Prism Sound Orpheus : Test Drive :

[TAFKAT]

I finally got a chance to have a test drive of one of these things today..

Nearly all feedback I have read has concentrated on its almost mythical sound quality, whereas my main focus was the stability of the ASIO driver within Nuendo on a clients new i7 / XPSP3 system I am in the final stages of configuring.

To say I was less than impressed is an understatement..

Latest driver , control panel and firmware V1.5 :

First off, lowest latency is 10 MS with its standard buffer setting ( buffer and latency settings are separate) - translating to 12 MS Input/23 MS output in Nuendo, no ASIO direct monitoring, but does have DSP mixing and monitoring. Loading a demo song and assigning the outputs in VST connections hung the application. 3 finger solute, and relaunch settled it and enabled some playback. Opening another song which required reassigning of the outputs, and again the system stalled, this time managing to recover after a few seconds.. WTF.

After some more fiddling I realized that after any change to the audio settings, you need to wait a few seconds before clicking on anything else, as the wheels are still spinning in the background. After realising this I managed to tippy toe around without further incident. I say tippy toeing , as I didn't feel overly confident in the driver to be honest.

Next up was to run some of the stress benchmarks up , and with just N3.2 installed at present I chose Blofelds DSP II. Session loaded and outputs assigned it was time to start loading up the MBC plugs. On enabling the first plugin I knew I was in for a ride, huge pause in playback of a few seconds, the more plugs the worse it got. I stopped at 4-5 as it was just unworkable. The ASIO meter was only floating around 5-10% , so no problem with the overhead its just the realtime capability of the driver is non existent. After being used to the realtime performance of RME and Lynx all these years, I was quite shocked at the behavior..

Now I know that the strength of the Prism is in the sound, but quite frankly as an Audio interface, IMO its not even on the radar.. :icon_no:

Anyone else with some experience to share.. ?

[shelter]

Never used one myself, can you post the manufacturers link?

BTW I never received a reply to the PM I sent you.

[TAFKAT]

Never used one myself, can you post the manufacturers link?

BTW I never received a reply to the PM I sent you.

Prism Sound Orpheus (http://www.prismsound.com/music_recording/products_subs/orpheus/orpheus_home.php)

Re the PM, hmm, let me check, must have missed it..

[shelter]

OK had a look at the Orpheus spec and pricing. I think I'll stay with my
Neve's. At $1100 a channel it's not what I'd expect.

[TAFKAT]

Part II ...

I came back to the system configuration this morning and tried a few other settings to no avail, I then installed a PCie Firewire card, TI chipset of course, no change in behaviour. Swapped out the cable as well. Nadda..

Quickly installed a RME HDSP card to check the system was performing normally to remove that variable, all working fine.

I then installed Nuendo 4 as well on the system as the client has requested both be configured. Check RME all fine. Swap to Prism, and try and set VST connections, no go. Quit N4, start N3 , reporting driver can not be initialised..

Power system down , double check cable connections, cycle Prism Off / On, restart system , hear the definitive click as driver is initialised as Windows boots, all good.

Launch N3 , check/assign VST connections all good, fire Demo and Stress sessions, purring , no hanging, pausing, sluggishness as before . Exit N3, launch N4 , double check VST connections , all fine, launch Demo and Stress sessions , purring. Right..., WTF moment again.

Lastly , swap back to original cable to come full circle, all purring..

Only change was the installation of N4 and re-seating the cable.. :eusa_think:

Adjust the Buffer setting in Control panel to 1/2 the original setting from 5000 to 2500 giving me latency setting increments per 2.5 ms, lowest buffer now 5 MS - Reporting 6.531 / 11.429 : System Purring..

Smiles, throws hands up, points arse to moon , stands on one leg and recites Stairway to Heaven backwards while doing this... :willy_nilly:

I am stumped , but for now I'll blame it on the moon..., and in hindsight, more than likely a dicky firewire connection :D

So whats the verdict on the driver now ?

Hard to compare with other low latency Firewire device like the RME Firfeace for example , simply because the input/output latency on this unit is configured so differently, so setting a 5MS buffer which is close to 256 on the RME will result in similar input latencies, but playback is almost double, which is closer to 512.

I did run up DAWbench DSP just for the Hell of it, resulting at around 135 RXC instances before crackles appeared, and that was only at around 80% CPU /ASIO loadings, both reading the same value which is not uncommon at the higher latency settings. This highlights one of the reasons I find measuring scalability with meters useless , as by the meter we have a good 20% overhead, but the system is already breaking up..

I also installed and loaded a UAD2 Duo card, and loaded that fully without any ASIO munching evident, so the system is about ready to go to its new home, and I can move onto the next adventure ... :phew:

[TAFKAT]

Part III ..

Well this saga has taken another twist...

After shipping the Orpheus unit back to the client , on reconnecting it to his Audio Notebook that had performed flawlessly previously , the same performance issues that I had reported in my first post when running N3 were now being experienced .

But only in Nuendo, other applications performed normally.

Well at least I wasn't going mad , sadly it looks like the reverse recital didn't hold... :eusa_whistle:

Question now remains what exactly is going on with the unit ?

We are starting to suspect that the firmware upgrade may be a mitigating factor, as prior to the firmware update the unit was rock solid according to the client.

I'll stay on the case and report back when we get a final resolution.