Where's zero-db?

The Soundblaster Live SPDIF Performance

or Where the heck is 0db at that slider?

	Why?

	I made this little test because I got very curious when I actually managed to record a clean digital signal into my computer with distortion. "Strange...", I thought, "...isn't the whole idea behind using digital signals that you should get a perfect copy of the original?" Well, not in Creatives world, it seems. When running this test I managed to get the card to amplify my signal with 2.5 dBs...

	What did I do, exactly?

	Well, I started with by making two recordings of a 1 kHz sine tone onto a DAT at different levels; one at 0dB and the other at -12dB. Then I recorded these two signals (or test tones) back into my computer using Soundforge 4.5 and measured the peak input levels. Both playback and recording where done at a 48kHz samplerate, since this is supposed to be the default rate of the card. Also, all signal transfers where made digitally.

The results

	First of all, I established the point of infinite attenuation on the slider. No suprises here, it's at the lowest position.	Signal: 0dB 1kHz sine
	Then I moved the slider slightly upwards just to see what happened. Now the peak indicator said -26dB. Quite a jump there...	Signal: 0dB 1kHz sine
	Then I started to move the slider upwards (very carefully using the up-arrow on the keyboard). The picture shows where I got to -0.2dB. One nudge up and I got clipping.	Signal: 0db 1kHz sine
	To get an idea of exactly how much the card will amplify the signal I switched to the other test tone and moved the slider to where the peak indicator said -12dB (the same as the signal). The odd thing was, that it wasn't the same place as where I got -0.2db earlier. Anyways, if I moved the slider all the way to the top, I got -9.5 dB, which means that in this case, the amplification of the signal was 2.5 dB.	Signal: -12db 1kHz sine
	So where should you put the slider at, then? Well, to be perfectly honest, there is no straight answer since the "0db" spot moves around quite a bit for being a digital mixer (well, for being a mixer at all, really)... Anyways, the picture shows the position where I got the highest possible signal without any clipping using any kind of signal.	Signal: any

Conclusion

The SPDIF interface on the card is by no means useful for any "pro"-applications. On the other hand, the card is aimed at the consumer market, and for consumer needs I think it will do just fine.

Except that Creative makes a deal about this SPDIF interface and how it can give you "perfect" copies. I say that's impossible with this card. I can't imagine why they implemented copy protection in the first driver release, since perfect copies just couldn't be made...

Another strange observation; If I set the mixer at the position recommended above and press record with the in signal being digital silence from the DAT (no tape playing, no AD in progress), the noisefloor peaks at -74dB... Quite a lot for digital silence if you ask me. There must be a lot of strange things going on inside the card...

If you have any feedback on this report, email me!

1999-01-10