Which cheap USB or FireWire audio interface for Dell E6400 notebook?

Globulus,
For my interface (USB, with audio and MIDI - the M-Audio Fast Track Ultra), for an ASIO buffer size of 128, my round-trip latency is 9.8ms. I can use Pianoteq at this ASIO buffer size, and the latency is fine. The MIDI part of this interface has a total round-trip latency of 12ms (for note-on and note-off events), and it appears to introduce about half this for normal playing, when I compare it to a MIDI port that is integrated on my motherboard. Despite this, my latency is fine using the USB for MIDI, at that ASIO buffer size.  Increasing my ASIO size fo 256, the LTU result is 15.9ms.

Perhaps just just pick a store that has a good range, and that has a good return/exchange policy. I kind of suspect that any USB or Firewire product that is working properly will be fine, and I wouldn't worry too much about test results.  I had a problem with my device, but it was only a problem that showed up in measurements - I hadn't noticed any problems just playing it normally. 
But yes, a PCI, PCI/e, PCMCIA or ExpressCard adaptor will probably give you the best possible latency.  I doubt very much whether an ExpressCard to PCMCIA adaptor would introduce any noticable latency - if it works properly, the latency will still be excellent. (just my suspicion - no experience with those adaptors at all)

It appears that ExpressCard to PCI adaptors are also available, e.g: http://www.magma.com/store/expressbox1.html
EDIT: Yikes, that thing is EXPENSIVE!!!

Greg.

firewire is the better option Globulus! Focusrite Sarrire Pro 24 is what I use! It has the famous Focusrite pre-amps to boot! Here's the link:

http://www.focusrite.com/products/saffi...re_pro_24/

hope this helps?

Chris

Update:

I bought an Edirol UA-25EX today:
http://www.rolandus.com/products/produc...ductId=970

This was the only make/model my (tiny and only) local shop had in stock, but the price was sufficiently competitive so I just went there and picked one up. Note that this is the new version from 2008.

I had read too many bad reports about driver problems with the M-Audio Fast Track Pro/Ultra devices but it was difficult to find anyone saying anything bad about the Edirol.

I have played around with it for an hour or so now. I set the host sample rate to 96khz and the buffer size to the second-lowest value, giving me an ASIO latency of 4.7ms.

I have the audio interface as well as my RD700SX hooked up to a Mackie VLZ Pro mixer so I can quickly switch between Pianoteq and the internal sounds of the RD700SX. In terms of latency, I cannot tell the difference between Pianoteq and the internal sounds. If I layer both of them (unmuting both on the mixer), I certainly cannot hear a double attack. When switching the host to 44.1khz or 48khz, ASIO latency becomes 7-9ms and I get the impression that there is a slight (but not annoying) delay compared to when using the internal sounds, but again, if I layer Pianoteq and the internal sounds with those settings, I cannot hear a double attack.

Even at 96khz (Pianoteq's internal sample rate set to 48khz) and the next-to-minimum sample rate, I cannot cause Pianoteq to skip a beat.

I do get the occasional glitch (maybe a few times a minute), but I blame my Dell E6400 drivers for this. DPC Latency Checker also displays the occasional nasty red spike, indicating that I might want to look at the driver configuration. I am certainly not going to return the Edirol on account of this. I know what crackles caused by too small buffers sound like and this is not it. I disconnected the USB cable (used for midi) between my RD700SX and the notebook and switched to using a 5-pin DIN midi cable between the RD700SX and the Edirol interface and this seems to have decreased the frequency of those spikes, but that could just be my imagination.

I will try it out on my desktop pc and check if I am getting similar spikes there, but I doubt it.

So far I do not have anything bad to say about the interface. Very quiet. No noise. No hum. Built like a tank. Vista and Windows 7 drivers available, should I ever decide to upgrade my OS. Balanced outputs. Expensive if you are just looking at the features/connections (of which it has very few), but in this case, you are paying for quality. The XP drivers are still at 1.0.0 which may be a reflection of the fact that this is a very simple *hardware* device with very little in the way of software control, which might have made things easier for the software guys. As an example, to change the sample rate, you have to disconnect the device, move a 4-position (44.1khz, 48khz, 96khz playback, 96khz record) switch on the back, then reconnect the device and restart your application(s). As I bought the interface for one purpose only, I am okay with this set-it-and-forget-it hard button approach.

@skip:

The E6500 is the same as the E6400, albeit with a 15.4" screen instead of 14.1".

As usual, there might, however, be differences in the selection of display resolution, graphics card, wireless network card, cpu, and so on.

I have a T9400 cpu, nVidia gpu, and Dell (Broadcom) WLAN card.

The main discussion about audio glitches due to DPC latency problems would appear to be this one (15 pages at the time of writing this):
http://forum.notebookreview.com/showthread.php?t=341918

As you can see, opinions differ as to what might be the cause. The answer is probably that there are multiple possibilities.

In my case, disabling the WLAN card in the Device Manager removed most of the dropouts, though I would still get the occasional (as in a few every minute) spikes when monitoring using DPC Latency Checker.

I was still using a pretty old driver and the old Dell ControlPoint software to manage my wifi connections. Updating the driver and letting Windows manage those connections pretty much solved everything.

DPC Latency Checker would now stay below 500us but frequently hit the 500us mark and on rare occasions produce a yellow spike, but none of the enormous red spikes.

By switching the Edirol interface from 96khz to 48khz I have found that DPC stays in the 200-400us range, only occasionally hitting 500us. I have not seen a yellow or red spike since switching to 48khz.

ASIO4All (2.9 Beta 1) does not work with the Edirol in 96khz mode. The native Edirol ASIO drivers only allow me to go down to 448 samples (4.7ms) in 96khz mode or 112 samples (2.33ms) in 48khz mode.

Using ASIO Latency Test Utility 3.7 I get the following results:

Edirol native ASIO driver:

48.0 khz: 112 samples/2.33 ms: 460 samples/9.58 ms
                224 samples/4.67 ms: 684 samples/14.25 ms
                336 samples/7.00 ms: 908 samples/18.92 ms
                ...

96.0 khz: 448 samples/4.7 ms
                896 samples/9.3 ms
                1344 samples/14.0 ms
                ...

ASIO4All 2.9 Beta 1:

44.1 khz: 64 samples/1.45 ms: 377 samples/9.58 ms
48.0 khz: 64 samples/1.33 ms: 364 samples/7.58 ms

Running Pianoteq at 48khz (internal sampling + host) with a buffer of 64 samples I now get glitch-free playback. Though: If I set polyphony to 256, hold down the sustain pedal, and torture the keyboard, I do get the occasional crackle, so I have lowered polyphony to 64.

I guess I will be running the host in 48khz mode, using the ASIO4All driver, but if Pianoteq does not use more than 48khz internally, I see no problems in that.

I do not know what the midi-to-playback latency is, but with 48khz/64 samples I cannot tell *any* difference in the latency between using the internal sounds of my RD700SX and Pianoteq through the Edirol interface. I won't be using it for voice recording or anything like that, so I am really not concerned about the round-trip latency reported by LTU.

Aside from the limitations I have mentioned (and the little blue power LED that will burn your eyes out), the Edirol interface seems to be a solid, high-quality performer. As for the Dell notebook, it was designed and meant for business use and anyone considering buying it for use as a DAW should check the relevant forums before clicking the Buy Now button.

FWIW: Running the DPC checker for 5 minutes, idle (with no soft-synth open or any ASIO activity at all), mine peaked at 415uS, which is just under the green line.

Config: Windows XP SP3, Dell Latitude E6500, P8600 Core 2 Duo @2.4GHz (locked on maxiumum speed), 4GB, Samsung HM251JJ internal SATA drive, Nvidia Quadro NVS 160M graphics, Intel Wi-Fi 5300 AGN.

It appears that the Windows Performance Toolkit is generally advised to diagnose DPC issues, so I may give this a shot to see if I can improve my latency at all.  (a bit academic though because I'm not experiencing any problems)

I have run the DPC checker with Pianoteq active, using ASIO4ALL and the internal IDT audio interface, and this does produce high latencies into the red. I increased the ASIO buffer size, which improved the DPC latency, however I experienced glitches. When I closed the DPC checker, these glitches went away.  So, it looks like this DPC checker itself can change the behaviour of the system. ;^)  (I wonder whether it's just the graphics in the DPC checker)

I have not yet run the DPC checker with Pianoteq & the Fast Track Ultra active.

Globulus: if you had the E6500, you'd have BOTH a PCMCIA and ExpressCard slot, rather than just one or the other. ;^)

Greg.

Thanks again! Yes, I think I'll do some actual piano playing before I get around to analysing my green bars. ;^)

RE: the PCMCIA & ExpressPort, it appears that you can choose which one you have on the E6400. I wonder whether you can swap it around afterwards? Not important for you now since you have the Edirol I suppose.

My E6500 is one of those too-good-to-be-true Ebay specials - I'm grateful the damn thing even boots. ;^)

Greg.