Kinda off-topic, but do you have any idea why my friend's Marc 2 has an ugly hiss when using DirectSound output to play a song (WAV, MP3, doesn't matter)? It sounds ok with WaveOut, very weird...

Same here and also for another member. Aknowledged by Marian's tech support.

Explanation : The Marian Marc has no proper DirectSound drivers for WinXP (nor probably for 2000 and NT).
Other annoying side effect : I can't play UT2003 at all with it since I upgraded from Win98 to WinXP. I had to get an old SB64 and set it as default sound device in order to play games.

On the other hand, these broken drivers allow to play bit perfect audio (no resampling) with wave out, while it should be impossible with WinXP (only in Kernel streaming and/or ASIO, or something like this).

That's me

They're very kind people there. But unfortunately they didn't have time to fix DirectSound. They did however send me RMAA reports of their own reference sample Marc2 when I asked (mine had some issues, apparently a monday-morning sample).

i'm going to buy a new audiocard and i thought marian marc 2 would be the best choice for me because i only want to record and work with live sound (no midi at all). i use wxp and freebsd. i don't know much about audiocards so please could anybody give me advice? is there a better alternative?

If you want to interface with pro audio gear, get the Echo Mia or MiaMidi, it features balanced I/O and pro levels.

If not, you could get a M-Audio Audiophile, which uses RCA connectors, unbalanced I/O and consumer levels.

A cheaper cousin of the later but still of equivalent quality is the M-Audio Revolution. It has the disadvantages of not having a digital in and using small mini-jack connectors. Also, current drivers make impossible to use the digital output with ASIO.

All these cards feature just 1 analog stereo in and out, except the Revo, which features 4 analog stereo outs.

Edit: according to some people's measuremets, all these 3 cards measure better in loopback configuration (out connected to in) than the Marian Marc card.

@sen: there's some chance of getting a card like the Audiophile working with BSD, since there are sources for Linux drivers available for its ubiquitous Envy-chip. Marian on the other hand use a non-audio generic-PCI chip, which is nice because it doesn't process the data in any way, but ofcourse there is no support for this hardly known card in the open source community.

@KikeG: if you're referring to the measurements of the Marc2 that I put up here - it seems I had a bad sample.

Yes, I was talking about those. Did you have your card replaced and got better results?

I had email contact with a Marian engineer, who tried to replicate my problem. His results were a lot better than mine. With 19 and 20 kHz tones at -13dBFS, I saw a highest -78dB peak at 18kHz. In the data he sent me, having the same tones at -13dB, the highest other peak was at 1kHz, and at -116dB.

I didn't have it replaced - unfortunately it was secondhand, and I didn't have any buyers receipt.

Edit: what's particularly funny about this thing is, that according to the guy at Marian, all data goes bitperfect to the DAC. Even volume adjustment is done at the DAC. So, if any intermodulation occurs, it's happening in the analog section (or in the DAC).
Now here's a test report of Hoontechs DSP24 Mk2. Same AK4524VF DAC, same IMD pattern as I had! However, Egosys' WT2496 also has this AK4524VF, and its test results show exactly the same pattern as the Marian engineer sent me: only a peak at 1kHz...
Either this AK4524VF had bad and good revisions, or it is very sensitive to the quality of the surrounding components... ?

Seems my friend has to get his card replaced, too. Analogue IN only wants to record in mono, with a ~-70dB noise on the other channel, and a very very faint 20 KHz tone. Don't ask me where that's coming from. Luckily he bought it in a real music shop and he even knows the owners, so a replacement is no problem. Then we'll see if the hardware was at fault.

I think the bad IMD can be due to poor analog input and/or output stages, not the DAC, as the Waveterminal measurements suggest.