IBM T-42 DAW, Edirol UA-25, RME Fireface 800 review

Hi, I’m setting up a new DAW system; the following information may be useful:

Previous sound system: Mac-based ProTools 24/888, Cubase, etc. (our sound guy’s choice). Currently putting together a PC-based system (relatively low cost, greater portability).

Software:

Calkwalk 4.0.1 PE
Native Instruments Komplete 2
Garritan Personal Orchestra
Sound Forge 7
Audition 1.5

Hardware:

Kurzweil K2500XS (all ROMS), from previous system. The NI and Garritan instruments may make this component into a basic keyboard controller.
Mackie HR824’s, from previous system (also used Event 20/20bas in previous system).
Mackie 1402VLZ, from previous system (also used 1202VLZ in previous system).

Notebook-based DAW:

IBM T-42 2378DXU (1400x1050 display, ATI Radeon 9600 Mobility), 1.7GHz Centrino, 1.5GB RAM, 7200RPM 60GB drive (250GB 7200RPM WD FW external), 18" 1280x1024 flat panel as 2nd monitor.
WinXP SP1 (won’t be installing SP2).
SIIG Cardbus FW800 (TI Chipset based): http://www.siig.com/product.asp?pid=392 , $65 at buy.com. There is an issue with the SIIG card and IBM Thinkpads (might affect other similar products), fortunately, there's info here for a simple BIOS change fix: http://www.siig.com/faq.asp?faqid=10050934

It appears that (some) cardbus firewire (800) cards do not supply enough power to run bus-powered firewire devices. An extra 12v power adaptor is required, which plugs into the side of the cardbus card. This is the case with the SIIG FW800 card. The Lacie FW800 card comes with an external power supply (and costs slightly more): http://www.lacie.com/products/product.htm?pid=10482 . Something to keep in mind if one is intending to use a FW bus powered device and must also install a FW cardbus card.

Tested audio hardware:

Audio quality test for A/B comparisons: Reaktor 4, running Leviathan23 (user download) and BlueMatrix (default), both using startup default settings. These ensembles repeat fairly clear, clean, and subtle sound.

1. Native SoundMax audio: better than expected quality going through Mackie mixer to HR824’s and listening through Sennheiser HD580 headphones.
2. Edirol UA-25: USB bus powered, works on a 1.1 or later bus (2.0 wouldn’t make it any faster (latency) as it only supports stereo in+out). Sound quality is better than built-in SoundMax (24 bit versus 16 bit probably helps), but not dramatically better (headphones). On speakers, the difference was more subtle. The UA-25 is a cool little low-cost recording interface/soundcard + MIDI I/O. Does not get very hot during operation (efficient). Unfortunately, my unit had noise problems on the line outputs (headphone out was OK).
3. RME Fireface 800: did not perform an A/B comparison, but sounds similar to UA-25 (both parts running at 44.1, 24 bit, providing noise free, clean output). I don’t need all the I/O’s (just doing sound effects and videogame background music), but the Fireface 800 came highly recommended as having fast and stable drivers.
4. On the way: Creative Audigy 2 ZS Notebook ($117). While it may not really be a 24 bit part (it’s spec’d at 192kHz, 24 bit), and is a consumer product (with consumer drivers), it provides digital and analog I/O, 7.1 output, etc., and may work OK with headphones when using the laptop standalone. We’ll see if it can provide lower latency and higher sound quality than the native SoundMax hardware (it supports ASIO2.0 and is advertised as low latency). I’ve read reports of problems with the ENE chipset cardbus controller (the IBM T-42 uses a TI chipset). For a review and more info:
http://www.extremetech.com/article2/0,1558,1684450,00.asp http://discuss.extremetech.com/n/main.asp?webtag=extremetech&nav=start&msg=64950.

Benchmarks using the Sonar3Test.cwp from this site:

Other than giving background processes higher priority, I have not performed any special tweaks/optimizations to this notebook (49 processes loaded).

Native SoundMax audio:
10ms (lowest possible), 60% CPU (WDM/KS 44100, 16 bit)

Edirol UA-25, USB 1.1
3ms (lowest possible), 70% CPU (WDM/KS or ASIO, 44100, 24 bit)

RME Fireface 800 (same results using FW 400 or FW 800 cable), ASIO:
23.2ms, 33%
17.4ms, 35%
11.6ms, 36%
5.8ms, 40%
2.9ms, 51%
2.2ms, 57%
1.5ms, 75%

WDM/KS

1.5ms (lowest), 78+%. Note: the ASIO control panel buffer size affects WDM/KS latency as well.

MME/32 bit

Cakewalk 4.0.1 could not open driver, however NI Reaktor 4 (for example) was able to open without issue. The RME documentation leads one to believe that their trick MME drivers are better/faster than the WDM kernel streaming drivers. However, it would appear on this IBM T-42 laptop that ASIO is the way to go.

More Fireface Info

After switching from ASIO/MME, the RME drivers have acted strangely (and prior to the latest drivers, generated a BSOD). The RME FW 800 has many setup/mixing options which is both an advantage and disadvantage. The UI-based mixer is not intuitive to use (some options require CTRL, SHIFT, ALT, or combinations). After reading through the manual, basic settings are easy (such as sending all inputs to one output). A wrong mouse click can set the output to 0dB (100% max volume). There is no master volume control for the outputs: you may want a line output attenuator or mixer in line with the speakers. I had planned on eliminating the mixer from this setup… Perhaps RME can provide a firmware update (FPGA hardware design) so that the front headphone volume can also run the line outs (similar to the Control Room/Phones output on the Mackie mixers). If using just two speakers and no headphones, a stereo-phono splitter can be used on the headphone output to connect directly to a pair of speakers. This would provide a volume control of the single output (via an unbalanced connection).

This unit also runs a little warmer than I’d like to see (power efficiency, longevity/reliability. However, it has a two year warranty, which is better than many other similar products). This unit ships with a fairly long FW 400 cable. I purchased a FW800 cable at Fry’s (9-9pin 10' cable, PPA Inc., #3893. $45). So far, I have not seen a difference in behavior/performance between FW400/800.


I’ll post an update for the Creative card once it arrives.

ORIGINAL: daverich

have to say my fireface arrived today and i'm shocked at the performance.

It really is incredible - I've not seen the dreaded UAD error message once today yet either!!!

Kind regards

Dave Rich.

Cool. What were you using before (sound quality comparison)? Do you like the Mixer (have you tried routing multiple ins to multiple outs)? Are you running a hardware mixer/attenuator, or straight to speakers?

Thanks,

John

I’ve read reports of problems with the ENE chipset cardbus controller (the IBM T-42 uses a TI chipset). For a review and more info:
http://www.extremetech.com/article2/0,1558,1684450,00.asp http://discuss.extremetech.com/n/main.asp?webtag=extremetech&nav=start&msg=64950.

For anyone having Cardbus problems with the ENE CB1410 Controller download and install this driver:

http://h18007.www1.hp.com/support/files/hpcpqnk/us/download/18388.html

I had a friend with an Averatec 6200 notebook that was having horrible problems running an RME FF800 or a FirePOD via a Lacie FW800 Cardbus PCMCIA card. "Serious distortion artifacts" was how someone experiencing a similar configuration problem described it. His Device Manager reported a "Generic Carbus Controller", but after careful searching we found the above driver for the ENE CB1410 controller that was actually in the computer.

After installing this driver Sonar produced pristine audio right out of the gate. We thought this notebook had turned out to be a really bad choice and were ready to dump it. After the driver update it now performs flawlessly.

Apparently, many people have had all kinds of audio nightmares with the ENE CB1410 controller and some people may be having problems and not even realize that this is what it is. Hopefully, this driver will solve a lot of problems for people in similar situations.