Topic: Deity PR-2 Pocket Recorder

[Sebastian]

The way these recorders attain better EIN is by having analogue gain that increases the signal at a much higher rate than it adds noise (i.e. it might make the signal 10 dB louder while only adding 1 dB of noise). So instead of thinking of the Roland R-05 as a recorder that can reach an EIN of -121.8 dBu(A), I like to think of it as a recorder that gives you clear enough gain to increase the level of the microphone signal so that it can stay as far away from that -98 dBFS(A) noise floor as possible. Or you can use an external preamp for that. Either way, the ADC's noise floor is absolute, you cannot escape it.

EIN is pretty much a marketing tool. It is usually measured at really high gain settings (sometimes in the 50's to 60's dB - that's amplification by a factor of 1,000). The performance of an amplifier at these gain levels does not say anything about its performance at considerably lower gain levels that you would use when recording live music. On top of that, those settings are usually "A-weighted". This means they are basically applying an aggressive EQ curve that further reduces a lot of the audible frequency spectrum before measurement - including noise.

Also, where does the rumor about the PR-2's gain stage being all digital come from? This sounds really odd and I really can't believe it.

[Gutbucket]

My speculation as I've not tested anything, but there is at least an analog input stage that provides some sort of gain matching to buffer and provide the appropriate level to the ADC.  Not sure if the "input gain" adjustment effects that or if its fixed and the gain adjustment is done digitally, but I imagine that at the very least the switching between mic-in and line-in sensitivity is managed in that analog gain stage.  Would be nice if PIP could be switched on when in the less sensitive line-in mode to accommodate hotter output mics like the DPA 4060s I initially tried to use into it.

Firmware update wishlist:
>Change metering to PPM or combo VU/PPM (or make menu switchable).
>Option to link both channels of input gain adjustment when in 2-channel mode.
>Switchable PIP on line-in.
>Option to slave mulitple PR-2s together with clock and transport sync via a simple wired USB connection (rather than via separate wirelessly timecode gen box).

Haven't tried the last one, may be possible already.

[dallman]

Apart from all this, I do see mixed reports here about the gain behavior of the Deity... Some say they get noisy recordings when gain set to 0, others say the gain is fully digital, so there is no need to set the gain to anything other than 0... I guess it would be good to get that sorted out...

I can tell you that it is sorted out. If at 24bit (with a live recording) you run gain at 0db you will get a crappy recording. If you set the gain the same as you would running 24 bit on any other machine , you will get as good a recording as you would get running 24 bit anywhere else.

32 bit is a different story and a mono recording with this deck. Many of the above comments do not specify what bit rate is being used. I am specifying and my data is based on my own results, so feel free to disagree, but if you have no personal experience I think commenting might just add confusion. If you choose to run 24 bit at 0db good luck, I hope it is a really really loud show. If someone is of the belief that there is no difference in 0db and +18db because it is digital, my advice is then run at +18. That way you can keep your belief and not get a crappy recording. 

[grawk]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

[dallman]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

It does sound good. I'll still continue to record my way, but this is a nice recording!

[grawk]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

It does sound good. I'll still continue to record my way, but this is a nice recording!

Thanks.  The PR2 and KK14s make a REALLY easy rig to get in the door and pull the heat.

[rumbleseat]

I hate to stir this pot after it seems settled, but I just did a short experiment:

Church CA-11 microphones into the PR-2 (5V PIP), placed in front of a speaker.  I played the intro to the song Aqualung by Jethro Tull at 95 dB (C weighted on REW with a calibrated mic)
So - it's loud, but not really quite up to concert level, which I've read is about 100 dB or so.

Recorded on the Deity in stereo, 24 bit, with the recording levels set at 0 dB, 15 dB, 24 dB, and 36 dB.
Sucked them into Audacity 3.7.1 and used Effect > Amplify to normalize each file.

Audacity added 34 dB to the 0 dB recording,  19.7 dB to the 15 dB recording, 10.6 dB to the 24 dB recording, and 0 dB to the 36 dB recording (this last one had clipped in two spots)
I then aligned the tracks and looped a "silent" section of the music near the start of the song.   (the Aqualung song (from the "Special Edition" album) was played from Spotify on my phone (Galaxy S10e) and cast to a Chromecast Audio attached to my receiver)

Are you still with me?

The silent section of the track recorded at 0 dB sounded like shit.  Like a bad cassette recording.
The track recorded at 15 dB was also noisy, but much quieter.
The tracks recorded at 24 dB and 36 dB were quieter still and sounded identical to each other with only a very slight amount of hiss.  I think this hiss represents the sum of all the residual noise in the entire signal chain.

It is not clear to me where the noise in the 0 dB and 15 dB recordings is coming from.  It's either inherent the Deity file, or Audacity adds it when boosting the volume.

TL:DR - For best results, please adjust the recording levels on the Deity PR-2 like you would on any other recorder.

[Rairun]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

I don't want to be unkind here because I think it sounds really good otherwise, but if I had made this recording, I'd be really bummed out that such a good capture had so much hiss. When I tape really quiet shows and then try to apply some sort of compression, I can often hear the hiss surfacing, disappearing and resurfacing. I try to go easy on the compression in that case. But I listened to "Ain't Got a Place" from your recording, and not only can I hear quite a bit of hiss throughout the entire file, but I can also tell the exact points at which the compressor was working due to the hiss getting louder or quieter (but never transparent).

I'm running the risk of sounding like a fool here if you go and say you didn't actually use a compressor and that I'm hearing things. But I don't think I'm wrong, it just sounds that clear to me.

[Gutbucket]

Dig McMurtry, will listen later..

To state the obvious, optimal gain required will depend not only SPL but also the sensitivity of the mics being used. Here is the sensitivity of some of the mics we've been discussing, arranged from low to high-

>Church Audio CA-11 (not sure)
>Neumann specs KK14 cardioid @ 3.6 mV/Pa ≅ -49 dBV (1 kHz into 1 kohm)   
>DPA specs 4061 omni @ 6 mV/Pa; -44 dB re. 1 V/Pa (±3 dB at 1 kHz)   
>AT specs 853 cardioid @ 11.2 mV/Pa; -39 dB  re 1V at 1 Pa (standard 3-wire config, but I think lower that than with the 2-wire 4k mod typical around TS)   
>DPA specs 4060 omni @ 20 mV/Pa; -34 dB re. 1 V/Pa (±3 dB at 1 kHz)  TOO HOT!

As mentioned, in my use DPA 4060 @ 20mV/Pa proved too hot for concert recording into PR-2, while 4061 @ 6 mV/Pa seems about right with the potential to accommodate high SPL content.

[Sebastian]

It is not clear to me where the noise in the 0 dB and 15 dB recordings is coming from.  It's either inherent the Deity file, or Audacity adds it when boosting the volume.

It's coming from the amplifier. And, like every other analog amplifier, its equivalent input noise is better with higher gain settings and worse with lower gain settings. Also, your Church mics might have contributed some amount of noise. It's usually best to conduct noise tests without real microphones. Instead, you can connect a dummy load resistor to the recorder's input.

I did my own measurements some time ago, although I must admit I only tested it at max. gain to verify Deity's too-good-to-be-true EIN claims. I should probably repeat these tests with zero gain sometime...

[Rairun]

Dig McMurtry, will listen later..

To state the obvious, optimal gain required will depend not only SPL but also the sensitivity of the mics being used. Here is the sensitivity of some of the mics we've been discussing, arranged from low to high-

>Church Audio CA-11 (not sure)

My CA-11s are 19 dB less sensitive than my Clippys with Primo EM272M capsules. The official sensitivity for these capsules is -28 dB (±3dB at 1kHz, 0dB=1V/Pa), but according to MicBooster their improved RF protection makes them 6dB less sensitive than the EM272Z1 capsules, which are also rated at -28 dB. So -34 dB? Which would place the CA-11s at -53 dB (or -47 dB, if we consider the official EM272M specs).

[grawk]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

I don't want to be unkind here because I think it sounds really good otherwise, but if I had made this recording, I'd be really bummed out that such a good capture had so much hiss. When I tape really quiet shows and then try to apply some sort of compression, I can often hear the hiss surfacing, disappearing and resurfacing. I try to go easy on the compression in that case. But I listened to "Ain't Got a Place" from your recording, and not only can I hear quite a bit of hiss throughout the entire file, but I can also tell the exact points at which the compressor was working due to the hiss getting louder or quieter (but never transparent).

I'm running the risk of sounding like a fool here if you go and say you didn't actually use a compressor and that I'm hearing things. But I don't think I'm wrong, it just sounds that clear to me.

Don't worry, I'm not thin skinned.  I put it up, I don't mind commentary.  I did use some compression in raising the volume so that the applause and audience noise between songs were less intrusive.  I do agree there's a bit of hiss, but to me, it doesn't distract from the music.  I will run with more gain next time I run the PR2 to see if it makes a difference.

[dallman]

https://archive.org/details/james-mcmurtry-20240915/20240915+T01+-+Painting+by+Numbers.flac

acoustic show with a pa.  Not a super loud show, and I recorded at 0 gain.  kk14->pr2

I don't want to be unkind here because I think it sounds really good otherwise, but if I had made this recording, I'd be really bummed out that such a good capture had so much hiss. When I tape really quiet shows and then try to apply some sort of compression, I can often hear the hiss surfacing, disappearing and resurfacing. I try to go easy on the compression in that case. But I listened to "Ain't Got a Place" from your recording, and not only can I hear quite a bit of hiss throughout the entire file, but I can also tell the exact points at which the compressor was working due to the hiss getting louder or quieter (but never transparent).

I'm running the risk of sounding like a fool here if you go and say you didn't actually use a compressor and that I'm hearing things. But I don't think I'm wrong, it just sounds that clear to me.

Don't worry, I'm not thin skinned.  I put it up, I don't mind commentary.  I did use some compression in raising the volume so that the applause and audience noise between songs were less intrusive.  I do agree there's a bit of hiss, but to me, it doesn't distract from the music.  I will run with more gain next time I run the PR2 to see if it makes a difference.

It will...

[Sebastian]

I was a bit bored today and decided to take a look inside the PR-2, hoping to find out what kind of ADC this thing uses.

The PR-2 contains two PCBs that are stacked on top of each other. I was able to identify a few ICs, but a some of them had their top markings removed. What a dick move by Deity! What I found out was that the whole thing is run by an ARM Cortex M4 32-bit MCU as well as the fact that they are using pre-assembled modules for the wireless (Bluetooth) connectivity.

Also, there is a Renesas DA7218 stereo codec/headphone driver IC! Section 9.8 of its datasheet describes the digital audio format. Something that stood out to me was the following:

The internal serialized DAI [Digital Audio Interface] data is 24 bits wide. Serial data that is not 24 bits wide is either shortened or zero-filled at input to, or at output from, the DAI’s internal 24-bit data width. The serial data word length can be configured to be 16, 20, 24 or 32 bits wide using the dai_word_length register bits.

Let this sink in! The ADC does not support 32 bits of audio data natively. It can, however, output a 32 bit data stream. So are we getting a 32-bit signal at all when recording in mono?. I'm not sure why they went that route. There are 32 bit codecs on the market.

Now take a look at table 29 in section 9.2.2. The DA7218's ADC does not support a sample rate of 44.1 kHz at all. And I had already been wondering why it wasn't possible to choose 44.1 kHz.

Also, since there was some discussion: The DA7218 has analog amplifiers, but they can be digitally controlled.

Also, there is an as-of-yet unidentified IC chip on the right side of the same PCB. The markings are just barely visible. I *think* the first line says "OPA", so this could be an analog op-amp, but I'm not sure. It's too hard to see. The second line seems to end with "75" (again, I'm not sure).

[Gutbucket]

Thanks for the peek inside. How easily did it come apart? Did it just snap back together?  If easy enough and unlikely to get damaged in a way that requires gluing back together I might upen mine up to see if I can better ID the unmarked chips for you.  Any insight into the potential of somehow physically wiring two or three of these together to share transport control and clock?