Logic Tone Generator good enough?

[brewery]

I've been using Logic's test tone plug in for a long time now to do record alignments and print tones on my 2 track.

Yesterday I thought that it would be more convenient if I had one audio file to play back all the tones plus bass sweep, so i bounced them all down into one file.

Everything seemed fine until I looked at the hi frequency tones in the sample editor.
50, 100 and 1k looked like nice solid sine waves, but both 10k and 15k looked really jagged and uneven.
The output is still solid on the Logic meters, and on the Studer the meters just rock a tiny bit, like they always do at high frequencies.

I guess it's not really a problem since the level is steady, and the tones sound like a clean sine wave - maybe it's even just the resolution of the sample editor.

Would this be a concern for any of you guys?

Thanks
Erik

[Lagerfeldt]

44.1 kHz sample rate?

[Nordenstam]

It's about how the program displays the waveform.

This is a sinewave close to 1/4 the sample rate as shown in sound forge with "connect the dots" approach to visualizing the wave:


This is what a similar frequency wave looks like as shown in Izotope RX's reconstructed waveform display:

[brewery]

Yup, sample rate was 44.1kHz, 24bit - didn't even think about how that would affect it..

I don't really mind what the wave form looks like as long as it gives me a proper alignment, and people can use my tones to align their machines to them (which so far always worked without any complaints)

Would you invest in an external signal generator, or should I just keep doing it this way?
(at 96k)

[brewery]

just noticed another weird thing:

close up the waveform looks like this:



but if you zoom out a bit it looks like the whole thing is oscillating at lower frequency:



If you can trust these graphics at all, it looks like the slower cycle is roughly 1K.


In 96kHz it looks a lot more even:

[Maelgad]

if to save processing power the program only draws 1 point per 4 samples, this then lowers the nyquist frequency by 4x, which would explain all the oscillations you're noticing.

Also i doubt the display would calculate oversampling and smoothing that can be done in the converter.

[glassmaster]

Originally Posted by El Kuto

Would this be a concern for any of you guys?

Thanks
Erik

Nope.

[Nordenstam]

Originally Posted by El Kuto

If you can trust these graphics at all, it looks like the slower cycle is roughly 1K.

You can't trust those graphics. They connect each sample dot with a line. What really happens at the output of the converter is much more complex than that!

Here's a picture that shows both the connect-the-dots approach to visualizing the wave(like your program does) and an interpolated view that resembles what happens on the output of your digi->annie converter.



The upper row is the same as you see in your program. The lower row is what really happens when the samples are converted to analogue. The picture was made using the Izotope RX software. Adobe Audition also does such a reconstructed waveform view. Download the demo of one of the programs and play around with it. That way of visualizing the waves are much more useful for looking at waves than programs that only does connect-the-dots waveform display.

This article may help: Intersample peaks