Re: Spectrum analysis in JAMin is incorrect.

On 03/29/2011 02:35 AM, Steve Harris wrote:
> Yes, it's meant to map to octaves, rather than linear frequency. This tends to be what musicians expect, but isn't necessarily appropriate for other uses.
Yeah, the display is *technically* correct if you're dealing with 
*energy or power* rather than *amplitude*. Since white noise has equal 
amplitude in every frequency band, the amount of power per band will 
increase with frequency at 6 dB / octave. That's what the graph is 
displaying.

But as a mastering engineer, what I care about is amplitude relative to 
0dBFS. I think it's more logical, based on the labeling of the graph on 
the Spectrum page, to assume it's measuring amplitude, not power. The 
FFT graph on the HDEQ page does display amplitude, so the two graphs are 
inconsistent anyway, and an engineer who has not tested the graphs' 
behavior could get confused or worse.

David
> - Steve
>
> On 2011-03-29, at 09:08, Dan wrote:
>
>> The frequency scale isn't linear, it looks like it's at least roughly
>> logarithmic, in which case it's quite possible that *is* what white
>> noise looks like.
>>
>> Best
>> Dan
>>
>> 2011/3/29 David Nielson<[email protected]>:
>>> After changing the update frequency for the FFT analysis in JAMin, I
>>> thought I'd see if I could fix another long-standing complaint of mine:
>>> the graph never even gets close to the top of the window, so while there
>>> might be really interesting things going on at low volume levels, I
>>> can't see them.
>>>
>>> It seems like the signal is getting some automatic gain control
>>> (compression) before the FFT stage, separate from the compression that
>>> JAMin is doing itself, to keep the visual of the signal visible. Passing
>>> several test tones through to see what they looked like, I discovered
>>> this in the 30-band spectrum page:
>>>
>>> http://www.naptastic.com/david/jamin-graph-wrong.png
>>>
>>> The input being analyzed is white noise. It displays correctly on the
>>> HDEQ page, but here, well, that ain't what white noise looks like. (You
>>> can ignore the peak indicators being in weird places; that's from the
>>> other test tones I ran.)
>>>
>>> So, I'm gonna jump back in the code right now, but I thought y'all might
>>> want to be aware, there are some gremlins in there that it seems no one
>>> has noticed before.
>>>
>>> ------------------------------------------------------------------------------
>>> Enable your software for Intel(R) Active Management Technology to meet the
>>> growing manageability and security demands of your customers. Businesses
>>> are taking advantage of Intel(R) vPro (TM) technology - will your software
>>> be a part of the solution? Download the Intel(R) Manageability Checker
>>> today! http://p.sf.net/sfu/intel-dev2devmar
>>> _______________________________________________
>>> Jamin-devel mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/jamin-devel
>>>
>>
>>
>> -- 
>> http://www.mcld.co.uk
>>
>> ------------------------------------------------------------------------------
>> Enable your software for Intel(R) Active Management Technology to meet the
>> growing manageability and security demands of your customers. Businesses
>> are taking advantage of Intel(R) vPro (TM) technology - will your software
>> be a part of the solution? Download the Intel(R) Manageability Checker
>> today! http://p.sf.net/sfu/intel-dev2devmar
>> _______________________________________________
>> Jamin-devel mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/jamin-devel
>


------------------------------------------------------------------------------
Enable your software for Intel(R) Active Management Technology to meet the
growing manageability and security demands of your customers. Businesses
are taking advantage of Intel(R) vPro (TM) technology - will your software 
be a part of the solution? Download the Intel(R) Manageability Checker 
today! http://p.sf.net/sfu/intel-dev2devmar