I've read the threads mentioning the PCM2906 but I have some technical questions, so I figured I'd start a new thread. It appears that this chip is supplies class-compliant USB audio so that makes it, in theory, plug and play with Macs and PCs which is just what I'm looking for. My questions:
1. Is there a chip that provides S/PDIF to USB conversion without A/D D/A stages? I'm only interested in converting to and from digital signals.
2. The chip is spec'ed at 48 kHz 16-bit. The datasheet suggests that a 24-bit word would be truncated to 16. Is that correct?
3. Is there a chip like this that can handle up to 24-bit 96 kHz S/PDIF signals?
TIA!
1. Is there a chip that provides S/PDIF to USB conversion without A/D D/A stages? I'm only interested in converting to and from digital signals.
2. The chip is spec'ed at 48 kHz 16-bit. The datasheet suggests that a 24-bit word would be truncated to 16. Is that correct?
3. Is there a chip like this that can handle up to 24-bit 96 kHz S/PDIF signals?
TIA!
Actually, I did a little research today. I found that the PCM2906 does in fact do SPDIF in and out, to and from class-compliant USB. The only problem with it is that it's only 48 kHz at 16-bit and I'm looking for something that can handle up to 24-bit 96kHz.
I found that the Tenor 7022 also does both as you can see by the miniSTREAMER implementation. It does 24-bit 96kHz to class-compliant USB and is only $55.
The problem with it is that's it's just a little old: it's the older 1.0 class-compliant spec which runs at 12 mbits/sec and according to this document, it's not the most recent implentation. AFAIK, jitter isn't relevant in a digital-to-digital transfer and 12 mbits/sec is in fact more than double the bit-rate required for 24-bit 96kHz stereo. But I'm debating the compatibility and reliability aspects of this earlier implementation.
So the hunt is still on... but clearly some solutions do exist.
I found that the Tenor 7022 also does both as you can see by the miniSTREAMER implementation. It does 24-bit 96kHz to class-compliant USB and is only $55.
The problem with it is that's it's just a little old: it's the older 1.0 class-compliant spec which runs at 12 mbits/sec and according to this document, it's not the most recent implentation. AFAIK, jitter isn't relevant in a digital-to-digital transfer and 12 mbits/sec is in fact more than double the bit-rate required for 24-bit 96kHz stereo. But I'm debating the compatibility and reliability aspects of this earlier implementation.
So the hunt is still on... but clearly some solutions do exist.
Cool, nice to know those solutions exist. I wasn't up to speed with all the TE7022 capabilities - I have it in a USB-SPDIF converter.
Perhaps you could get in touch with the SDR USB 2.0 Audio Widget developers, they might be able to suggest a route towards a solution for you.
http://www.diyaudio.com/forums/digital-source/185761-open-source-usb-interface-audio-widget.html
Perhaps you could get in touch with the SDR USB 2.0 Audio Widget developers, they might be able to suggest a route towards a solution for you.
http://www.diyaudio.com/forums/digital-source/185761-open-source-usb-interface-audio-widget.html
Silicon chip has a project featuring the PCM2902 as a microphone/line to USB converter featuring direct SPDIF interface. It's a neat little chip
Silicon Chip Online - USB Stereo Recording & Playback Interface
You might like to checkout the following chips. These are custom chips with NDA requirements. (On the other hand, our audio-widget projects are based on open-source firmware and schematics.) As we explained in the audio-widget thread, we do not go for SPDIF interfaces because of the high jitter. However, if you want SPDIF, you can implement it with the Tenor chip below. There may be a board with the Tenor chip that you can get for US$170. As for the C-Media, it is for UAC2 (ie up to 192/24) but I'm not sure of the SPDIF capabilities.
Alex
CM6631
and
TE8802L
> - USB2.0 Audio Class v2.0 and v1.0
> - 2-Inputs supported by one I2S pairs with 128/256 Fs.
> - 2-Output supported by one I2S pairs. With 128/256 Fs.
> - Adaptive/Asynchronous Mode supported
> - Resolutions support 16/ 24Bit with sampling rates support
> - 8/11.025/24/22.05/32/44.1/48/88.2/96/176.4/192KHz
> - Built in one IEC60958 professional 24 bit/96KHz S/PDIF RX, with I2S
> pins as well.
> - Built in one IEC60958 professional 24 bit/192KHz S/PDIF TX, with I2S
> pins as well.
> - Eight GPIO pins with interrupt.
> TE8802L provides a very flexible USB2.0-based audio & multimedia solution.
> Interleaved 2-channel input and output with 16-bit/24-bit resolution and
> 8/11.025/24/22.05/32/44.1/48/88.2/96/176.4/192K sample rate are supported.
> TE8802L can be partitioned into 4 major blocks, which include the USB core,
> MCU processor, DMA control, and the peripheral control for I2S /
> I2C/SPI/UART/MIDI interfaces.
> TE8802L has TOC (PWM) output to control external light and integrated 4:1
> SAR ADC for analog input to control Volume UP/Down or Mixer Volume etc.
Alex
CM6631
and
TE8802L
> - USB2.0 Audio Class v2.0 and v1.0
> - 2-Inputs supported by one I2S pairs with 128/256 Fs.
> - 2-Output supported by one I2S pairs. With 128/256 Fs.
> - Adaptive/Asynchronous Mode supported
> - Resolutions support 16/ 24Bit with sampling rates support
> - 8/11.025/24/22.05/32/44.1/48/88.2/96/176.4/192KHz
> - Built in one IEC60958 professional 24 bit/96KHz S/PDIF RX, with I2S
> pins as well.
> - Built in one IEC60958 professional 24 bit/192KHz S/PDIF TX, with I2S
> pins as well.
> - Eight GPIO pins with interrupt.
> TE8802L provides a very flexible USB2.0-based audio & multimedia solution.
> Interleaved 2-channel input and output with 16-bit/24-bit resolution and
> 8/11.025/24/22.05/32/44.1/48/88.2/96/176.4/192K sample rate are supported.
> TE8802L can be partitioned into 4 major blocks, which include the USB core,
> MCU processor, DMA control, and the peripheral control for I2S /
> I2C/SPI/UART/MIDI interfaces.
> TE8802L has TOC (PWM) output to control external light and integrated 4:1
> SAR ADC for analog input to control Volume UP/Down or Mixer Volume etc.
Thanks to all for these most useful suggestions!
Trevor, the 2902, like the 2906, is only 48 kHz 16-bit.
alexlee188: synced and locked, what difference would jitter make to a SPDIF signal feeding a SPDIF input? The jitter would have to be catastrophically bad to cause a skipped or misread bit. Is this the case in some designs? I'll be checking out the audio-widget thread with great interest. Thanks.
Trevor, the 2902, like the 2906, is only 48 kHz 16-bit.
alexlee188: synced and locked, what difference would jitter make to a SPDIF signal feeding a SPDIF input? The jitter would have to be catastrophically bad to cause a skipped or misread bit. Is this the case in some designs? I'll be checking out the audio-widget thread with great interest. Thanks.
Thanks for those links alexlee188. It seems people have spent some serious time analyzing SPDIF jitter figures but it also seems to me that there's a lot of "hair splitting" going on. Most concerns are in the context of reproduction, ie: live translation of an SPDIF signal to audio. Naturally, the market and discussions are more heavily biased towards reproduction.
My own interest is capture: recording a digital stream as digital data. Direct translation to audio is of secondary importance and in that context, as long as bit-for-bit copying is possible, jitter shouldn't have any influence over the recorded data, unless the jitter is so bad that data gets corrupted in the transfer. As long as the SPDIF clock is stable enough for the receiver's PLL circuitry to accurately track, it should be fine.
There's a simple test to know if a digital stream was captured accurately: take a digital test file, send it to SPDIF and record it in the target system as a digital file. Then bring source and copy files to an audio editor, and making sure they're time aligned, phase-invert one and mix them. If they are accurate copies, they should null completely to 0, ie: complete silence.
My own interest is capture: recording a digital stream as digital data. Direct translation to audio is of secondary importance and in that context, as long as bit-for-bit copying is possible, jitter shouldn't have any influence over the recorded data, unless the jitter is so bad that data gets corrupted in the transfer. As long as the SPDIF clock is stable enough for the receiver's PLL circuitry to accurately track, it should be fine.
There's a simple test to know if a digital stream was captured accurately: take a digital test file, send it to SPDIF and record it in the target system as a digital file. Then bring source and copy files to an audio editor, and making sure they're time aligned, phase-invert one and mix them. If they are accurate copies, they should null completely to 0, ie: complete silence.
That's fine
The jitter problem occurs when you try to playback the stream via the SPDIF and use the SPDIF recovered clock for the DAC.
Anyway for capture, you migt also like to checkout out sdr-widget which has 192/24 capture using the AK5394A ADC. It has the lowest quiescent noise and highest dynamic range compared with all commercial soundcards we tested, except for one card, the EMU1212M, which has a slightly higher dynamic range.
Alex
The jitter problem occurs when you try to playback the stream via the SPDIF and use the SPDIF recovered clock for the DAC.
Anyway for capture, you migt also like to checkout out sdr-widget which has 192/24 capture using the AK5394A ADC. It has the lowest quiescent noise and highest dynamic range compared with all commercial soundcards we tested, except for one card, the EMU1212M, which has a slightly higher dynamic range.
Alex
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.