• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Buffalo II

Status
Not open for further replies.
Jump to Latest
#941
I know, really all it does is make the chip not DIY friendly.

BTW despite what some are saying ESS has not relaxed their NDA policy in the least. I am in contact with them regularly. Distributors have been clouding this issue giving people permission to do things when they have no authority to do so. The NDA is not with the distributor, it is with ESS. The distributor is only responsible to make sure it is signed, they have no authority to alter the agreement. *ONLY* ESS can authorize anything regarding the datasheet and the NDA.

Cheers!
Russ
 
Last edited:
#944
Russ...

Cool to hear that you have been experimenting with filters a bit. I, and likely many other, DIYers am (are) not knowledgeable at all when it comes to controller interfaces, and writing custom code. It would be really cool if TPA could offer some custom filter options for the B-II someday (I understand that you are busy). I suppose these could be incorporated in the controller PIC chip for distribution?
I have quite a bit of experience listening to the different filters on the PS Audio PerfectWave DAC (Wolfson 8741 filters) and although the differences are subtle, they certainly are musically relevant. I suspect with the high resolution of the B-II, the differences in filters could make for some nice fine tuning of the sound.
 
#946
Dual Mono Buffalo Configuration

I'm trying to figure out how to adapt a dual-mono Buffalo DAC configuration to the existing I/V section in my CDP without adding external dual Ivy I/V modules as per TP hook-up diagram. Am I correct that there is a software strap on Buffalo that allows all of the internal DACs in one module to be dedicated to one channel? Thus one Buffalo module would provide left channel output and the second module would provide right channel output? (The remaining output connections would be unused.)
 
Last edited:
#947
Russ White said:
While it is possible it is not advisable.

....
You can you one dual trafo and share the same secondary that is used for V+ on the bipolar supply for the digital supply. Be very careful here, if you use the other secondary you will make smoke.
...
Click to expand...

Yes. When playing with the secondaries, have a fuse before the transformer. If the fuse blows, you connected the secondaries wrong...
 
#948
Russ White said:
Unfortunately I am afraid if I persue this I am going to get into NDA trouble. 🙁

The short of it is all the info you need is in the DS.

I have created several filters using Octave. I have been reluctant to promote that feature for DIY until I get permission to discuss that portion of the datasheet.

Cheers!
Russ
Click to expand...

Can you share the coefficients since they are generated with another application?
 
#949
Russ White said:
I know, really all it does is make the chip not DIY friendly.

BTW despite what some are saying ESS has not relaxed their NDA policy in the least. I am in contact with them regularly. Distributors have been clouding this issue giving people permission to do things when they have no authority to do so. The NDA is not with the distributor, it is with ESS. The distributor is only responsible to make sure it is signed, they have no authority to alter the agreement. *ONLY* ESS can authorize anything regarding the datasheet and the NDA.

Cheers!
Russ
Click to expand...
There may be hundreds/thousands of signed NDA agreements, how can they know who has disclosed the documentation to the internet? Once there, none of us have signed any NDA 😎
 
#955
Russ White said:
I am sure I could, but that would not be very meaningful to the people here trying to learn the how and why of it.

But I may in the end just do that.
Click to expand...

I suppose you could share at the high level their characteristics (linear phase, minimum phase -no pre-ringing, brickwall, apodizing - roll off before 20K and fully attenuates at Fs/2, and others I don't know about). It would be nice to play with minimum phase and apodizing linear/minimum phase filters.

Thanks.

PS: When playing with the 5 filters in OPUS s/w mode, it was hard to differentiate them.
 
#956
glt said:
I suppose you could share at the high level their characteristics (linear phase, minimum phase -no pre-ringing, brickwall, apodizing - roll off before 20K and fully attenuates at Fs/2, and others I don't know about). It would be nice to play with minimum phase and apodizing
Click to expand...

So long as the filter response at audible aliasing frequencies is well down, then who cares whether Fs/2 is fully attenuated?

For example, take a low-pass filter running at 88.2 kHz providing oversampling for a zero-stuffed 44.1 kHz signal. These are usually designed as half-band filters (which have the property of only needing to actually filter the odd samples for oversampling since the even samples are merely the input), with Fs/2 at -6 dB, and rolloff beginning at 20kHz. By symmetry the filter will reach -120 dB at 24.1 kHz, for a transition bandwidth of 4.1 kHz - pretty darn sharp at 88.2 kHz, and with a fair bit of pre-ringing.

Now consider a different filter, not half-band, with rolloff beginning at 18 kHz, -36 dB at 22.05 kHz, -100 dB at 25 kHz, and -120 dB at 26 kHz. Note that aliasing from 25 kHz will wrap down to 19.1 kHz, but once again at -100 dB from the original signal. The aliasing at 22.05 kHz is of no audible consequence since very few people can hear at all up there, and it's almost -40 dB anyway. This will likely not be a major issue. The filter is half the length at 93 taps vs 187 taps for the first example, and the impulse response damps down to 0.01 in about a third the time of the first filter.

Long story short, aliasing is not a major issue so long as it's kept well down in the baseband. Besides, if ultrasonic aliasing were important, people would run screaming from half-band filters since they're only 6 dB down at Fs/2.

Where people miss the boat with FIRs is in thinking that extreme slopes are always a Good Thing. It ain't necessarily so. Someone once asked me to design some lowpass filters with a 1 kHz transition band from flat to maximum attenuation. I listened to them, then told my manager: "These things sound like cr*p. For lack of a better term, they're *scratchy*. Please let me design them with a wider transition band." He agreed; they sounded much better than the first cut and required less computation (which saved my butt when we started running out of cycles).

Similarly with speaker crossovers: I encourage people to use the shortest FIRs possible for the job, but that's a discussion for another forum.
 
Last edited:
#958
glt said:
I understand your argument. I was merely repeating what is being used out there. Care to design some filters for Bufflao II for the rest of us to play with?
Click to expand...

How do you think I know this stuff? 🙂 Those numbers are not from thin air. Only problem is, I need to know how the DAC uses its filters before going further, as throwing a bunch of coefficients at the thing won't do any good if the structure differs from my assumptions.
 
#959
Dual Mono Config

David Garretson said:
I'm trying to figure out how to adapt a dual-mono Buffalo DAC configuration to the existing I/V section in my CDP without adding external dual Ivy I/V modules as per TP hook-up diagram. Am I correct that there is a software strap on Buffalo that allows all of the internal DACs in one module to be dedicated to one channel? Thus one Buffalo module would provide left channel output and the second module would provide right channel output? (The remaining output connections would be unused.)
Click to expand...

Russ, as I am trying for the group buy today, any thoughts on this question are appreciated.

Thanks,

Dave
 
Status
Not open for further replies.
Top Bottom