Dynamic range and equalization

[Zamiel]

hello, this is a technical question and maybe not so easy to answer. we know that, when we use an eq, the software reduces the dynamic range, the "bits", of the original stream in order to make calcuations and prevent overflow. does this happen even if we just apply attenuations? thank you

 

[DonH56]

hello, this is a technical question and maybe not so easy to answer. we know that, when we use an eq, the software reduces the dynamic range, the "bits", of the original stream in order to make calcuations and prevent overflow. does this happen even if we just apply attenuations? thank you

Attenuation means to reduce the signal level so yes. But reducing the level is not the same as reducing the dynamic range if the signal is still above the noise floor. If the music has a dynamic range of 70 dB (more than most sources), and the noise floor is 10 dB, then if you reduce the level from 40 dB to 110 dB down to 20 dB to 90 dB (1/4 the volume), you are still 10 dB above the noise floor so the dynamic range of the music is unaffected.

Note it is no different in practice than reducing the level using an analog volume control. The signal level is reduced and thus closer to the noise floor no matter how you attenuate. Whether the noise floor is analog or digital (usually a combination) makes no practical difference.

 

[kemmler3D]

hello, this is a technical question and maybe not so easy to answer. we know that, when we use an eq, the software reduces the dynamic range, the "bits", of the original stream in order to make calcuations and prevent overflow. does this happen even if we just apply attenuations? thank you

The answer is yes but it's worth mentioning that very often DSP converts the audio to 32 or 64 bits for processing. So per @DonH56 comment, this all takes places place miles above the noise floor.

 

[DVDdoug]

we know that, when we use an eq, the software reduces the dynamic range, the "bits", of the original stream in order to make calcuations and prevent overflow.

DSP is normally done in floating point which for audio purposes has infinite dynamic range (something like -700 to +700 dB).

If you boost you can end-up going over 0dBFS and then clip when it's converted back to integer for the DAC. If you lower the level and normalize for 0dB peaks before converting to integer you are still using all of the bits.

If you attenuate digitally so you peaks are lower than 0dB, then you aren't using the full dynamic range. If your peaks are at -6dB you aren't using the most significant bit. But it's usually insignificant especially with 24-bits.

Within reason (and if you aren't clipping) digital volume adjustments are considered lossless. They are an everyday part of audio production and nobody even thinks about it...

Pros generally record at 24-bits and often at around -18dB which means they are only using 21-bits. (But mixing is summation so they gain-back bits.)

 

[Chris A]

we know that, when we use an eq, the software reduces the dynamic range, the "bits", of the original stream in order to make calcuations and prevent overflow.

Do we know that? It depends on whether or not there are accompanying phase shifts (i.e., are you using minimum phase EQ vs. linear phase, etc.), and whether we're using EQ boost and attenuation filters.

does this happen even if we just apply attenuations?

If attenuations only are used, and induced phase shifts are minimal, generally speaking, the average output result will have lower average output levels. However, whether or not you actually have audibly lower dynamic range (crest factor) is a function of the EQ filters applied (center frequencies) and whether or not the audible frequencies in targeted audio stream/track are subjectively affected.

For instance, it's possible to apply an infrasonic filters to eliminate rumble from a captured analog music track from phonograph records--and not materially affect the core dynamic range of the musical frequencies themselves.
______________________________________________________________________________-

In general, though, if you are wanting to have real-time control of resulting music track dynamic range, there is a way to do that, in either real time while playing back music tracks (using something like foobar2000) or in batch mode using something like Audacity.

In general, I find that the Fabfilter Pro-Q4 plugin seems to be a bit more intuitive for visualization than the Pro-MB multiband upward expander/compressor, but care must be taken so that "breathing" that you can get with an overly upward expanded track doesn't become audible. You can do both static and dynamic EQ with these plugins.

Chris