TOPIC:

Dear All

A new build for Device console v1.1.14 was just pushed to our servers. It will fix a host of small bugs from the previous build. In an effort to restore a more stable behavior, we completely rewrote the boot up mechanism to improve on the issues we noticed on the field. Please upgrade to latest build by opening device console with an internet connection, a warning that a new build is available will pop up. Once upgraded and DC rebooted, it will prompt you to refresh/upgrade the firmware on the first connection. Clicking REFRESH XMOS/DSP will do similar. 

Once again an immense thank you to all the valuable community feedback and logs we've received on our support portal from each of you who have been patiently helped behind the scenes. As mentioned before, some issues aren't easy to spot as they are unique to a specific configuration, hardware platform, timing of boot up. We are extremely grateful for all your help to positively advance/solve an issue together. 

Please contact us if you have any questions/Issues. 

DevTeam

FWIW I've installed this yesterday and it seems to be working fine for me (that said, 1.1.12 and 1.1.13 were also OK here).  
 

I have been trying to upgrade to 1.1.13 and 1.1.14. Both times it says do you want to upgrade, I say yes. It says reboot I say yes and nothing happens.
Mac on Intel build

@devteam — please elaborate on the 40 bit filter thing you mentioned earlier here .

Nowhere in the Device Console changelog does it speak to 40 bit anything, and some of us are under the impression the hardware in the SHD is only capable of 32 bit at most. 

Where in the signal flow does 40 bit exist?

Along with answering here, I would suggest you update to the signal flow overview to show where things get resampled, upsampled, and downsampled and to what specs. 

I have been running 1.1.13 for sometime now with no issues.  I attenuate the input (prior to ASRC) to avoid intersampling peaks and it seems to work fine.  I don't want to loose this ability so I have not updated to 1.1.14.  What are the concerns with the version I am running?  Again, I have had no issues myself.   

@Devteam: can you elaborate on the improvements made on the filters section? Do I need to reload / redefine the biquads?

MihaiP wrote:

@Devteam: can you elaborate on the improvements made on the filters section? Do I need to reload / redefine the biquads?

I just sent them a ticket today asking for them to follow-up. I sure didn’t expect them to disappear off their forum for a month right in the midst of all the big software changes and customer questions and reported problems. Who does that?  

Sorry we had a very busy few days before the news release and last minute testing... We're always happy to answer all tech support queries you have so make sure to send a support tix if any urgent question.  

In answer to the question on the new build and filter precision at 40bit, we basically "extended the precision". That's how we're achieving better noise floor performance when doing equalization. Grab an APX and you'll see a difference. You might even hear it or maybe not.. When doing floating point math, precision matters. If you "extend" the precision, you get better performance. You don't have anything to do from your side. Think of it of running an application at 64bit that used to run at 32bit. Nothing changed from the user perspective and 32bit is indeed the path. You can however get extended precision on some specific math equation that matters (i.e. the filters). That takes lots more resources but that's what we did here for all platforms. 

Hoping this clarifies. 

devteam wrote:

Sorry we had a very busy few days before the news release and last minute testing... We're always happy to answer all tech support queries you have so make sure to send a support tix if any urgent question.  

In answer to the question on the new build and filter precision at 40bit, we basically "extended the precision". That's how we're achieving better noise floor performance when doing equalization. Grab an APX and you'll see a difference. You might even hear it or maybe not.. When doing floating point math, precision matters. If you "extend" the precision, you get better performance. You don't have anything to do from your side. Think of it of running an application at 64bit that used to run at 32bit. Nothing changed from the user perspective and 32bit is indeed the path. You can however get extended precision on some specific math equation that matters (i.e. the filters). That takes lots more resources but that's what we did here for all platforms. 

Hoping this clarifies. 

I appreciate the response, but to be honest this is going over my head. 

1) You know none of us has a super expensive APX analyzer, but you do. Any chance you want to show what you’re talking about with screenshots from your measurements?

2) Is this just relevant to PEQ or Dirac as well?

3) At any point is the signal itself upsampled to 40bit, or just that the filter precision is 40bit and it’s applied to 32bit? I don’t know enough to know whether that’s ridiculous, but you’ve only confused me with what you said.

Why don't you try googling the topic?  There are plenty of references on this.
Or, you could check with your buddies over at ASR and they'll explain it to you.

That said, how well the math extends to real-world performance with real-world hardware is certainly a valid inquiry.

Dave.

dreite wrote:

Why don't you try googling the topic?  There are plenty of references on this.
Or, you could check with your buddies over at ASR and they'll explain it to you.

That said, how well the math extends to real-world performance with real-world hardware is certainly a valid inquiry.

Dave.

You’re a pleasure, as always. Thanks for descending from your lofty perch. 

My questions are asking about their specific implementation, and asking for clarification where they weren’t all that precise. 

I know this pains you, but there are a lot of people around here who don’t know everything about everything like you do, and who I think would benefit from miniDSP spelling this out a little. Counting myself as one.

Here you go....I'll do the search for you.
search.brave.com/search?q=extended+precision
en.wikipedia.org/wiki/Extended_precision

Dave.

dreite wrote:

Here you go....I'll do the search for you.
search.brave.com/search?q=extended+precision
en.wikipedia.org/wiki/Extended_precision

Dave.

Super helpful. Thanks. I really appreciate the section where they explain how miniDSP is managing this is in the SHD, and how it relates to Dirac filters and anything else we’ve previously understood about the SHD’s operation that may have changed with the latest firmware.

It’s great you take time out to let people know when their questions are beneath you, and to constantly protect miniDSP from just simply explaining themselves to their customers. Keep up the good work. 

'Sorry, but your agenda is well established here.
You can't expect me (and maybe some others) to take you seriously based on your track record.

Dave.