[tangent]

Over the past few months, I've bought a stereo-link USB DAC, a standalone DAC, a standalone ADC/DAC, and a prosumer-grade sound card that can be pressed into service as an ADC/DAC. I decided to run alls these DACs off against each other, with a focus on using them to improve the sound quality of low-end digital sound systems.

The first DAC I bought was the Creek OBH-14. This little DAC/preamp combo uses the same basic enclosure as the familiar Creek OBH-11 headphone amplifier. The attenuator knob is also identical between my two units. The OBH-14's preamp functionality gives you an analog and a digital input, the latter in both S/PDIF and Toslink flavors. The OBH-14 level matches the digital and analog inputs nicely, so if you hook both up to a CD player, you can toggle between the OBH-14's DAC and the CDP's built-in one with a button press. The OBH-14 has two analog outputs, one at line level and the other controlled by the attenuator on the front of the unit; cranked all the way up, the two outputs match in level. I bought this DAC/preamp to improve my main stereo system's sound and flexibility, a job it has performed admirably.

Then, a month ago, I bought a stereo-link USB DAC for listening to MP3s at work. (See this thread for my review of the stereo-link's headphone section.)

Most recently, I bought an ART DI/O and an M-Audio Audiophile 2496 sound card, to improve my home recording capabilities. I'm using the DI/O mostly as an ADC, to feed the Audiophile card's S/PDIF input because I get better results than when I let the sound card do its own sampling.

The Test

I tested the DACs alternately hooked up my main stereo system and to my best headphone system. The main stereo is a Cambridge Audio A300 integrated amp driving a pair of Sound Dynamics RTS-3s through AudioQuest Type 2+ speaker cables ($25/meter). The headamp setup is a Corda HA-1 driving Sennheiser HD-580s.

My main source was high-bitrate MP3s played through the Audiophile 2496, except in the case of the stereo-link which connected to the computer in its own way. As a side test, I hooked the two standalone DACs up to my Pioneer DV-434 to make sure the MP3 quality was indistinguishable from the original CD.

For the tests with the main stereo, I used a 20' length of Radio Shack Gold Series cable to get from each DAC to the integrated amp, since my computer is in a different room from the stereo. (I told you this was a low-end test!

) This test wasn't important enough to me to move my PC into the next room or buy a 20' pair of high-end interconnects.

My digital cable of choice was a Hosa 10' S/PDIF cable. This is cheap professional wire, not fancy audiophile wire. Many people report problems with the ART DI/O dropping out when used with inferior digital cables, but I had no problems with this $10 cable. (FWIW, the construction is remarkably similar to the thicker variety of Radio Shack Gold Series wire, except that only the contact portion of the RCA termination was gold plated -- the grip part was left with a nickel finish, which is of course perfectly reasonable.)

For test material, I chose Loreena McKennitt's "Book of Secrets", a drum-heavy work with good dynamics and several parts that tend to show up problems in playback equipment.

How'd It Sound, Jack?

I tested the stereo-link first, though my main stereo system. My immediate impression was of fairly uncontrolled, boomy bass. This is forgivable considering the price of the unit ($130). I also didn't see that problem through the Corda/Senn combo. I suppose the headphone system kept tighter control over the sound than my mid-fi stereo system was able to.

Next I tested the OBH-14 and the ART DI/O. I saw very little difference between these two DACs in my system. The only significant difference I found is that the ART DI/O showed some "texture" in some bass drums on one track, and then only when hooked up to the stereo system. I have no idea if this texture is in the recording or if the DI/O took some liberties with the audio.

Finally I tried the Audiophile card's analog outputs. The main thing I found is that it has all the detail of the standalone DACs, but none of the smooth musical power. The output seems clinical, simply playing back the audio in a very literal fashion. The sound card doesn't damage the sound, really, it just lacks the qualities expected in an audiophile stereo component. For playing back music for entertainment to decent speakers or a good headphone setup, it's better to use an external DAC than to use the on-board one. If you're just using typical PC speakers or you're using the card in a studio setting, the on-board DAC/ADC is probably fine.

Another thing I noticed about the Audiophile is that it has fairly low-level outputs, like most portable players and most other sound cards. I had to turn the Corda up to about 3 o'clock or so to get the same volume level that the OBH-14 and the stereo-link were putting out at 12 o'clock.

ART DI/O Notes

My DI/O is completely untweaked: the power supply is stock, I did not terminate the digital output, I'm using cheap 6' Rat Shack 1/4" to RCA cables, and I left the tube in since I use the DI/O for ADC. I will be trying several tweaks soon, but I didn't want to delay this review. I did try tweaking the inputs and outputs with terminators, giving the ADC path a live load while playing back, etc., and nothing changed the sound. I consider this to be a feature.

The DI/O analog outputs are indeed "hot". When I had it hooked up to my integrated amp, I had to turn the volume down to 2 to match the volume that the other two DACs achieved with the amp set to 5. On the Corda the stereo-link and the OBH-14 wanted the volume control set to 12 o'clock, but the ART DI/O had to be dialed down to 8 o'clock or so. (The '0' position on the Corda is at 7 o'clock.)

Specifications

stereo-link: 44.1 kHz 20 bit sampling, USB input, RCA output pair, 1/8" amplified headphone output. $130

Creek OBH-14: 48 kHz 24 bit sampling, S/PDIF and Toslink digital inputs, one analog input (RCA), two simultaneous analog outputs (RCA), one "direct" and one controlled by the front-panel attenuator. $350

ART DI/O: 96 kHz 24 bit sampling, S/PDIF digital input, 1/4" mono pair analog input (for ADC), 1/4" mono pair analog output (for DAC). $200

M-Audio Audiophile 2496: 96 kHz 24 bit sampling, S/PDIF digital input and output, RCA analog input and output pairs, MIDI I/O. On-board mixer and patch bay. $180

The Final Analysis

The stereo-link is a fine budget choice for people wanting to improve their PC's sound output: it has excellent bang for the buck. For your $130, you get a decent DAC and a passable headphone amplifier. The only reasonable low-end "separates" alternative I can think of is: a S/PDIF capable sound card, a DAC like the ART DI/O and a small DIY headphone amp like a CHA47 unit. If you penny-pinch, you could probably put something like that together for $250, which would give you a better system than the stereo-link. Which do you want more: to get fine sound or to save money?

The Audiophile 2496 card does not make a good entertainment device. It's fine for recording applications, but for entertainment, you'll want to add an outboard DAC to it.

The choice between the ART DI/O and the Creek OBH-14 isn't clear. If you need ADC, the only choice is the DI/O. If you need a small preamp or volume control or Toslink, the OBH-14 is the only choice. Personally, I plan on keeping both.

 

[dhwilkin]

Interesting, thanks for the review! This is the first time I've seen several low-cost, but high-quality, DACs compared to each other. This should be helpful to many people just getting started and on a tight budget.

 

[Tim D]

Can you elaborate on how the D/IO's A/D section is stronger than your soundcard?

I was able to find a Yamaha-clone PCI card for $20 with a fully functional SPDIF coax out a long time ago when Creative Labs was probably still hooked on ISA. No digi-in though. Already seamless support with any Win OS I've seen too.

IMO the D/IO is also cool in that it can function in either PC or stereo environment for me anyway I need it.

 

[tangent]

Quote:

Can you elaborate on how the D/IO's A/D section is stronger than your soundcard?

I haven't done a serious quality comparison like I did for the DAC section.
Mainly I use the DI/O as an ADC because of the superior usability of knobs vs. point/click/drag with the sound card's mixer program. I also like the DI/O's "tube warmth" feature -- it does indeed change the sound in good ways.

I won't have a good excuse to do a real quality test until I need the DI/O for some other purpose, in which case I'll do a quality comparison to see if I need a second DI/O, or whether I can get good enough results with the sound card's ADC.

 

[zoboomofo]

Thanks for the review Tangent. Very useful since many of us are looking at your gear choices as potential upgrades. I should add that the Creek dac ( one with a toslink in ) might not be the only choice for those of us with only toslink out transports - as chych suggested to me. You could get still get a DI/O plus the digital interface for < 350 ( $70 for a Fostex COP-1/96).

 

[tangent]

Well, I just got done tweaking and re-testing the ART DI/O, and found exactly zero change.

I replaced the 6' cheap 1/4" to RCA cables with some 1' instrument-grade cables (the kind electric guitarists use), I put a 75 ohm terminator on the digital input, and I swapped the stock power supply with the 3.4 amp Radio Shack one you hear so much about. No change.

That's not really true. The old interconnects seemed to cause a very-low-level click occasionally on audio peaks, which the new ones seem immune to. And, the new power supply certainly is more convenient since it isn't a wall wart -- it has cables coming out of both sides of the box. These were both good upgrades, but they're more in the way of usability than audio improvements. Sigh...

I had been thinking about getting a set of the Bolder ART DI/O cables, but now I'm wondering if they'll change anything, either.

 

[ai0tron]

If noise is removed it sounds like a good improvement to me.

 

[tangent]

I've since discovered that the low-level pops I was hearing were caused by the ART DI/O. When you have it set to use its own internal clock, these pops happen, but when you have it set to slave to the external clock (i.e. that generated by the CD transport) the pops go away. Apparently using the internal clock is best when using the DI/O as an ADC, but you want external clocking when it's acting as a DAC. I'm guessing that the pops were due to discrepancies between the DI/O's internal clock vis a vis the transport's clock.

Still, the new cables are prettier, and for my purposes 1' cables are a whole lot nicer than the clutter caused by 6' cables.

 

[Tim D]

How long have you been testing under those sampling modes? Perhaps you should try re-evaluating your original comparisons under ext sync mode. 88khz mode for example has a slightly different tonality than ext sync besides the occasional pop...I happen to prefer ext sync. There is a thread in AA for which Art D/IO rep's have claimed the sampling modes are more of an A/D side option. And that ext sync is recommended for pure DAC use as cited in the manual.

I actually had a different power supply besides the stock which was for my USR 56k modem...I didn't notice a huge difference if any so I'm not going to jump out for the RS model yet. In fact there are other mods entirely that you can do for power supply besides just swapping transformers (thats all the wallwarts are).

 

[tangent]

Quote:

How long have you been testing under those sampling modes?

I don't know if I did all the tests for this thread's comparison with internal clocking or not. All I know for sure is that I can reliably make those pops appear and disappear by messing with the sync options. I don't remember having popping problems during the DAC test, so it may be that I just happened to settle on external synch.

Another explanation might be because most of the tests used my Audiophile 2496 as a source, whereas I disovered these popping problems with the DI/O connected to my Pioneer DV-444 (top model in the line containing the popular DV-343). I was using the DV-444 as a source during the past few days while testing all my Sennheiser cans against my new AKG 401s (which is another thread entirely). Once I corrected the DI/O's sync problem during the headphone test, I completed the test using ext sync, at 44.1 kHz. So far as I could tell, the disappearing pops were the only thing that changed when moving to ext sync.

I may try 88.2 kHz as you suggest, but the DI/O is back to acting as an ADC and I don't want to tear that setup apart yet again.

 

[Hirsch]

I can use 88 KHz for playing back CD's using DIO as DAC, but if it's HDCD, forget it. The signal from the source is not in sync with the DIO, and you get noise.

Here is my admittedly novice understanding of what's going on: A digital signal will normally have only one master clock. One unit acts as the master, and the other syncs to it. Most standalone CDP's don't have the capability to slave to an external clock, and why should they? Their timing is determined by the sampling rate of the source disk (and the quality of their own clock). Since the signal timing is determined by the source, it makes sense for the DAC to simply slave to the source. It can then take whatever sample rate the CDP dishes out (up to the limits of the DAC).

By setting a sample rate on the DIO, you're configuring the digital signal to run with two master clocks. This works as long as the source is sending a sample when the DIO expects to receive it...

The pops you heard were signal being sent to the DIO that it didn't expect, or perhaps the absence of a signal it did expect. They do not represent a problem, but rather a digital signal that was not correctly configured. Sometimes this can work. I use a Theta TLC jitter filter between my CPD and DIO. Consequently, my DAC DIO is getting a very clean signal, and I can run at 44.1 or 88.2 with no problems (usually).

As a contrast, my Theta Chroma 396 DAC can only run in external sync. It has no other sampling options, nor does it need them.

In going A to D, the DIO is converting an analog signal to a digital one. Here, sample rate settings are critical (how else would the DIO know what timing to use?) The receiving end of the digital signal (sound card?) should sync to the DIO, which would normally be the master clock in an A to D conversion. (This is one of the reasons that SoundBlasters and most consumer audio cards suck for digital recording. Instead of syncing to the external source, they receive the samples and reclock the sampling rate to 48 KHz.)

Note that if you use external sync while performing ADC, you're losing one of the major benefits of having the DIO as the convertor. The DIO has a very good clock, and it's not subject to the same electrically noisy environment that your sound card exists in. If the DIO is set to external sync here, the sound card's clock will determine the timing of the digital signal. As the DIO's clock is better than that in most sound cards, this will result in less accurate timing. A pro sound card usually has the ability to sync to an external source. In that case, you want to leave both the digital inputs and outputs of the sound card connected to the DIO (which will be set at the sampling rate for your conversion), and let it act as the master clock for your sound card. This should give you the best digital sound in the A to D mode.

This got longer than I intended. The short version: as a DAC, use external sync. For ADC, set the sampling rate and set your sound card to external sync (if you can). This should result in correctly configured digital loops.

 

[KR...]

Quote:

I can use 88 KHz for playing back CD's using DIO as DAC

This is how I have my DIO set up (it's not really my DIO, I am auctioning it). I get amazing results this way, I don't really like it as an ADC. But, it makes a killer DAC, I am using it with my DVD Player, as you know DVD players have very good transports.