feature request

Hi there, I am not sure if feature requests are place correctly here - if not, please move.

I am using Spectralayers in a totally different context to what most of the users here do, I reckon: For analysing bird and bat calls. In this regard I am missing a few things that should not be too difficult to implement, I hope:

[*]The first thing I desperately need is a tool that allows me to measure time differences. It is just annoying that I always have to switch to Audacity, to be able to do this.

[*]For Bats I use sample rates up to 250 kHz (even 384 kHz), so the enforced down-sampling to 192 kHz is problematic for my use. If you look to extend the application for your product, please consider this in future updates.

Thanks, Alfred

Hi Alfred,

Indeed something to measure time and freq differences could be useful. Will probably be added to a second patch. Meanwhile you can select a time range, type R to add a region matching this time range, and go to Project > Time Markers & Regions to see the length of the new region.

Regarding the 192 Khz limit, it’s mostly there to ensure real time performances on all computers (as 192 Khz is supposed to be plenty enough for normal audio use), but you can actually unlock this limit by going to the Windows Registry (Win+R, regedit) : in the registry go to HKEY_CURRENT_USER\Software\Steinberg\SpectraLayers Pro 6.0, then click Edit > New > String Value, call it maxsamplerate, and set its value to 384000. Restart SL, you should now be able to open 250Khz and 384Khz files natively.

Hi Robin,
thank you so much for your response! This is extremely helpful. For the first time since I have started using (and loving) Spectralayers 4, there is not only someone who listens, but provides helpful support, I am glad SL has ended up where it has.
In detail:

  1. The registry hack re sample rate solves all my problems in this regard.
  2. The work around concerning measuring times is even more useful, than you might think. Using the functionality, which you have outlined, allows me to measure the length of multiple calls by inserting as many regions as I like. I can then export them to EXCEL to calculate an average call length. This is more than I had hoped for. Additionally I can easily calculate an average call repetition time from the start time deltas. A bit of Excel macro work needed here, but that is a onetime effort.
    Really Nice!

Now, coming to a few more feature requests - you called for it :slight_smile::

  1. Would it be possible to show times (in particular the duration) with .1 ms resolution as opposed to the 1 ms we have now? Maybe you have another registry hack for this? This request is due to the high sample rate (and the short duration of the bat calls in particular) I am dealing with.

  2. When analysing Bat calls, one of the most important parameters for ID purposes is the so called “Frequency of Maximum Energy (FME)” within a call. When I use the statistics function of SL for a box, which I have set around a bat call, I get, amongst other pieces of information, the level for the maximum peak in dB. All I need now is an indication of the frequency for which this level is shown. Could this be done in a future version, please?

  3. SL has the function to show a graph along a linear line drawn in the 2D spectrogram plot. A simple line is not sufficient to reflect bat or bird call structures or any natural sound actually. The ideal solution would be the capability to draw “any” line in the spectrogram and get the levels along that line contour. Another possible solution would be to use a “spline type” curve, which could be fitted to the signal to be measured. I would also be happy, if there would be an option to compose a more complicated line from piecewise linear segments, it would probably be OK for my purposes.

  4. Another wish is to be able to not only draw a line and get the spectrum for this, but use a box instead and average the line spectra that fall inside this box and show it as an averaged spectrum curve.

Edit:
5. Forgot another really important one: Is there any chance SL will be able to handle BWF TIme Code any time soon?

I leave it here for the moment. I am not expecting this to fall out of the sky, but maybe you can add some of this to your “to do” list for the future. In general I see that you will have basically two groups of users for SL. On the creative side of things are those who want to modify and shape sound to their liking and on the other people like me doing forensic investigations of a given sound, not modifying it. I know that the creatives are the vast majority of your user right now, but I believe they could also benefit from some of the stuff that’s important to me and the other user group might grow in the future.
In any case, keep up the good work.
Alfred

Hi Robin,
forget my number 1. from the email before. I am now using the spectral region, which gives me actually time as well as frequency and I use “samples” for the time axis. Defining the start and end of the calls in the time domain is better anyway. After doing my EXCEL maths based on samples I can always go back to milliseconds with any resolution I want using my sample rate.

Alfred

Interesting job you have.

Hi CoinOP, no job - just a hobby :slight_smile:. A lot of people think I am a bit odd …

I was to quick: I can select samples, which are shown under Project > Spectral Markers and Regions, but when I export the data will be converted back to time as opposed to remain samples. My request would be to export using the format chosen for display.

Is there a similar trick to increase the frequency resolution on the Mac version as well?

Yes, you can edit the following file:
~/Library/Preferences/com.steinberg.SpectraLayers Pro 6.0.plist
there’s a section starting with and ending with , inside that section insert:
maxsamplerate
384000

Not at all. I love animals…afraid of bats though. I live across the zoo in Amsterdam Holland. As soon as it’s about to get dark a few fly up and down the street. When you open the window they fly by really close. It’s fascinating none the less.

Tnx Robin!