view all articles in
Lavry Black AD10/DA10 Digital Converters This pro-featured, low-cost line of A/D and D/A packages will leave listeners converted.
by John Gatski, 12.15.2007 John Gatski is the Publisher & Executive Editor of Pro Audio Review.
Dan Lavry has developed quite a deserved reputation as a quality digital audio converter designer. He has been lauded for his namesake Lavry Blue and premium Gold converter series, as well as key designs within other high-dollar brands over the years, including Ultra Analog, Apogee and dB Technologies.
Most high-end professional gear users say Lavry’s products have consistently been at or near the top for accurate sound. Now the new low-cost Black Series proves that great conversion is just not confined to high-dollar products.
Fast Facts Applications
Studio, mastering, project studio
Operation at 44.1, 48, 88.2 and 96 kHz at 16-/24-bit; + Dither Mode; AES, Word and Internal clock settings; three processing modes
AD10 and DA10 Black Series, priced at $1,400 and $1,000, respectively
Lavry Engineering, Inc. | 360-598-9757 | www.lavryengineering.com
PRODUCT POINTS - AD10
- Excellent converter quality
- Stereo Imaging
- 16-bit + dither mode
- Combo balanced inputs
- One button control of all modes
- Not enough metering (starts at -13 dB)
The AD10 offers unbeatable converter performance in a reasonably priced, nicely featured package.
PRODUCT POINTS - DA10
- Sound quality
- Headphone amp
- Digital-controlled analog gain control
- Sample rate indicators
The DA10 is a great sounding, low-cost converter that is sure to find its way into many mastering and editing rooms.
The AD10 and DA10 Black Series, priced at $1,400 and $1,000 respectively, are pro-featured digital converters that take advantage of the latest converter chip technology.
The AD10 operates at 44.1, 48, 88.2 and 96 kHz, at 16- or 24-bit. There is also a mode called 16 + Dither, which is essentially the 24-bit signal dithered for CD work. The converter includes three-clock settings: AES, Word and Internal.
Unlike any low-cost converter I have ever seen, Dan Lavry has added three processing modes to the AD10: Tube, Transformer and Complex. These modes, according to Lavry, offer Digital Alias Free emulation, a patented process that is said to remove the typical aliases that occur in digital conversion and are claimed to add non-musical harshness. The “Tube” mode also adds increased even-order harmonics, while the “Transformer” mode adds odd-order harmonics to the audio. The “Complex” mode combines the tube and the transformer.
Lavry’s AD10 sets a new mark in function-knob minimalism. There is just one toggle function switch to control sample rate, Word length, the processed modes, clock options and each channel’s gain. The settings are accessed by a series of toggle ups and downs to get the right modes.
Lavry has mentioned in conversation that the reduced number of knobs and switches clears up more room for handy features such as sample rate indicators, Word Clock length etc., plus the gain control is very simple and allows for as much transparent gain as possible. Though Lavry claims it very easy to operate, I personally found it a bit unintuitive in the beginning, and never quite got totally comfortable with it.
The front panel metering indicates left and right gain in the gain set, and gain setting modes with adjustable peak hold. For the gain setting while monitoring, the segments go from -13 dB to 0 dB full scale. In the gain set mode, the gain is increased via the meter from right to left, using the 0 dB to -13 dB segments.
Other front panel niceties include a push button on/off button and easy to read green LEDS for the various function options (sample rates, clock, sound modes and Word length [Bits]).
The back panel features the typical connections found on most professional converters, including balanced input via XLR and 1/4-inch combo jacks, AES/EBU digital output via XLR, S/PDIF via optical and a Word Clock via BNC. A supplied XLR-to-RCA converter jack cable allows connection to S/PDIF coax digital products.
The only internal user adjustment is to disable the peak hold function. It is easy to get to the units innards for an inspection of its quality layout, and to make the peak hold parameter change if desired.
The DA10 operates at 44.1, 48 or 96 kHz in the narrow or crystal modes. In the wide mode, you can playback 176 or 192 kHz audio, but the audio is downsampled to 110 kHz.
Unlike the A/D, the D/A has more switches, including selection of Input (XLR, Optical and S/PDIF), PLL (Wide, Narrow or Crystal) Polarity and Mode (stereo or mono). The panel also includes such welcome features as a LED status of sample rates from 44.1 to 96 kHz, and the power switch.
The volume is enabled by a toggle, a digitally controlled analog control. A digital readout indicates the relative gain from 0 to 56. The headphone output circuit is discrete, and both headphone and rear analog can be output simultaneously.
The rear panel sports the analog and digital input jacks. An internal jumper allows for setting each SLR output for unbalanced operation with XLR to RCA or 1/4-inch cables.
To test the audio quality of the AD10, I set up a session for stereo recording of a Martin Custom OO-28. The gear setup included a TASCAM DV-RA1000 master recorder, two Shure KSM-141 microphones, a TRUE Systems P2analog microphone preamp, Alpha Core XLR and interconnects and the Lavry AD10. The 24-bit guitar cuts provided a good reference to bring out the nuance, space and overall realism a quality converter should provide.
To get a further indication of the converter’s accuracy, I also copied several prerecorded high-resolution material, including SACDs, DVD-As and other recordings made through the output of an Esoteric DV-50 universal player.
The newly recorded audio was edited and burned to a 24-bit linear PCM DVD-V via Toast 8, and played back through a number of DACs, including the DV-RA1000 internal DAC, the Lavry DA10, Benchmark DAC1 USB and Bel Canto DAC2 upsampling DAC.
The converter line output playback included either headphone monitoring via a Benchmark HP-2 or routing via a Coda high-current preamp to the Pass X3550.5 amplifier, then to the reference Legacy Focus speakers.
For comparison purposes, I also recorded all the same material through a Benchmark ADC1 USB and the internal converters of the TASCAM DV-RA1000. Most of the reference recordings were done in the (Clear) mode with no harmonics processing.
Maybe I am a bit slow, but I have to mention that I found the one-button-does-everything approach that Lavry designed into the AD10 to be unintuitive. Having to cycle through the various functions with up pushes and down pushes never really felt that comfortable to me — especially in switching from one channel to the other to set the gain.
Second Opinion In April of 2004 I reviewed and subsequently bought a six-channel Lavry LE4496 (Blue) as my stereo/surround monitor digital-to-analog converter. At the time, it raised the bar for features and quality in its price range. Since then, Lavry’s competitors have rolled out incremental improvements in their own mid-priced DAC range. Thus, the DA10 has followed suit.
What the DA10 now brings to the table is an increased sense of refinement. As a user of the Lavry DA924 Gold DAC, I have always found it apparent that there is a difference in the top tier. The Blue DAC sounds somewhat harder in the upper-midrange than the Gold. The latter offers unmistakably superior purity and resolution, while sounding smoother and more natural.
The DA10 takes the Lavry mid-price range a step further towards the Gold. Highs are smoother and lows are better defined. Sonic textures, ambience cues, imaging and reverb tails are more faithfully reproduced, as well. Both the Blue and Black DACs are easy to accept as reference quality. Only a very small mental adjustment is needed, and it happens unconsciously and quickly. The Black DA10 sits comfortably between the Blue and the Gold, and has now become my primary monitoring DAC.
The real surprise, however, came with auditioning the AD10. I found its quality to be competitive with analog-to-digital converters costing four to five times as much, the differences in resolution being barely more than subtle. Listening to analog, as well as high-resolution SACD sources through the AD10 produced just about as good an analog-to-digital conversion as I’ve ever heard.
The one nit I’d pick is that there is no input metering below -13 dB; recording a wide-ranging classical piece produced no activity on the AD10’s meters for most of the selection. I briefly tried the tube and transformer harmonic modes, but found them both to be too noisy for all but very loud material. For straight-ahead conversion, however, the AD10 is a winner.
— Alan Silverman
Also, it was sometimes bit disconcerting that the meter begins at - 13 dB. Some material I was recording had levels lower than that, and sometimes I thought the converter was not working without any visual representation of those levels.
With my nitpicks out of the way I can talk about how good the Lavry sounds.. I found the Lavry AD10 to be one of the most revealing A/Ds I have ever used. Through every D/A I played the AD10 guitar recordings, and I could aurally spot tinges of room reverb, guitar pick or finger pick attack that I didn’t pick up from other A/Ds. The TASCAM internal A/D definitely did not have as detailed width, and while the Benchmark was very close, under careful repeat listening, I could hear just the tiniest bit of increased ambiance through the Lavry.
In listening to copies of Big Band and intricate jazz recordings that I copied from prerecorded material, the Lavry AD10 copied exactly what the DV-50 output!
Overall, the Lavry AD10 recorded audio — played back through any of the DACs — presented no harshness or edginess. Folks who have criticized PCM in the past should listen to this latest generation of converters; they are not hard-edged in their sound.
I should mention that although the Lavry revealed just a bit more inner detail width than the Benchmark ADC1 USB, it was slightly leaner in the bass. I would categorize the Lavry as more analytical, and the Benchmark as slightly more analog in character.
I then explored the Digital Alias Free processing modes. Since I prefer accuracy over processed audio, the processed modes did not interest me that much. The Tube mode did add warmth, but at the expense of transient quickness; the guitar plucks and cymbals sounded slower. The Transformer and Complex mode did not appeal to me at all, seemingly harder sounding than the Tube or Clear modes. Others, however, may find them useful in their work. They really don’t add too much to the cost of the converter.
I listened to the DA10 with the aforementioned AD10-recorded music, as well as playback of other A/D recorded material .The Lavry and Benchmark DAC converters were similar, but I found that, like the A/D, the DA10 was slightly more analytical in the treble subtleties with reverb tails; transients sounded just a bit more separated in the mix..
To my ears, the DA10’s mysterious “Narrow” mode (Lavry would not comment on what the circuit is exactly doing) increased the upper end separation even a touch more than the “Crystal” mode, but we could not see any measurement differences between the “Crystal” and the “Narrow“ modes (See Bench Tests on page 46).
The Lavry DAC has an excellent headphone amplifier, outputting quality audio back through my AKG 701s and Ultrasone 2000s with plenty of drive power.
Additionally, I must commend Dan Lavry for adding the sample rate LEDs. It is immediately useful to the operator to see what sampling rate is at play. Most DACs do not have input sample rate display.
The DA10’s only negative is more of a suggestion: add actual RCA unbalanced outputs to give it greater utility.
The Lavry Black DA10 and the Lavry AD10 Black series are fine examples of converters employing chip technology and unique design twists to produce incredibly detailed converters for not that much money.
The AD10’s accuracy rivals, or perhaps exceeds, the performance of other A/Ds at any price. The DA10 is an excellent DAC that makes a perfect companions to the AD10. Both are highly recommended.
BENCH TEST - Lavry AD10 A/D Converter
Bench Measurement Data
Input voltage for 0 dBFS, 2.79-12.4 V/11.1-24.1 dBu
Unit reaches digital full scale before input overload occurs
Digital output audio signal polarity relative to audio signal input
44.1 kHz Fs, +0.0, -3.0 dB 20 Hz – 21.6 kHz;
96.0 kHz Fs, +0.0, -3.0 dB 20 Hz – 47.5 kHz
Total Harmonic Distortion
At –0.1 dBFS, Measurement bandwidth = Fs/2;
44.1 kHz Fs, < 0.0009% 20 Hz – 20.0 kHz;
96.0 kHz Fs, < 0.0008% 20 Hz – 17.5 kHz;
0.0012% @ 20.8 kHz
44.1, 96.0 kHz Fs
< +/- 1 dB 0 to –120 dBFS
< +5 dB @ -140 dBFS
Signal to Noise Ratio
Input termination 600 ohm 44.1 kHz Fs
Wide band (Fs/2) 115.7 dB
A weighted 118.1 dB
96.0 kHz Fs,
Wide band (Fs/2) 112.0 dB
A weighted 118.0 dB
44.1, 96.0 kHz Fs
A weighted <10 Hz – 20 kHz BW, 118.5 dB
44.1, 96.0 kHz Fs
20 Hz @ 0 DBFS, THD+N in 400 Hz – 20 kHz BW -113.5 dBFS
44.1, 96.0 kHz Fs
Ch1 > Ch2
20 Hz – 4 kHz > 110 dB
20 kHz > 104 dB;
Ch2 > Ch1
20 Hz - 20 kHz > 110 dB
Gain set at “-9” with 4.42V signal input > 0 dBFS, “sound” at clear, 24 bit unless otherwise noted.
(click thumbnail) Fig. 1 Frequency response as a function of sampling rate. Red = 44.1 kHz, Blue = 96.0 kHz. Bench Measurement Commentary
The Lavry AD10 is an interesting, versatile, and great performing A/D converter that has some interesting simulation modes for tube and transformer type characteristic sounds.
To keep the number of measurement files and data presentation to something reasonable, testing was done at 44.1, & 96.0 kHz sample rates.
shows the high frequency response rolloffs for the sampling frequencies of 44.1 & 96.0, kHz. The presentation is from 10 kHz upwards to best see the high frequency end of the response. The low frequency response of the unit (not shown) was essentially the same for all sample rates and was down about 0.05 dB at 10 Hz.
THD+N (Total Harmonic Distortion plus noise) at both sample rates with the measurement bandwidth set to Fs/2 is shown plotted in Fig. 2.
The rather abrupt rise in distortion above 10 kHz at the 96 kHz sample rate is caused by second harmonic type distortion which is why is suddenly starts to reduce above 20 kHz or so as the approximately 45 kHz bandwidth starts to rapidly attenuate the second harmonic above this when the fundamental gets past 20 kHz.
THD+N as a function of signal input level below 0 dBFS really exhibits some very low values away from full scale, especially at the 44.1 kHz sample rate.
(click thumbnail) Fig. 2 THD+N vs. frequency and sample rate. Red = 44.1 kHz. Blue = 96.0 kHz.
In characterizing the effects behavior of the AD10, I made a plot of how the amount of effect distortion varied with input level for the three effect settings of tube, xfrm (transformer) and complex. It appears that the amount of these distortions is a linear function of input level when plotted as db down relative to full scale. Interestingly, the amount of these distortions is relatively the same in magnitude. It is to be noted that the tube and complex modes cause actual clipping on the plus half cycle of the signal from about –2 dBFS up to 0 dBFS whereas the transformer mode does not. Another thing of significance to note is that the amount of distortion of any of the effects at any particular input level is very constant with frequency over the audio band — not a particular attribute of transformers as their distortion is mostly in the low frequency end of the audio range. To show the harmonic addition of the effects, I plotted the spectrum of the added distortion at – 2 dBFS for each effect of a 1 kHz test signal. The tube mode gives a nice spectrum of evens, the transformer a nice spectrum of odds, and the complex Is like the xfrm with the added second harmonic. Also, the noise apparent floor comes up with the effects selected vs. the clear normal mode — apparent as I am not sure if this is the Audio Precision spectrum analyzer doing this or the AD10. I suspect it is the AD10.
Input/output linearity was quite good for the AD10.
Finally, channel separation, as is typical, shows the same for both sample rates. What is most atypical is the fact that the usual decrease in separation with increasing frequency is very absent in this design.
— Bascom H. King
BENCH TEST - Lavry DA10 D/A Converter
Bench Measurement Data
Output Levels @ 0 dBFS
Internally strapped as received, 12.2V, 24 dBu
Headphone Output Level @ 0 DBFS
Front panel level at maximum
High impedance loading 6.1V
Both channels loaded
with 50 ohm 6.096/734 mW
With Polarity switch in normal position
Pin 2 hot Inverting
Pin 3 hot Non-inverting
Line Outputs 150 ohm
Headphone outputs 0.074 ohm
44.1 kHz Fs, +0.0, -3.0 dB 20 Hz – 21.1 kHz
96.0 kHz Fs, +0.0, -3.0 dB 20 Hz – 45.8 kHz
192.0 kHz Fs, +0.0, -3.0 dB 20 Hz – 55.5 kHz
Total Harmonic Distortion
22 kHz measurement filter:
44.1, 96.0, 192 kHz Fs < 0.0011%, 20 Hz – 1 kHz
44.1, 96.0. 192 kHz Fs < 0.0026%, 20 Hz – 20 kHz
80 kHz measurement filter:
44.1, 96.0, 192 kHz Fs < 0.0014%, 20 Hz – 1 kHz
44.1, 96.0, 192 kHz Fs < 0.0085%, 20 Hz – 20 kHz
44.1, 96.0, 192 kHz Fs
+2.0/-0 dB 0 to -120 dBFS
+6.0/-0 dB @ -130 dBFS
Signal to Noise Ratio
44.1, 96.0, 192 kHz Fs:
Wideband, 96.0 dB
22 kHz BW, 109.0 dB
A weighted, 112.0 dB
44.1, 96.0, 192.0 kHz, 114.0 dB
44.1, 96.0, 192.0 kHz, 108.0 dB
44.1, 96.0, 192.0 kHz Fs
Ch1 > Ch2:
110 dB 20 Hz – 300 Hz
81 dB @ 20 kHz
Ch2 > Ch 1:
110 dB 20 Hz – 600 Hz
82 dB @ 20 kHz
Measurements were made at 44.1 & 96.0 kHz sample frequencies and with bit depth of 24 bits triangular dithered. Digital signal input was via AES/EBU. The PLL switch was set to narrow for 44.1 & 96.0 kHz Fs and wide for 192 kHz Fs. The adjustable output was set for maximum output level.
(click thumbnail) Fig. 1 Frequency response at 44.1, 96.0 & 192 kHz Fs. Red = 44.1 kHz, Magenta = 96.0 kHz, and Blue = 192.0 kHz. Bench Measurement Commentary
The Lavry DA10 is a compact, versatile D/A converter that will accept input sampling frequencies from 30 – 200 kHz
Frequency response for the sampling frequencies of 44.1, 96.0 & 192 kHz is shown plotted in Fig. 1.
With frequencies extended down to 10 Hz, the response was down about 0.05 dB at 10 Hz. As in the recently measured Benchmark DAC 1 USB, non-standard sample rates and frequencies above 96.0 kHz are handled by a sample rate converter running at a sampling frequency of 110 kHz. In the case of the DA10, these sample rates are handled by setting the PLL switch to wideband. So, in figure 1 for the 192 kHz input sample rate, the response rolls off at about 55 kHz instead of the usual 90 kHz that would result from a true sample rate of 192 kHz.
Total harmonic distortion plus noise in both a 22 & 80 kHz measuring bandwidth as a function of signal frequency and sampling frequency was virtually identical for all sampling frequencies. This is shown in Fig. 2
for a sampling frequency of 44.1 kHz and for both of these measurement bandwidths. Extending the signal frequency up to 45 kHz for the 96 & 192 sampling frequencies results in distortion continuing to increase beyond 20 kHz.
I tried to see if there was any discernable difference in the amount of distortion as a function of the PLL mode set to the three settings of narrow, wideband, and crystal. There was just a tiny bit more distortion and noise in the wideband setting when the measurement bandwidth was set to 80 kHz. The results were the same for the narrow and crystal modes.
(click thumbnail) Fig. 2 THD+N vs. frequency and bandwidth measurement at 0 dBFS. Red = 22 kHz BW, Magenta = 80 kHz BW. Sample rate = 44.1 kHz
THD+N of a 1 kHz test signal as a function of level down from 0 dBFS full scale is tested for a sampling frequency of 44.1 kHz and a measurement bandwidth of 22 kHz. Results were essentially the same for all three sampling frequencies.
Deviation from linearity was similar at all sampling frequencies, but a little worse at the higher rates.
Channel separation was essentially independent of sampling frequency but not direction.
It was in some new jitter effect measurements that I thought that I might see some difference in the audio output behavior in the three PLL modes. I have devised a jitter effect measurement where a 500 Hz 1 UI is injected onto the data stream on a full scale 1 kHz signal. The idea is to see what affect this has in the distortion of the D/A converter’s audio output. Here distortion sidebands of 500 and 1500 Hz show up at appreciable percentages in one D/A converter tested in the past. I applied this test signal to the DA10 and made measurements for each of the three PLL modes. Much to my amazement, there was absolutely no effect at any of the PLL settings with all three measurements essentially looking alike. A spectrum of the distortion output of one of the PLL modes is plotted in figure 8 and all that shows up are the harmonics of the full scale 1 kHz audio signal — more than 100 dB down.
— Bascom H. King