This review page is supported in part by the sponsors whose ad banners are displayed below

A quick forum scan showed that first Isabella owners had previously owned the Dodd; had considered the ModWright LS 36.5 (the implied preference for the Isabella was shared by an Asian distributor who also compared it to the 36.5); and had auditioned the Audio Horizons TP2.1 and Eastern Electric MiniMax before deciding. For the standalone Isabellina, the first owner on record had, over a 2-year period, owned or listened to the Channel Islands VDA2, Benchmark DAC1, Zhaolu 2.5, Bel Canto S300iu (integrated amplifier with DAC input), Electrocompaniet ECD-1, Citypulse DA 7.2, Scott Nixon JFET NOS USB DAC, iRock NOS USB DAC and AudioSector/AudioZone NOS USB DAC before declaring the Red Wine Audio piece superior to the AudioSector, his prior favorite.

This demoed very competitive spirit against known performers. Vinnie meanwhile explained why he believes his non-oversampling Isabellina converter doesn't suffer the soft bottom and top traditionally associated with the zero-sampling breed: "We are able to offer a very low output impedance for the Isabellina because we are using a discrete transistor class A output stage. Some NOS converters use a passive I/V stage (this results in high output impedance). Tubes also generally have higher output impedance (e.g. Wavelength). I've seen opamps used as well but once we tried using a discrete transistor stage run in Class A, we found it to be the best with our NOS solution." The Isabellina's output impedance is 100 ohms, with an output voltage of close to 2. Vinnie then added: "Between you and me, I have extensively heard both the Dodd and Modwright (as well as a few other preamps in the price range and some beyond) as I was in the development phase of the Isabella and though I'm biased of course, I have to agree with the early commentators. While I was disappointed that the Isabella took so long to finalize, I now am glad that I held out and got it to sound exactly how I wanted. I'm sure I lost sales because people didn't want to wait any longer but I thought it was better to wait and get it just right than to come out with it early and then offer 'new upgrades' in a few months."

Vinnie also offered more insights into how he views digital in general: "It seems the conventional wisdom is to make digital sound more like analog by resampling to 24/192 to spread out all the quantization noise, jitter etc. In general, oversampling/upsampling and the associated digital filtering can yield excellent S/N specs and other impressive specifications on paper; but I haven't found one to sound musical and 'analog like'. They sound 'digital' (not necessarily a bad thing and 'digital' does not always means 'harsh' or 'edgy') but they do not sound like an analog source. The only DACs I have found to approach this are the 16-bit, non-oversampling, digital-filterless iterations that implement R-2R (resistive ladder) conversion. This is what we are using in our Isabella and Isabellina. These types of conversion chips are no longer made so we hunt down new-old-stock. But the DAC chip used is not the whole story. The digital input stage (receiver), analog output stage and power supply of the conversion progress are all essential to the sound. We carefully designed the Isabella and Isabellina to have extraordinary synergy with our amplifiers (30.2 and 70.2). Our amplifiers do not use the Tripath 2024 chip. We actually use two chips (these are located under the board so you do not see them when you open the enclosure); the Tripath TC2001 (input stage) and an output driver chip by ST Micro. I custom-designed the input and output stages (you won't find anything close to this in the data sheets) and this, along with the high-current SLA battery power supply, has much to do with the final sound."

Back to the Isabella. In use, her touch-sensitive membrane power switch requires a mere short push to fire up, a longer one to power down - with mild transients each. A red LED between the 6922s lights up immediately to bathe the area beneath the rail-guided sliding glass window in liquid blood. 40 seconds later, there is sound and that red LED extinguishes. Switching between battery and AC power is instantaneous. This allows useful comparisons. Toggling the gain switch does likewise. Even the shortest tiniest prompt on the remote rotates the volume up or down by the same repeatable 5mm. In theory, you'd think that to get in-between the non-step steps, you'd want to adjust manually. In practice, the remote is likely all you need if you've chosen the proper gain setting.


Syncing Vinnie's MacBook to the Isabella couldn't have been easier. Fire up the MacBook, click on the infamous apple in the tool bar's extreme left, select System Preferences, then Sound, then USB DAC rather than Internal Speakers. C'est ca. To rip your CDs to Apple Lossless, go into iTunes, select Preferences, then Advanced, then Importing. Now select Apple Lossless encoder. Set it to automatic and check 'auto retrieve CD track names from internet'; 'create file names with track numbers'; and 'use error correction when importing CDs'. Set the iTunes volume control to max. Then go into iTunes, Preferences, Playback and turn off Sound Enhancer and Sound Check to not compromise the outgoing data. Also, turn off the equalizer under View, Show Equalizer.


Naturally, to get CD artwork and track titles, the MacBook must be connected to the Internet. It will import CDs offline but they'll be filed simply as numbered tracks and you'll have to input the meta data by hand. The li'l white MacBook remote makes using a laptop as a music source far more 'non-PC' than anti mouse/pad/keyboard types would assume.

The inbuilt DAC option
Comparing the Mullard-fitted analog RCA outputs of my Raysonic Audio CD-168 to linking its digital RCA output via Stealth Sextet to the Isabella's digital BNC input showed equality (competitive performance) but not sameness. The differences operated in the areas you'd expect between upsampling and filter-less converter designs which sell in a similar price class. The Raysonic had more metal inside Lila Down's heavily modulating rancheras voice of Una Sangre. Its bass was even more forceful. Image outlines were more highlighted or crispified as well. Though you'd never accuse the Sino Canadian player of etch, there was a smidgeon of it compared to the Isabella's own converter. Detail freaks might gravitate to the CD-168, believing it to be a tad more resolved. Others would demure, pointing at the Isabella as just as resolving, albeit without the subtle sheen the Raysonic lays on. Equality but not sameness. On well-recorded cymbals [various ECM albums], the CD-168 had more swish and sparkle but not to extent the Red Wine's non-oversampling nature would have predicted. Quantitatively, the Raysonic mostly seemed a bit more endowed down low while the Isabellina DAC had a more organic feel - like a fine brochure image without any applied gloss layer.


Substituting the Raysonic CD-168 with the Opera Audio Reference 2.2 Linear player resulted in nearly perfectly matched equality. Like the Isabellina converter inside the trimmed-out Isabella, the Linear is a Kusonoki-style design without digital filtering which then runs a tube buffer on its outputs. Having also reviewed its massive kidney-shaped Droplet brother, I'd in the end come down in favor of the cheaper Linear. Rhythmic fidelity -- timing snap and tap -- was higher. The upshot here was that a MacBook + Isabellina DAC get you not just into the general ballpark but become virtual stand-ins for Opera Audio's best NOS machine. It was to get better still...


The next comparisons involved the digital outputs of my Ancient Audio Lektor Prime and Esoteric/APL HiFi UX-1 carried again on the Stealth cable; vs. Isabella's silver-braided USB cable plugged into Vinnie's loaner MacBook. In other words, I compared data extraction. One was traditional, using the two best transport mechanisms on the market - Philips Pro-2 in the Ancient Audio machine; Esoteric's biggest heaviest VRDS sled in the Peychev-modified universal machine. The other was PC, ripping the same CDs to a MacBook's HD via Apple Lossless, then retrieving those CD files magnetically over the USB interface. Important to mention too is that said interface juxtaposed a very expensive Stealth Audio S/PDIF cable vs. Vinnie's silver-braided USB link; and that all the two expensive dedicated audio extractor drives had as an opponent was whatever cheap 'n' cheerful slot-grabbing computer drive Apple puts into its MacBook. In short, expensive drives + expensive cable doing traditional CD spinning and laser extraction in one corner; cheap drive + cheap USB cable doing magnetic streaming in the other.

Here it got very interesting. Magnetic playback via USB sounded more open, immediate and direct than laser retrieval over the Philips top loader. While top digital in many cases has gotten close enough to involve hair splitting during comparisons, this difference was audible enough. I had an unambiguous preference for the MacBook which removed a layer of intermediate hesitation and distance. Some film of betweeniness disappeared which reinserted itself promptly each time I switched Isabella inputs back to S/PDIF. Granted, by the time the massive Esoteric sled entered the picture, I could no longer tell whether I was hearing USB or S/PDIF - but that's $6,000 OEM for the raw transport assembly when ordered in lots of 50.*

______________________
*While definitely intended, the obvious comparisons -- between MacBook/USB and the analog outputs of the Prime and UX-1 -- couldn't be made. The Prime suffered a bad channel regardless of 6H30 to suggest a defective part, likely a capacitor; and the UX-1 retaliated with a dead channel. Both machines need fixing but worked swell as transports.


The "got interesting" paragraph reads too cut 'n' dry not to warrant highlighting what it means: an $1,100 MacBook and quality USB cable outperformed a Philips CDM-PRO2-fitted machine with ultra-quality digital cable as data extractor. Gone are the days when using a computer implied automatic inferiority to quality digital hifi components. Nowadays the PC -- or Mac in this case -- can come out ahead. Talking cash, a digital transport running the Philips Pro mechanism should set you back at least twice what a suitable laptop-as-server will. Yet dedicated transports with the Pro drive can demand up to $20,000. The lowly computer platform's financial victory can be a lot sweeter still. Now add the savings of a USB vs. top S/PDIF cable. More sugar. Who's your daddy now? If you're new to the hifi game and don't already own a quality CD player, I see no justification or rationale for buying one. Unless you're completely immune to the music-server conveniences that is. Did I already mention that Apple's tiny remote really makes interacting with a laptop on the shelf usually occupied by a CD player exceedingly graceful for us old-fashioned 2-channel dinosaurs?

Taking da(c) stock. When it comes to his three digital inputs, Vinnie is sanely correct to focus our attention on the USB socket. It's no coincidence that it's 1 on the input selector dial. While you can equal it by preceding the BNC socket with a gawdfully expensive CD transport, I very much doubt you can exceed it. Apple's data extraction, the Apple Lossless codec and Vinnie's USB receiver circuitry are simply too well done. The smart money thus parks all its bets on USB.


The Isabellina DAC is squarely competitive with what's in the Raysonic CD-168 and Opera Audio Reference 2.2 Linear players. So at $1,500 for the inbuilt Isabellina option, that's how I'd recommend spending a $2,500 CDP budget: a grand on the MacBook, the balance on the Isabellina. If you had twice to spend and did not care about the iTunes conveniences, you might come out ahead sonically if you linked a traditional CD player to the Isabella's analog inputs. Whether you really will, by how much; and how much it would cost; I can't assess until my two reference players are repaired.


About higher-rez files: "You can feed the Isabellina DAC a digital signal that was sampled higher than 44.1kHz. For example, even if the music was sampled at 96k or 192k, it will lock onto it and use that sampling frequency. It does not oversample, however, so if you send in 44.1k, you get 44.1k out (not some integer multiple of this sampling frequency). With regards to word lengths, if the files were 18, 20 or 24 bits, it will truncate to 16 bits. So it always plays back in 16 bits but the sampling frequency depends on the music file." In other words, upsampling transports like the Ensemble can interface with the Isabellina converter. My Zanden Model 5000s NOS converter would not, presumably because of its choice of receiver chip. "The receiver chip locks the incoming S/PDIF signal which, in today's case, converts it to I-squared-S and then feeds the digital-to-analog converter chip. The USB stream undergoes no S/PDIF conversion at all."