View Single Post
Old 22nd May 2008 | Show parent
Gear Head
Dear Mickaël

I read your test and appreciate you are trying to do this. However as you noticed, doing a proper AB test is not as easy as it looks. This has to be done in a very controlled environment with preferably a few people observing every step you take.

There are a few flaws and unknowns in your test setup.
Let’s start with the source, the Fantom. For what I understand, this is an instrument (consumer) output level device. Which means, not enough to drive a converter expecting a professional level. As you mentioned the Rosetta 200 needed a lot more signal than you had available. So you fixed that by raising the level in Pro Tools, correct?
With that setup, you basically run the AD converter at a much lower bit depth than it potentially can. With a 16 bit source of the fantom you are now running an even lower number of bits on the AD converter because you are not getting to full scale. Raising the level afterwards in the digital domain (in PT) does not enhance the bit depth (resolution), it just makes it louder. Including all the artifacts (like distortion) of lower word length conversion.

You are apparently not aware of the feature to adjust the input level (and output) of the Rosetta 200. That way you would have a more optimal conversion. Please read the manual for the correct procedure

I assume the output of the Fantom is unbalanced. If so, make sure the unused pin (usually pin 3) of the XLR input is tied to ground. If not the transformer behavior like input of the Rosetta 200 will result in a loss of low frequency response and an overall “muddiness” in the lower mid range (as you noted)

Make sure the soft limit is off. (Very often a forgotten function)

It would be better to use the digital output of the Fantom with a high quality DAC for your conversion. The built in DAC’s in electronic instruments are usually not that great.

Far better would it be to use a natural source, like they recently did at the Fox stages to evaluate the best converter for their orchestral recordings. Blind listening tests were done comparing the converted version to the source (a life string quartet). I guess I don’t have to tell you who won that contest…

The bottom line is indeed comparing to the source ( I do that all the time with every new design). What does it do with your image, your spatial information, your depth? If there is a huge tonal difference like you describe, there is usually something wrong with the setup.

As Dr. Joy Degruy Leary noted, he did not hear the differences you noted with his Rosetta 800. They are indeed very much alike with the exception of the option to adjust the gain of the 200

Best regards,