Music Hall pa2.2 ADC/Phono Stage Review

[dtaylo1066]

Outrage from a company of Music Hall's reputation.

 

[xanalog]

If it were a $150 device, we could cut it some slack...

Sir, you are being too gracious!

Many MM/MC amps (per @welwynnick's Post# list) shame this MusicHall pa2.2, ...price; no matter.
MiniDSP ADept pokes its thumb in MusicHall's eye!
Its novel feature appears to be those dual (Var./Fixed) analog RCA outs. meh!

 

[Multicore]

Outrage from a company of Music Hall's reputation.

What exactly is that reputation? And on what was it built? I'm sincerely curious. I've no info on either question.

 

[AtticReclus3]

I remember buying an older unit to this when it was this or the fostex h4a whatever all in one, this was back in 2014-2016.. these guys haven't improved it seems, shame.

 

[AudioSceptic]

This is a review and detailed measurements of the Music Hall pa2.2 ADC (digitizer) USB Phono Stage. It is on kind loan from a member and costs US $449.
View attachment 480158
Controls are easy to use, being just toggle switches. An external adapter is provided for power:
View attachment 480159
Nice to see independent input for general digitization. You also have independent inputs for moving coil/magnet. I only tested the unit as digitizer. Not sure if line out is an analog pass through or for output from USB input (did not test this aspect).

There were no drivers on the company website so I used my ASIO4ALL emulation over Windows class driver.

Music Hall pa2.2 Measurements
I started out just testing it as a line input. The input saturated rather early at just 1.5 volt input:
View attachment 480160

It is not going to win any awards as an ADC, finishing near last with all the distortion it has:
View attachment 480161
But that is not the worst problem. That title belongs to timing mismatch between the two channels. Left and right sine waves should be on top of each other and they are not. If I increase the test frequency to 20 kHz, the delay mismatch becomes much closer to the wavelength of the tone:
View attachment 480162
In other words, by the time you get to 20 kHz, the two channels will be completely out of phase! This could be a firmware issue and fixable. As is, it makes the rest of the review academic but let's test a bit more in case they fix this:
View attachment 480163
Company has rather detailed specs and shows an incredible value of 120 dB which would be state of the art. Needless to say, we are not even in the same planet. Maybe that is a chip spec.

ADC input bandwidth is a bit above audible band:
View attachment 480164

192 kHz did nothing different from 96 kHz.

Let's switch to MM phono input:
View attachment 480165

Strange that the low frequency noise is flat. SINAD of 45 dB would tie it for the worst phono stage ever tested! And our channel mismatch as far as timing remains.

RIAA equalization is respectable:
View attachment 480167

Headroom is not, and strangely, is worse at lower frequencies:
View attachment 480168

Digging in, we see that it is the negative supply that runs out:
View attachment 480169

I lost the will to live, I mean to test more....

Conclusions
I expect competent implementation from Music Hall. The pa2.2 bucks that trend, demonstrating clear implementation bug and performance is quite sub-par. Other than good RIAA equalization, I can't think of anything positive about it. If it were a $150 device, we could cut it some slack but at nearly $500? No way. You can buy an audio interface with a phono stage for much less money and far better performance.

I can't recommend the Music Hall pa2.2. I hope the company at least fixes the timing problem.

------------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

Is this the worst product you've tested so far? It would be unacceptable at any price but for $449 it's outrageous.

 

[Subwaffler]

Seems pretty obvious to me that its faulty. Why not try a second sample before piling in on Music Hall?
Also, how does it behave through the line outputs?

 

[ZioAcido]

So.. MusicHall went shopping on Alibaba one afternoon.. took the published specs from NoName OEM as word of God, tossed the purchased internals in a chassis, and put a finger in the air to come up with the $500 price tag..
*deep sigh*..... what an utterly hopeless piece of kit.

 

[Subwaffler]

I hope the member who loaned it is being banned for not noticing it is faulty (or rubbish)!

 

[phoenixdogfan]

Thanks for the fine review, Amir. We're being too harsh here. This unit would clearly have been state of the art in 1959. Imagine pair this with a triode tube amp sporting Western Electric 300Bs and a set of Voice of the Theater horns. Pure heaven. For 1959.

 

[AnalogSteph]

But that is not the worst problem. That title belongs to timing mismatch between the two channels. Left and right sine waves should be on top of each other and they are not. If I increase the test frequency to 20 kHz, the delay mismatch becomes much closer to the wavelength of the tone:

In other words, by the time you get to 20 kHz, the two channels will be completely out of phase!

A 1 sample offset is a fairly common problem that's usually accompanied with L and R channels being swapped (which may have been corrected for in software). In a nutshell, whenever audio is sent around using a serial data stream (e.g. I²S), data alternates between left and right channel samples, and when synchronization between sending and receiving end is slightly off for some reason, the right channel at time t gets combined with the following left channel sample at time t+1 to form a new (L,R) sample pair. Should be fixable in firmware if you can find out where things go wrong.

I first came across such a problem near 2 decades ago on a SB Live! 24-Bit.

 

[FraPia-62]

I hope the evaluated unit is simply a defective sample and not indicative of normal production standards — though if that’s the case, one has to question where quality control fits into the process. Otherwise, it feels like we're moving backward in terms of performance.

Back in 1990 — 35 years ago — I received two Burr-Brown demo boards for evaluation, each featuring a PCM78 ADC, a PCM56 DAC, and a Philips TDA1535 sample-and-hold amplifier. If memory serves, the boards were labeled DEM1123. Aside from the different sample-and-hold stage (TDA1535 instead of SHC5320), they were identical to the DEM1122 boards shown in the PCM78 datasheet.

After modifying the boards to allow direct access to the PCM56 digital interface and completing the setup with a 96 kHz sampling clock and a reconstruction filter for the PCM56, I measured SINAD performance (24 kHz bandwidth) as follows:

PCM78: SINAD = 88 dB, improving to 89 dB with 2× decimation.
PCM56: SINAD = 92 dB.

It’s worth noting that both the PCM78 and PCM56 were considered cost-effective solutions at the time — not flagship parts — and were typically targeted at budget DAT recorders and CD players. Higher-spec alternatives included the PCM75 and PCM53.

By today’s standards, many would likely expect such legacy devices to appear at the far right of the SINAD performance bar graph — firmly in the red — behind even entry-level sound cards or USB audio dongles. However, this is not what I’m observing. Given the performance potential of even the lowest cost modern converters, I would argue that achieving significantly better results should be well within reach with a reasonable level of design attention.

Out of curiosity — does anyone still have access to a DEM1123 board and the ability to remeasure it using modern test equipment? It would be interesting to see how these early devices compare under today’s measurement conditions.

 

[Subwaffler]

Has it been confirmed that the unit is faulty yet?