as close to the maximum as possible without ever hitting it. The confusion everybody has is because typical analog audio distribution systems are monitored with a VU meter, but digital input levels are monitored on a peak meter calibrated to dBFS. There is no actual formula. There are general rules, standards, and opinions, but in all cases the audio should never hit analog clipping or 0dBFS.
The first problem then is knowing the actual audio level. There are different meters, different reference levels. The most common studio standard is to use a VU meter with the "0" reference = +4dBu (a real signal level on a connector or wire). Because the meter is not peak responding, normal program that moves the meter to the "0" point has peaks about 10dB higher, at +14dBu. Then there is headroom, or room for error. If there is anything live going on, there could be peaks much higher yet, so there should be another 10dB or more before the maximum of +24dBm (peak) is hit. If the audio is not live, or very well controlled, you might not need so much headroom.
Once you establish the above, you can select what reference level hits a specific point in the digital world of dBFS. 0dBFS is the digital maximum. Knowing absolutely nothing about the audio, you'd match 0dBFS to the analog clipping point, in the above case, +24dBu = 0dBFS (both peak of course). That will mean that analog program deflecting a VU meter to "0" VU will hit -12 to -10dBFS, with 10dB of headroom. If you know the maximum peak level of the analog audio, you can go to a higher dBFS point. But note that using a line-up tone will result in different numbers because the tone is steady and continuous, not varying program material, because the meter movement is slow and won't display a short peak value correctly. That's a big point of confusion, using a tone vs program, and using a VU meter vs peak dBFS meter. Tones move a VU meter higher than peak program hitting the same level. So a line-up tone would put "0"VU at -20 to -22 dBFS, assuming the levels above.
Any digital level you choose implies an assumption about the actual audio levels both of the program and the system capability. Not necessarily wrong, but you need more information.
A popular but incorrect assumption is to make all audio peaks come as close as possible to 0dBFS. Incorrect because that allows no room for error. It's achievable when using digital files only (using normalization), but not avisable for any sort of live audio.
And the last variable is the program distribution end point. What is it? A digital STL or transmitter? A stream? Mass content distribution like YouTube? They all have different reference levels and level standards, and measured with different means. YouTube, for example, will reduce a file audio level so it matches their -8 LUFS standard (but they won't turn low audio up). That level is HOT, and is why YT is a very loud audio source. LUFS is an entirely different means of audio level measurement that doesn't match VU or dBFS. So you can't make the final decision unless you know the final destination.