Re: ESI UGM96 soundcard mini-review
- I agree, I remember reading somewhere that this thing can't do 24 bits at 96kHz due to USB bandwidth. Also, it costs $90 USD, the lowest I can find at the moment.
Have you been able to take in 24-bit at 96kHz in WinXP?
--- In email@example.com, Sid Boyce <sboyce@...> wrote:
> On 02/01/10 16:53, Rob wrote:
> > Hi Group,
> > Now I have two working SR systems, and an understanding of PowerSDR-IQ,
> > thought I would try different sound cards. Here's a look at the UGM96,
> > using an Edirol FA66 as a yardstick.
> > The advantage of the UGM96 hits you when you realise how small and cheap
> > it is, see http://www.esi-audio.com/products/ugm96/
> > <http://www.esi-audio.com/products/ugm96/>
> > The test system was:
> > Samsung Q320 laptop with Windows XP SP3
> > PowerSDR-IQ V1.12.20
> > Softrock RxTx V6.3 with MoBo V4.3 beta
> > Alinco DM330 power supply (switch mode)
> > Unlike the multi channel Edirol, the UGM96 has only a stereo in and out.
> > So I had to use my PC for the sound output/input and have the VAC tab
> > active in PSDRIQ.
> > It took some fiddling with settings to get the UGM96 working smoothly. I
> > am using the V2.3 windows driver from ESI, which is newer than the one
> > on the install CD. The buffer size in the ESI system-tray panel needs to
> > be increased to 256. In PSDRIQ the latency was set to 1024, with the
> > ASIO driver.
> > PSDRIQ gives a stern warning about not using an approved card, but apart
> > from having to fiddle with VAC, it's not a show stopper.
> > The patchbay software supplied by ESI is confusing. There is still no
> > documentation on how to use it. Leaving it "unpatched" seemed to work.
> > I noted no big difference on Rx or Tx performance between the two cards.
> > Both cards give spurious spikes in the panadaptor. The zero spike of the
> > UGM96 was bigger, but the wideband spikes smaller. I usually strap the
> > FA66 to the Softrock chassis to reduce the zero spike anyway. The Alinco
> > power supply causes some spurious "humps" but not too bad.
> > The input op-amps on the UGM96 are JRC4580, which are not really high
> > performance chips. They could be replaced by alternatives.
> > The CPU load with the UGM96 was higher, but shouldn't be a problem on a
> > reasonably specced PC. The main operational issue is the buffer size.
> > When tuning, there's a slight delay between hitting a frequency and the
> > audio coming through.
> > There's a picture of my SR Mk.2 with the UGM96 attached in the photos
> > section.
> > Conclusion: UGM96 plays well with PSDRIQ, within the limitations of a
> > two channel, unsupported card. For base station use, the FA66 has the
> > edge. For portable use... I will fit the UGM96 inside my SR, making an
> > integrated USB-in, antenna-out box.
> > 73s,
> > M0RZF
> Reviews elsewhere suggests that it would only do 24-bits @ 48K, or
> 16-bits @ 96K.
> Which of those 2 are you using?
> 73 ... Sid.
> Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
> Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
> Specialist, Cricket Coach
> Microsoft Windows Free Zone - Linux used for all Computing Tasks
- Hi Gents,
My result was rather unscientific, based on comparing the highest and lowest signals, and overall listening perfomance against the FA66. Since few musicians use Linux, it's hardly a surprise that the makers target their soundcards at Mac/Windows.
I can't see much difference in the noise floor when using 48k sampling, yet the USB can't quite handle duplex at 96k/24 bits.
The ESI ASIO driver works as well with PSDR as the Edirol one. But I have no way to verify the bit depth, apart from looking at the noise levels. Even with 48k@24bit it still would use useful. Please don't quote US $ prices, they make us jealous!!
Graeme, on the FA66 comment - most of the sockets on the FA66 aren't grounded to the metalwork. The zero spike drops dramatically if I strap the case (grounding screw provided) to the Softrock case. BUT the broadband spikes increase.
I need to look again at the area of soundcard spurious. The spikes are below noise on 40/30m but will be a problem on the higher bands.
> I agree, I remember reading somewhere that this thing can't do 24 bits at 96kHz due to USB bandwidth. Also, it costs $90 USD, the lowest I can find at the moment.
> Have you been able to take in 24-bit at 96kHz in WinXP?