Variac said:MODERATOR"S NOTE:
Many recent posts on this thread have been moved to this new thread:
"ESS Sabre Buffalo DAC"
http://www.diyaudio.com/forums/showthread.php?s=&postid=1521516#post1521516
because they are discussing the Buffalo DAC board and associated products created by Twisted Pear Audio. Discussions regarding these products should take place on the new thread, while posts regarding the ESS Sabre DAC generally should still take place here.
Thank You,
Variac
Thanks for doing that... had to be a fun task
downsampling
... I agree with You. Indeed Sabre's upsampler is a very good one! Still, I have one upsampler and I want to play with it... so I wonder if it's possible to "cut" the 384 fs digital signal (after external upsampler) in two, in that way it will allways be 192 fs going directly into the Sabre: one sample used, one ignored and so on. Can I do that? What sonic impact will have this approach?
Unfortunately I can't change the upsampler's output SR: it will allways be above 192 KHz!
On my board all the chips/modules works synchronous (External upsampler's Master Clock and Sabre's MCLK) so, I hope the clock delays are reduced to minimum.
Best regards!
... I agree with You. Indeed Sabre's upsampler is a very good one! Still, I have one upsampler and I want to play with it... so I wonder if it's possible to "cut" the 384 fs digital signal (after external upsampler) in two, in that way it will allways be 192 fs going directly into the Sabre: one sample used, one ignored and so on. Can I do that? What sonic impact will have this approach?
Unfortunately I can't change the upsampler's output SR: it will allways be above 192 KHz!
On my board all the chips/modules works synchronous (External upsampler's Master Clock and Sabre's MCLK) so, I hope the clock delays are reduced to minimum.
Best regards!
Re: downsampling
You could try that and it would most likely be OK because the FS is well above the audio band.
Theoretically speaking...
In the theory on decimation you need to filter first. Say you were going to decimate 44K data to 22K. You would need to filter everything above 11K in the original before you decimate, or you will get nasty images in the result.
In your case I assume the data has no audio in range above or near the final FS/2 or 96000 Hz. So it should be OK. Try it and tell us if it works
excelon said:
You could try that and it would most likely be OK because the FS is well above the audio band.
Theoretically speaking...
In the theory on decimation you need to filter first. Say you were going to decimate 44K data to 22K. You would need to filter everything above 11K in the original before you decimate, or you will get nasty images in the result.
In your case I assume the data has no audio in range above or near the final FS/2 or 96000 Hz. So it should be OK. Try it and tell us if it works
Remember to check in here for Buffalo discussion:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=128137&perpage=25&pagenumber=8
I wasn't automatically subscribed. the idea was that everyone subscribed here would be subscribed there..... Better go there and subscribe or at least make sure you are...
Mark
http://www.diyaudio.com/forums/showthread.php?s=&threadid=128137&perpage=25&pagenumber=8
I wasn't automatically subscribed. the idea was that everyone subscribed here would be subscribed there..... Better go there and subscribe or at least make sure you are...
Mark
Re: Re: downsampling
Based on the results Tony Faulkner had with his averaging
down sampling program to get from 176.4 to 44.1, it might be worth
trying this same approach. Just take the average of every two
samples, forget any LPF.
cheers
Terry
rossl said:
You could try that and it would most likely be OK because the FS is well above the audio band.
Theoretically speaking...
In the theory on decimation you need to filter first. Say you were going to decimate 44K data to 22K. You would need to filter everything above 11K in the original before you decimate, or you will get nasty images in the result.
In your case I assume the data has no audio in range above or near the final FS/2 or 96000 Hz. So it should be OK. Try it and tell us if it works
Based on the results Tony Faulkner had with his averaging
down sampling program to get from 176.4 to 44.1, it might be worth
trying this same approach. Just take the average of every two
samples, forget any LPF.
cheers
Terry
Would like to use this DAC for my new HT/ music setup
You guys know of any PRE/RECEIVER or CHIP
that can processe most multichannel digital signal
and output seperate digital signals ?
( one in wich we could setup the number of channels to be played would be perfect )
can this DAC accept seperate channel signal for 6-8 channels ??
else i could use 6 seperate DACs
You guys know of any PRE/RECEIVER or CHIP
that can processe most multichannel digital signal
and output seperate digital signals ?
( one in wich we could setup the number of channels to be played would be perfect )
can this DAC accept seperate channel signal for 6-8 channels ??
else i could use 6 seperate DACs
Dual Stereo instead of Dual Mono
Hi Russ,
After reading this comment from Dustin I got this idea for getting lower Zout(for voltagemode) and noise figures from the Buffalo :
I thought of feeding two Buffalos from your SPDIF receiver and using a separate clock, common for the two modules.
Then parallell the outputs from the two DACs to get one stereo output.
This would give us a Dual Stereo (as Dual Mono seems to be a nono).
If doing this with four Buffalos it would give us even better results in theory........
I know this migthbe a gigantic overkill(if it works), but will it work IRL?
Hi Russ,
After reading this comment from Dustin I got this idea for getting lower Zout(for voltagemode) and noise figures from the Buffalo :
I thought of feeding two Buffalos from your SPDIF receiver and using a separate clock, common for the two modules.
Then parallell the outputs from the two DACs to get one stereo output.
This would give us a Dual Stereo (as Dual Mono seems to be a nono).
If doing this with four Buffalos it would give us even better results in theory........
I know this migthbe a gigantic overkill(if it works), but will it work IRL?
Hi!
I am doing some "serious" PIC programming, also to communicate with Sabre DAC.
I can set all the registers at init time but then it comes to set the volume for all 8 DACs and I want it to do it as fast as possible.
I can simply do something like:
But I do not like it since it does i2c start/stop procedure. Is it possible to use?:
This would be more efficient.
I2C states that is should work, but I guess Sabre does not increment register address automatically :S
Cheers,
Matej
I am doing some "serious" PIC programming, also to communicate with Sabre DAC.
I can set all the registers at init time but then it comes to set the volume for all 8 DACs and I want it to do it as fast as possible.
I can simply do something like:
PHP:
for (i = 0; i < 8; i++) i2c_send(0x90, i, volume_value);But I do not like it since it does i2c start/stop procedure. Is it possible to use?:
PHP:
i2c_send(0x90, 0, array_of_volume_values);This would be more efficient.
I2C states that is should work, but I guess Sabre does not increment register address automatically :S
Cheers,
Matej
HI MatejS,
If you want to go faster, trying runing the I2C at higher than 400kHz, even over 1Mhz it should be fine, over thing that will limit you it the rise time cause by ther pulllup. If your bold, and need extreme speed for whatever reason, when sending break the I2C protocol a little by driving the buss high upon sending and just ignore the acknowledge bit. This semms overkill to me, but it might just work. No garauntees though.
Dustin
If you want to go faster, trying runing the I2C at higher than 400kHz, even over 1Mhz it should be fine, over thing that will limit you it the rise time cause by ther pulllup. If your bold, and need extreme speed for whatever reason, when sending break the I2C protocol a little by driving the buss high upon sending and just ignore the acknowledge bit. This semms overkill to me, but it might just work. No garauntees though.
Dustin