The X32 is 16 Bit?

[MJKlein]

Something isn't right here: https://trello.com/c/PKpWYAXs

[pmaillot]

The usb recording internal format is not 16 bits, but 24 bits in 32bits containers, per microsoft multichannel wav standard.

Those guys at Matr... whatever were just trying to get their proposal adopted.

-Patrick

[MJKlein]

But Adam literally said that the guts of the X32 is only 16 bit. What is going on? That was in response to my question about USB recording at 24b/48k. Weird.

[pmaillot]

But Adam literally said that the guts of the X32 is only 16 bit. What is going on? That was in response to my question about USB recording at 24b/48k. Weird.

Processing can be using a 16bits CPU, I don't know. But that doesn't prevent from managing/manipulating larger data sets such as 24 or 32 bits.

[MJKlein]

But Adam literally said that the guts of the X32 is only 16 bit. What is going on? That was in response to my question about USB recording at 24b/48k. Weird.

Processing can be using a 16bits CPU, I don't know. But that doesn't prevent from managing/manipulating larger data sets such as 24 or 32 bits.

True. I was talking with Dennis about what Adam said. As you know, he printed out all the schematics for the X32 and has gone through it from stem to stern. This is what he said:

When he says converter he is talking about the ADC on on each input board. There is a total of four per X32 each chip has 8 inputs, made by cirrus logic. 44/48 I am sure about based on the sample clock on the board. The bit rate is anyone's guess because it is configurable to many bit resolutions. Short of what Behringer states, and what is really coming out of the ADC chip, can only be determined by testing the stream, or intercepting the chip config data. If he says it is 16 then it is 16. Because I also thought it was low because of some thing I read in the data sheet for the chip and in some app notes, is no way they can get good THD+N for eight channel running 24bit, so I thought must be 20bits. 18 is HDCD, 16 is redbook, nobody would know unless they probed the stream.

Then he showed me the CirrusLogic ADC and said:

This is the ADC but they don't run it fast and high res, for 2 reasons: 1. it trashes the signal fast and high res, and 2. DSP processing power. They would need 16 ADC per x32 and a bunch more DSPs, and the US has export restrictions on some of the really good high bit stuff.

So, it makes sense. So, what exactly is 24 bit about the signal on the digital outputs?

[pmaillot]

But Adam literally said that the guts of the X32 is only 16 bit. What is going on? That was in response to my question about USB recording at 24b/48k. Weird.

Processing can be using a 16bits CPU, I don't know. But that doesn't prevent from managing/manipulating larger data sets such as 24 or 32 bits.

Take this back as far as I know... the X32 uses 2 x ADSP21371 DSP chips; specs below:

SHARC Processor ADSP-21371/ADSP-21375

• High performance 32-bit/40-bit floating point processor optimized for high performance audio processing
  Single-instruction, multiple-data (SIMD) computational architecture
  On-chip memory, ADSP-21371—1M bits of on-chip SRAM and 4M bits of on-chip mask-programmable ROM
  On-chip memory, ADSP-21375—0.5M bits of on-chip SRAM and 2M bits of on-chip mask-programmable ROM
  Code compatible with all other members of the SHARC family
  The ADSP-21371/ADSP-21375 processors are available with a 200/266 MHz core instruction rate with unique audio-centric peripherals such as the digital applications interface, S/PDIF transceiver, serial ports, precision clock generators, and more.

... and loads of SMT32F103RBT6, also 32bits processors.

-Patrick

[pmaillot]

So, it makes sense. So, what exactly is 24 bit about the signal on the digital outputs?

The AD/DA chips used are
4x CS5368, a 192 kHz capable, 24 bits, 8-Channel A/D Converter (on inputs it seems),
2x CS42438, a 192 kHz capable 6-In, 8-Out TDM CODEC (on Aux/Ctrl), and
3x CS4385 which are 24 bits 192 kHz capable 8-Channel D/A Converters for outputs I believe.

[MJKlein]

That explains why Dennis said he could make a breakout system that would give users access to high resolution audio for sidechain processing, and with fader and other channel controls intact.

Dennis also said that he understands how to add a daughter board to the X32 to increase the DSP capability and allow up to 192 kHz operation. I need to get him to join the forum.

[curtbl]

@MJKlein Ok not your last statement form Dennis about adding a daughter board to the X32.... got my attention. THis could be a game changer for both the X/M32 and Wing if we can ever get further into that schematics...

[MJKlein]

We're still working on the video port solution.