Article: Berkeley Audio Design Alpha DAC Review

[audiozorro]

I don't think I agree with "if your 24-bit DAC has a sample rate converter (SRC), the SRC inherently converts 16-bit data to 24-bit data before it undergoes D/A conversion, so it would be pointless to convert it to 24-bit manually".<br />

<br />

The recommendation is to use a digital volume control on a 24-bit file instead of a 16-bit file. This happens before any DAC, such as the Berkeley Audio DAC. As discussed previously, using the digital volume control on a 16-bit file will degrade the file before it reaches the DAC, regardless of the DAC used.

[Bob Stern]

I was thinking of a DAC that had its own digital volume control.<br />

<br />

You're right that the best way to implement digital volume control in computer software is to first convert 16-bit to 24-bit before applying the digital attenuation.<br />

<br />

Sorry for my confusion.

[emmodad]

can't seem to find this info anywhere, and have not yet seen an openup / teardown of the Alpha DAC to know precisely which components are used in its conversion architecture. Can anyone confirm:<br />

<br />

1/ if the Alpha uses a commercial D/A IC, and if so, which one (it's also possible that there may be several of the same IC used in a parallel configuration)? If not, any info on what hardware is performing the D/A conversion function, ie a module with any visible identification?<br />

<br />

2/ does the architecture include any obviously-visible device which could perform signal processing (FPGA, digital filter chip, DSP devices ie ADI SHARC or Blackfin, or TI or other) as a preprocessor to the D/A IC/module? Most probably this is the case, as (at mimimum) vol control and HDCD decode functionality must be performed<br />

<br />

3/ a DAC "box" can *accept* up to and including 192 kHz inputs, but the internal D/A conversion ICs do not necessarily operate at this sampling frequency. One example is the Benchmark DAC1, which *accepts* 32-192 kHz, but resamples any input sample rate to approx 110 kHz before conversion as part of their jitter-minimization strategy and to run the ADI D/A chip at an optimal performance-vs-sampling-frequency operating point (which was an ingenious system solution for the time when the DAC1 was designed, using D/A chips which are now of course many years old in design). Does anyone have info confirming that the Alpha DAC will actually operate at 176.4 and 192 without resampling? Does the user guide contain any info in this regard, and any performance-vs-operating-frequency characteristics? The website has only some leading- but definately marketing-speak information.<br />

<br />

Any pics of internals to be posted?

[zettelsm]

Chris and Computer Audiophiles,<br />

<br />

Just a brief note to say that my Berkeley Audio Alpha DAC arrived yesterday -- the missing piece to put my high-rez music server in operation. I had a wait of a bit less than six weeks, due in part to Berkeley revising production for EIC certification.<br />

<br />

The Alpha has been thoroughly discussed here and in other reviews, so I won't go on and on. Suffice to say that it completely lives up to its billing. High resolution files via the Alpha are stunning -- easily the best digital playback I've had the pleasure to hear.<br />

<br />

There was some concern about the power supply noise of the early versions of the Alpha DAC. Evidently that has been successfully addressed in the latest DACs now shipping; my Alpha is completely silent and I've noticed no external hum from the unit.<br />

<br />

There is nothing original in my music server system -- all of the ideas pretty much originate from Computer Audiophile postings, and I'm very appreciative of have such a great forum to get this information out here. Keep up the good work!<br />

<br />

Steve Z<br />

<br />

Mac G4 Powebook/Thecus N5200BR Pro/iTunes > Weiss AFI1 firewire to AES > Berkeley Alpha DAC

[The Computer Audiophile]

That's really great to hear Steve!

[riderforever]

Hi Chris,<br />

what's the power amp used in this review? BTW, do you a think the Berkeley AlphaDAC will work fine as preamp for a Mc275?<br />

<br />

Thanks

[jonmarsh]

Just the end of August, after reviewing this any several other reviews, all giving the Alpha high marks, I bought one for myself, unheard- and have no regrets at all- the performance is impeccable, and has raised CD playback to levels I once thought not possible- a typical example being the Sheffield CD of the Harry James "King James Version", taken from the 1976 recording; for the first time, this disk starts to approach what the vinyl delivered, which is no mean achievement. In this regard one can say it truly demolishes anything else I've heard at $6K or less. Imaging, transient envelopes, midrange textural trueness, are essentially in a class of it's own at this price. Obviously, the recordings that benefit the most are good ones, but there are other recordings that have not been perceived in the past as all that good because they are so demanding or punishing in some respect, but are now reproduced much more realistically and satisfyingly. <br />

<br />

The Alpha DAC does deserve the appellation of "closer to the mic feed" and I find it money well very well spent. Mine is currently being used with a Cambridge Audio CD player as transport and with AES/EBU output from a Fireface 800 connnected to Mac Mini, at various data rates and source types. Next to investigate is a card which converts DSD to 24/192 for SACD, to see how the Alpha works with that material. <br />

<br />

Thanks for your review Chris.

[audiozorro]

however what are you trying to accomplish when you say you need "to investigate is a card which converts DSD to 24/192 for SACD, to see how the Alpha works with that material"?<br />

<br />

What is your source for DSD material? Are you looking to rip or digitize SACDs?<br />

<br />

I had been contemplating the Korg MR-2000S to digitize vinyl at either:<br />

<br />

-PCM audio format: 88.2 kHz @ 24 bit, 96 kHz @ 24 bit,176.4 kHz @ 24 bit, 192 kHz @ 24 bit<br />

<br />

or<br />

<br />

-1-bit audio format: 2.8 MHz @ 1-bit DSD, 5.6 MHz @ 1-bit,DSDIFF format (extension .dff), DSF (extension .dsf),WSD (extension .wsd)<br />

<br />

The Korg AudioGate software can convert the DSD formats to a lower PCM audio format.<br />

<br />

I assume if you are looking to tap into the SACD layer this would be the legal and relatively inexpensive way to go barring the expense of needing an excellent SACD player. From the reviews and user feedback of the Korg MR-2000S, the DSDIFF result is indistinguishable from the original SACD played back on an excellent SACD player. Of course you lose a little something if you convert to PCM.

[zettelsm]

I guess I have similar questions as Audiozorro -- what is your source for DSD material? I've been doing a fair bit of research and haven't found a cost-effective way of extracting the DSD information from a commercial SACD. I own an Esoteric X-03 player which can be set to output the encoded DSD stream from an SACD via iLink (firewire), however short of pro gear to decode and/or do sample rate conversion, I haven't seen anything that will let me capture the stream as a file, let alone do decoding and SRC.<br />

<br />

I have a Korg MR2000S, the latest firmware and software (which, BTW allows burning a DSD DVD that can be played) but AudioGate will not capture streaming PCM or DSD. 2L Recordings has about 25 minutes DSDIFF music files at their website that can be downloaded and experimented with (and AudioGate will recognize and manipulate them).<br />

I'd love to find a source of DSD music files. Live music recording with the Korg isn't very feasible for me here in rural Montana -- there just aren't many concerts around here. <br />

<br />

For archiving vinyl I highly recommend using DSDIFF, either sampling frequency. DSD can be converted to high resolution PCM without the decimation issues and degradation that can occur with PCM to PCM SRC. I haven't compared the Korg/AudioGate "DSD disk" against DSD converted to 24/192 or 24/176.4 yet, but vinyl recorded to DSD and then converted to 24/192 and 24/176.4 and played via my Berkeley Alpha DAC is very, very good. I haven't developed a firm preference yet for 24/192 or 24/176.4 when converting from DSD, though I have read some theoretical and one or two practical comments that express a preference for 24/176.4 when doing SRC from DSD.<br />

<br />

I'd like to learn about others' experiences with DSD and DSD SRC to PCM.<br />

<br />

Best regards,<br />

<br />

Steve Z<br />

Montana, USA

[StephanLJ]

Hi Steve et al.<br />

I was facing similar problems since the only method of getting music off a SACD is to convert to PCM. I found the solution below and it works great for me. Output is 24/96 and with Lynx, Amarra and alpha DAC the result is more than acceptable. For recording I use Amadeus Pro.<br />

<br />

http://freerider.dyndns.org/anlage/HiresAudio_E.htm<br />

<br />

Greetings Stephan

[jonmarsh]

Yes, my interest is possible methods of SACD capture for the music server (Mac Mini). I do have a high quality SACD player already, and with Pure Vinyl it's possible to do a good capture at 24/192 and create music tracks in AIFF. Pure vinyl also has the ability to create fairly high quality 44.1 from the high rez capture and has an upsampling playback engine that can integrate with iTunes (Mac), though I haven't tried that out yet. Many things on the to-do list at this point, as I'm also finishing up a major speaker project. <br />

<br />

But this weekend I'm buying a used Pioneer DV-79 with the intention to install a board from these people: <br />

<br />

http://audiopraise.com/vanity/overview.php<br />

<br />

which would do a DSD to PCM conversion and output at 24/192- through SPDIF- which I will try with the Berkeley Alpha DAC. I suspect that may be the way to go, and as mentioned earlier, would use Amadeus Pro or another of the audio stream capture utilities available for the Mac to capture the incoming bit stream. Possibly a dicey concept, but I'd really like to get my SACD's on to the music system, especially the single layer ones with no hybrid RB layer which I can't back up in any form currently and can only play on the SACD player (Marantz SA-11). <br />

[rlodad]

Chris,<br />

I'm confused again. Ugh. In rereading the Alpha DAC review, you state (I think) that the DAC rolled off transients when used with the MacPro and Lynx AES16e card compared to Media Monkey and the legacy card in the Dell PC. If so, why do you still use the mac as your server? Have the newer iTunes or Amarra made the issue better? I bought a Berkeley, and need to decide on server type. Currently have a MacPro dual core 2Ghz with all my music under iTunes. Should I use this or boot under Windows 7 and use Media monkey or J.River?<br />

<br />

Thanks in advance.

[The Computer Audiophile]

Hi rlodad - Good observation and question. Currently I use a Mac Pro OS X Snow Leopard & Win 7, Mac G5 OS X Tiger, MacBook Pro OS X Snow Leopard, Dell Windows XP, and two custom built Linux servers I am working on. They all have different characters to the sound. I need to use all of them to keep up on everything that's out there :~)

[rlodad]

Thanks, Chris, but let's cut to the chase. Do you prefer one over the others with your current system? Amarra vs. "virgin" iTunes vs. Win 7? Can you give any subjective listening qualities of each/ We know you think the transients are rolled off some in whatever version of iTunes you were using for the berkeley review, any change with new iTunes. I know you are avoiding the subject, how about just a hint?<br />

<br />

:^)

[The Computer Audiophile]

Ha! I'm really not trying to avoid the subject :~) I just don't want people to think there is one "best" solution for all systems. Right now my go-to server for sound quality is the Mac G5 running OS X Tiger, iTunes & Amarra, and outputting audio via a Lynx AES16 PCI card.<br />

<br />

Each of my systems has weak points. I think the G5 has fewer weak points or maybe just weak points that aren't as apparent in my system.

[rlodad]

Chris,<br />

Thanks for the information. Any idea why the G5 might sound better in your system than the Intel MacPro? We are all assuming it is not because you're a dummy and built a poorly matched system, oh great leader... I have a 2gHz Intel Mac Pro (32bit, bummer...) and am wondering if you have any subjective evaluation of the difference in sound b/t the two and why the G5 might sound "better" or different.<br />

<br />

Thanks again.

[The Computer Audiophile]

Hi rlodad - The following (on-going) thread has more insight into this topic than I could ever write myself. Check it out if you haven't already. The Mac Pro is definitely capable of wonderful sound that's better than physical disc spinners, but the G5 has a little something extra. The Mac Pro seems to sound a little dull compared to the G5 and my WIndows XP music server. I stress that this is in my system because a big factor may be the Lynx card. The AES16e is the only card to fit in the Mac Pro whereas the AES16 PCI cad fits into the G5. I'll have to do much more listening with non-AES DACs to better understand the differences between Mac Pro and G5, but that takes a ton of time.<br />

<br />

<br />

http://www.computeraudiophile.com/content/Magic-Words-Power-Mac-G5-versus-Mac-Mini

[bill13]

I have Berkeley Alpha DAC & LYNX AES16 computer interface card.<br />

<br />

The Alpha DAC, even by itself with inexpensive CD player digital out, sounds much better than my Wadia DAC - a surprise to me.<br />

<br />

High res files, like 24/96, etc., thru computer LYNX AES16 is best I've heard so far - price no object.

[jeroen020]

Hi Bill,<br />

<br />

Which Wadia are/were you using ?<br />

<br />

[wattage]

OK, The Linx card seems to be the way to go for desktop applications but how are people attaching to the Berkeley if you have a laptop? Firewire through ? to what? USB to ? to what? S/PDIF if you are lucky enough? I'm trying to get some ideas as I should be able to get my hands on a DAC next week to try in my home system but I have an older laptop with Firewire and USB only. Thanks in advance for any replies. I'm just trying to get some indication as to what most people who are lucky enough to have this DAC are driving it with.

[The Computer Audiophile]

Hi Wattage - Patience will help you on this one. Great Asynchronous USB 24/194 to S/PDIF converters are on the way in a month or two from top notch manufacturers. I know quite a few people using optical TosLink out of MacBooks waiting for these converters.

[riderforever]

any update on the Berkeley USB to S/PDIF converter? And have you ever tried the M2tech hiFace, that seems a bargain at that price?<br />

<br />

Have a nice day,<br />

Fabrizio

[VandyMan]

I've always wondered why no one makes a software HDCD converter i.e. after you rip an HDCD disc it would trans-code the data and create a 20-bit (or 24-bit) file. That way you would not have to worry about the LSB being lost in the playback chain. In my case, since I use digital room correction and have tons of HDCD discs, it would be really helpful!

[JeffreyW]

The HDCD indicator light on my Berkeley Alpha lights on 44.1 HDCD recordings from Reference Recordings, but it does not light when playing back the Hrx Sampler which came with the Alpha. I am playing the original Wav files off the DVD. The sound of the Hrx Sampler is fantastic, but no HDCD indicator light.<br />

<br />

Does this mean my system is not bit perfect on Hrx? Also does anyone know if the 96/24 and 88/24 Reference recordings use HDCD<br />

<br />

<br />

Mac Mini (March 2010 model), firewire to Weiss AFI1, AES/EBU to Berkeley alpha. Dartzeel pre and main amp, Wilson Maxx2<br />

TNT HRX

[jonmarsh]

No, it means the high res HRx recordings don't use HDCD encoding; his was a process developed for CD's.