jerryfreak
Active Member
Background
This review is in regard to the Deity PR2 Stereo Pocket Recorder.
As an amateur field recordist (AKA "taper") who focuses on live concert recordings, this has been a long-awaited item in my community, Deity is a company that focuses on production audio and offers various mics, transmitters, timecode boxes, etc to that industry. A few years back they offered their first two-channel portable recorder, the BP-TRX. After a long long wait (over a year from product announcement to release), if finally came out in summer of 2021, and while it was feature-rich, a friend who gave it a shot was less than thrilled with it for our application relative to other 2-channel devices that were already on the market, particularly in regard to metering/usability. Turns out it just wasn't a good match for our application.
Flash forward to April 2023 when Deity announced the PR-2 at NAB, and had a prototype in hand. The unit seemed to be the perfect device based on its claimed features and specs. 32 bit float... 30 dB of gain... low self noise... can power lavs at 5V. All in a compact form factor with great battery life. The 5V mic power is of particular interest because it allows for the operation of high-quality microphones at a voltage which allows for the high sound pressure levels seen at concerts to be recorded without distortion. These range from sub-$100 lavs all the way up to permanently polarized 20mm condenser microphones that record at studio-quality
I was skeptical based on my recollection of the "vaporware" release of the BP-TRX, and sure enough the summer 2023 release date got pushed back a full year and the PR-2 finally started shipping in July 2024. As someone who has used 32-bit float recorders from Zoom and Sound Devices, i personally found the feature to be of somewhat limited use as in general the 24-bit modes offer adequate resolution to capture anything i record, and I am monitoring the recording as it happens to set proper input level. There is lots of hype around 32-bit recording and while it has its uses, I still have yet to see it implemented in a way that expands the dynamic range available between the input noise and the max input level at overload, which as weve seen with numerous reviews, tops out in the 115-125dB range. In any case, 32-bit float feature is useful for a device such as this, as tapers often record concerts where the device needs to be out of sight and out of mind for various reasons, and the elimination of the 0dBFS wall with 32bit increases the chance of a successful recording in a measurable way.
Unfortunately, somewhere in the revision cycle between prototype and production, the feature set got downgraded to where 32-bit float recording would only be available for mono recording. Considering the existence of 32-bit float single channel recorders from Zoom and Tentacle this would reduce its appeal a bit among tapers, but a good 24-bit recorder is still a good recorder so it is worth a look
Test Setup
Measurements were done using the unbalanced out from either a Topping D90 or a D90SE. The D90SE has the option to do a higher voltage level to improve resolution, but in all configurations the dynamic range of the DACs exceeded the range of the PR-2 and produced similar results. The max unbalanced output of 2.1V at 0dB on the topping is approximately 8.6dBU. i rounded it to +9dBU in the following measurements to make the analysis simple since i was relying on the Topping's output level display which is fortuantely presented in actual dBU
Measurement software was the pro version of RMAA 6.4.5
Microphone tests were done with a single DPA 4015 wide cardioid capsule, and an MMP-GS preamplifier.
General Impressions
First impressions were "wow - its tiny!" a mere 41g without batteries, in fact adding some NiMH's fattened it up to 100g. For those obsessive about weight, given the long runtime on AAs, i think that AAAs with adapters would shave weight and still offer plenty of runtime. non-rechargeable lithiums even more so.
Its form factor is compelling. the nearest competitor is the Sony PCM-A10, which is only slightly bigger in volume, and offers respectable performance for our application. As with most small 2-channel recorders, the plug-in-power ("PIP") the A10 offers is an anemic 2.5V, and can only power limited models of microphones. Many tapers choose to use an external battery box to power these mics, which is commonly at least as large as a 9V battery. Measuring the voltage of the input showed that it indeed delivered 5.0V +/- 0.1V so it should be adequate
The unit has a solid feel, and buttons are responsive. Menu is clear, and the important functions like gain are right at the top where it belongs. In an active recording situation its easy to make adjustments as double clicking the click-wheel takes you right to gain adjustment, and a double click of the power/back button takes you right back. Screw-locks on the 3.5mm TRS input and output are a nice touch. Gain adjustments are in steps of 3dB. I personally favor discrete level adjustments as they are much easier to fix in post as opposed to slower fade-type adjustments.
Most importantly, starting and stopping recording, as well as power-on and -off are all via long press, and a triple press of the power button will lock and unlock the unit. The fact that these actions are all very purposeful, gives me high confidence that once the unit is recording, it will stay actively recording, a critical aspect.
One feature that is sadly missing on the US version is the ability to monitor while recording. This is presumably due to the Zaxcom patents on combination transmitter/recorder products. The non-US version offers this feature and its unclear whether the US version can be upgraded via firmware to enable this feature. As of now Deity only offers a single firmware update for both versions, not specific firmwares for each model
General Operation
The PR-2 offers both mic- and line-in selection, at the same input levels, the difference being the mic option allows 5V power to the mics. no measurements were done with the D90 in mic-in mode because i wasnt comfortable sending 5V power to the output of my DAC
Gain range is from -12 to +36 dB
the unit operates in either 24 bit or 32-bit float in mono mode, and exclusively 24-bit in stereo mode
Level Meters
Level meters were usable, with a caveat.... you need to understand a bit of calibration. at levels approaching 0 dBU, the unit when set at 0 dB of gain will indeed show levels approaching 0 dBFS with a sine wave. Actually recording real music acts is a bit more challenging. the unit is lacking a peak/peak hold level indicator, a serious omission IMO. I'll add some videos later but i was able to "calibrate" the unit by playing back a wave file of a live concerrt recording which had the usual RMS level near about -20 dBFS and peaks near 0 dBFS. When playing this into the PR-2 at a level where 0 dbFS = 0 dBU, the limit of the device, the meters require a different eye. When watching the levels youre really looking at the -20 dB and -10dB lines.... in a real recording with dynamic levels, this rule of thumb works: -20 line on the meters corresponds to about -12 dBFS peak… the -10 line is the equivalent of -6 dBFS. So to record with an acceptable level of headroom i would target peaks above the -20dB line, but below the -10dB line. treat that -10 dB ine as your zero and youre in a safe range.
Measurements
First order of business was to determine the operating range of the unit. Maximum input level is right about 0dBU, with massive distortion occurring above +3 dBU. This is a compromise relative to the Sony A10 or the Tascam DR2D which can safely accept levels up to around +10 dBU. Anyone using the PR-2 to capture a feed from a mixing console would be wise to have an attenuator cable in their kit just in case.
Once the maximum input level was determined, i was able to define a test range. I did an RMAA measurement at each of the gain levels of the unit (-12dB thru +36dB). The negative gain settings were not useful in increasing the maximum input level of the unit.
For 24-bit stereo measurement I set the gain level to record close to 0 dbFS for each pass, so for example a -36 dBU input signal used +36 dB of gain, a 0dBU input signal had a gain of zero, and an input level of +6 dbU was tested at gain values of -6, -9, and -12dB).
I may graph these and insert them later, but we can clearly see the "sweet spot" is in the range of -18 to -6 dBU input level, a good range for a unit designed to take mic input. In general the results are passable but not quite state-of the-art. Here are some tests done on other comparable recorders near their respective max input levels. The PCM-A10 we already discussed. The Sony ICD-UX560 is a smaller version of the A10 which is limited to recording at 16/44 PCM. The Rode AI Micro is a dongle-type device for iphones. Those were kind of one-off tests and i didnt bother to determine the optimum input range of those devices. Both of these are legacy products and can be found for well $100 and its disappointing that the resolution of the deity does not exceed theirs. We see the A10 and DR2D with higher resolution due to their much higher acceptable input levels
these devices were also tested at lower input levels of -20 dbU and -36 dBU
Here we see much more respectable showing from the PR-2, it is right there with the A10 and DR2D. It actuallyy exceeds the A10 at low input levels, perhaps that A10 measurement is flawed and needs redone. I did an additional test at all input levels with the gain set to record wave files at -12dBFS to see if i could squeak out any additional performance (waves wer enormalized in post to brign them up to an acceptable level for RMAA to analyze), but here we see we are still limited by input noise.
32-bit measurements were a bit trickier as it is mono only, and the 32-bit float recording is at an arbitrary recording level. To compensate for this I did the RMAA test, and then took the recorded wave file and duplicated the track into a stereo track, then adjusted volume to get the peaks near 0dB so RMAA could analyze it. I omitted the stereo crosstalk in this test.
We do see much higher resolution in the 32-bit mode, presumably the unit is stacking the ADCs from each channel to cover different input ranges, which explains why stereo recording is not available in stereo mode. We also see that the gain setting is essentially useless as the results are the same and resolution is entirely determined by the input level in relation to the noise floor of the unit, we see a linear relationship. I dont see the unit approaching anything near the 123dB dynamic range they claim, with the input level limited to 0 dBU
Real-world Mic tests
For the mic tests i compared with my current favorite miniature recording setup, the DPA MMA-A ("AKA "d-vice), a 2018 product, now discontinued, which is designed to power the DPA microdot-compatible series of microphones. DPA claims a spec of 114 dB dynamic range and as a world-renowned microphone manufacturer for 50+ years now i tend to believe them. The unit uses a preamp chip with a THD+N below -136 dB into an AKM5552VN ADC. Has always served me well. Unfortunately I don't have good measurements on this unit due to the non-defeatable 5V bias voltage, and it also requires a mic input load to even turn on, so it would be tested thru a blocking cap/resistor circuit. ive built these in the past as proof of concept to see if the unit could take line-in, and it did work, but ive long since lost that breadboard monstrosity.
Back to the test... I used a 114dB calibrator to simulate a 114 dB input to the microphone, The MMA-A was used with Apogee Metarecorder software on an iPhone. the MMA-A has +30dB of gain, At 114 dB, an input level of 75% in metarecorder yielded something close to 0dBFS
so i tested the PR2 with this same calibrator and it was able to capture the 1khz test wave cleanly at 114 dB and it was well within its acceptable input range. at 0 dBU, the 114dB input at the microphone generated a signal of -20 dBFS on the deity, and adding 18dB of gain put it at -2 dBFS as expected. In this case it is similar to the MMA-A in input and gain range, and should be able to take the loudest signal i would expect to record without overloading the input
So in an attempt to determine the relative noise level of the PR-2 input, i had to get creative. After testing the two setups to get as close as possible to 0 dBFS with a 114 dB tone, i then took the mic, wrapped it up in a t-shirt and then a thick sweatshirt and put it in my closed car at night in my rural neighborhood. The closest i could get to "silence" and my tinnitus was telling me all about it...
So i looked at the recorded noise in a wave editor then compared its RMS level relative to the peak level of each signal (which were all within a dB or two peak level)
MMA-A: 86.3 dB
PR-2 in 32bit: 84.6 dB
PR-2 in 24 bit: 85 dB
what this suggests to me is that the noise floor of all these units is below the combined microphone self-noise and minimal ambient noise
so i decided to test the same mic with a higher resolution setup,a Grace Lunatec V3 which has 70 dB of very clean gain and a low noise floor.
result of that was..... 86.1 dB
id call that essentially the same and while i wouldn't expect the PR-2 to have the same dynamic range and noise floor of the purpose-built preamp unit, its good to see that its noise floor seems to be well below any needed in real-world recording situations
Conclusion
I like the unit for portable recording, particularly if you have a setup that requires 5V mic power such as the DPA mics, sennheiser, Audio Technica, etc. It offers a good combination of size, weight, power consumption, usability, and acceptable performance at mic-in levels. If i was using the device with a hotter line signal from an external preamp or mixing console i would consider using other portable units such as the Sony A10 or Tascam DR-2D.
Cons are of course the lack of real-time monitoring, and lack of peak hold meter
A great effort by Deity imo, and hats off to Andrew Jones the man behind these products . Id love to see them implement stereo 32-bit recording at the resolution we are seeing with the 32-bit mono mode. I realize a unit of this size necessitates compromises, but for example they could replace the 2xAA with 2xAAA and still have 10+ hours of operation and more board space for a second 32-bit circuit as well as something like a switchable 20 dB pad to allow for higher input levels. Well as they say, perfect is the enemy of good, and this is certainly good enough for my application.
Exclusions
I didn't test any of the following features: Timecode, bluetooth/app connectivity, automatic gain, playback/output measurements, low cut filter, or limiter as i dont have a use for these features
Attachments
Last edited:
