Been testing my DCX2496 with various SPDIF/TOSLINK inputs, and have had mixed results. Wondering if it is possible to tap into I2S, if it's even available at this stage in the DCX (prior to DSP).
My thought was to use a an external USB/I2S ASRC converter and tap into the output of the existing CS8420. It is not easy to understand the data sheet of the CS8420; I know it has I2S capability but not sure what it's set up for.
Would this address any jitter problems that I current have with SPDIF inputs? I am not familiar enough with I2S to fully understand what I am gaining, but my understanding is it corrects the input signal perfectly, thereafter only being subject to the system clock to which it is feeding.
My thought was to use a an external USB/I2S ASRC converter and tap into the output of the existing CS8420. It is not easy to understand the data sheet of the CS8420; I know it has I2S capability but not sure what it's set up for.
Would this address any jitter problems that I current have with SPDIF inputs? I am not familiar enough with I2S to fully understand what I am gaining, but my understanding is it corrects the input signal perfectly, thereafter only being subject to the system clock to which it is feeding.
There are a number of clock kits and mods for the DCX. You could look into those either as a turnkey solution or as inspiration for your own if you want to get a lower jitter signal into the ASRC.
BTW, what sort of problems are you having with the digital input? It's AES, but tends to work OK with unbalanced SPDIF in. There is not Toslink, unless you've added one.
BTW, what sort of problems are you having with the digital input? It's AES, but tends to work OK with unbalanced SPDIF in. There is not Toslink, unless you've added one.
Yes, I've read about all the clock mods on the DCX. Thing is, my problem appears to be with the quality of my SPDIF input, not with the internal clock on the DCX.
What I found is by ear and measurement various sources resulted in significant difference with analog output quality. An AppleTV's toslink in combination with Twisted Pear's converter was the worst (no fault of TP, I believe). A MacBookPro's toslink was improved quite a bit, with only a few glitches here and there (sounds like a hiccup in the audio stream). Finally, my M-Audio SPDIF is essentially perfect, so my assumption is I am dealing with an issue of SPDIF quality.
I would rather not have to use my M-Audio for a source, so need to consider a source with an appropriate converter that eliminates the SPDIF quality problem. I could use a Mac Mini, which has USB output.
The recent thread on the FIFO is interesting, but doesn't seem to really be going anywhere. Bypassing the CS8420 entirely is a consideration, but I don't know how difficult that is.
What I found is by ear and measurement various sources resulted in significant difference with analog output quality. An AppleTV's toslink in combination with Twisted Pear's converter was the worst (no fault of TP, I believe). A MacBookPro's toslink was improved quite a bit, with only a few glitches here and there (sounds like a hiccup in the audio stream). Finally, my M-Audio SPDIF is essentially perfect, so my assumption is I am dealing with an issue of SPDIF quality.
I would rather not have to use my M-Audio for a source, so need to consider a source with an appropriate converter that eliminates the SPDIF quality problem. I could use a Mac Mini, which has USB output.
The recent thread on the FIFO is interesting, but doesn't seem to really be going anywhere. Bypassing the CS8420 entirely is a consideration, but I don't know how difficult that is.
What about installing one of the MiniDSP Ministreamers with the possibility of using the I2S output, bypassing the CS8420 entirely?
miniSTREAMER | miniDSP
I could try to post at the MiniDSP forum if that makes more sense. $55 seems reasonable if I can get I2S, using the internal clock on the DCX.
miniSTREAMER | miniDSP
I could try to post at the MiniDSP forum if that makes more sense. $55 seems reasonable if I can get I2S, using the internal clock on the DCX.
zigzagflux,
I am also interested in bypassing cs8420 in behringer with I2S input,
so I watch this thread with care.
I am also interested in bypassing cs8420 in behringer with I2S input,
so I watch this thread with care.
Scratch that idea; if I bypass the CS8420 and go I2S, I will need to ensure a 96 kHz stream, which I don't have. Since the MiniStreamer does not upsample, all I can do is use the SPDIF output and keep the CS8420. But if the SPDIF is improved over my existing sources, that is fine too.
What can be expected of jitter quality with the MiniStreamer, or for that matter anything that runs USB to SPDIF? Not familiar enough with what goes on in those products.
What can be expected of jitter quality with the MiniStreamer, or for that matter anything that runs USB to SPDIF? Not familiar enough with what goes on in those products.
Zigzagflux,
go on this thread I have very valuable answers
http://www.diyaudio.com/forums/digi...chronous-usb-i2s-interface-8.html#post2717529
Hope it will help
Zoran
go on this thread I have very valuable answers
http://www.diyaudio.com/forums/digi...chronous-usb-i2s-interface-8.html#post2717529
Hope it will help
Zoran
it sounds to me like your hiccups are probably caused by incorrectly setup device drivers wrt buffer size
It would not surprise me; to be honest I wouldn't even know how to change things around to test this. You would think running optical out of a MBP would be driverless in a sense. What options do I have to change this? I can go into the AudioMidi app and change bit number and freq, but that didn't seem to make a difference (either worked or it didn't).
After further reading, it looks like the CS8420 is set up in hardware mode 1. Output is slave I2S 24 bit, which makes sense. Mode is 1A, which I don't understand. I would be interested in going to hardware mode 2, but this requires a number of board level changes to the CS8420 pins. Serial input can be set for I2S with IF2 setting. There is also a CUVEN choice that I don't understand. So it certainly looks possible, but would take someone with more knowledge than I to make sure compatibilities are sound.
After further reading, it looks like the CS8420 is set up in hardware mode 1. Output is slave I2S 24 bit, which makes sense. Mode is 1A, which I don't understand. I would be interested in going to hardware mode 2, but this requires a number of board level changes to the CS8420 pins. Serial input can be set for I2S with IF2 setting. There is also a CUVEN choice that I don't understand. So it certainly looks possible, but would take someone with more knowledge than I to make sure compatibilities are sound.
I may be a little late in responding, but wouldn't the twisted pear Metronome fit the bill? You can convert he incoming I2S signal to a constant 96khz, effectivly bypassing the CS8420. Please let me know if it is possible, as I would be very much interested in building a i2s in, i2s out dcx2496 to feed three external dacs...
link for Twisted Pear Metronome:
The Metronome ASRC Module
TP ASRC in the DCX
I am currenlty using the Twisted Pair ASRC as the input to my DCX. I use the Analog Devices ASRC rather than TI since it can have 5V IO. One difference with the DCX from most applications for the ASRC board is that the DCX uses the ASRC in slave mode. This means the ASRC supplies the master clock and eventually also the data. The bit clock and work clock come from thr DCX to the ASRC. The TP board has switches that alow this configuation to work.
I am currenlty using the Twisted Pair ASRC as the input to my DCX. I use the Analog Devices ASRC rather than TI since it can have 5V IO. One difference with the DCX from most applications for the ASRC board is that the DCX uses the ASRC in slave mode. This means the ASRC supplies the master clock and eventually also the data. The bit clock and work clock come from thr DCX to the ASRC. The TP board has switches that alow this configuation to work.
TP ASRC
The attached pic shows the basic layout. Much of the concept come from the freerider mod available here:
Improved digital input stage for the Behringer DCX2496
ALthough this is a bit different with the crystal on the TP board also supply reference clock for DCX. The DCX internal reference oscillator needs to be disabled. There is a PDF on the Tent Labs site that shows how to do this.
The attached pic shows the basic layout. Much of the concept come from the freerider mod available here:
Improved digital input stage for the Behringer DCX2496
ALthough this is a bit different with the crystal on the TP board also supply reference clock for DCX. The DCX internal reference oscillator needs to be disabled. There is a PDF on the Tent Labs site that shows how to do this.
Attachments
cs8420 and 192khz?
Dear all,
how to explain i input a dcx2496 with spdif 192khz/24bit and it works perfectly triggered, without cliks and pops
I am sure it is 192khz because i measured it exactly at the input.
My behringer is y 2017 maybe they’ve changed the asrc? But it seems not. Or the cs8420 downsamples any higher aes incoming flux to 96khz and upsamples any lower ones?
I read the datasheet and it seems only the serial (i2s) input is limited to 96khz but for the aes input nothing seems to be specified... the mistery remains...
thanks for your reply, GM
Dear all,
how to explain i input a dcx2496 with spdif 192khz/24bit and it works perfectly triggered, without cliks and pops
I am sure it is 192khz because i measured it exactly at the input.
My behringer is y 2017 maybe they’ve changed the asrc? But it seems not. Or the cs8420 downsamples any higher aes incoming flux to 96khz and upsamples any lower ones?
I read the datasheet and it seems only the serial (i2s) input is limited to 96khz but for the aes input nothing seems to be specified... the mistery remains...
thanks for your reply, GM
Last edited: