Bogus claims Re: OTB summing from Black Lion

[Nick Sevilla]

His response:

First, the summing done in the digital domain requires the use of DSP. As you may know, DSP processes are not conducted in a vacuum, and they are not all the same. They are subject to real-world limitations, such as DSP math processes, logic switching and current movements, gate input capacitances, cross-talk related anomalies, resonant poles on the supply line, etc. When we say "clock-based errors," we are not referring sample or master clock signals, we are talking about issues created within DSP by logic switching noise, eye pattern errors, cross-talk issues, and other real-world imperfections that manifest themselves in our otherwise simple and ideal "sum all samples at a time" scenario.

This my friend, is a guy who knows computer lingo, but not what it actually means.
B_S... pure and simple.
Here's a layman's explanation:
He is trying to convince you that a CPU, or the other mani computer chip, the Math chip, which deal exclusively in binary code, and also has ERROR CORRECTION... is somehow making errors large enough to affect audio processing.
And THEN he is connecting this to a word clock chip on a modified converter box, which lives OUTSIDE the computer, and does NO DSP WHATSOEVER TO YOUR MIXED DIGITAL SIGNAL.
DSP = Digital Signal Processing. Got it? DIGITAL means ones and zeroes.
ADC = Analog to Digital Converter. Got it? ANALOG signal converted to ones and zeroes.
They are different processes.
I say BS. Marketing.
Tell him to prove it in a lab... and NOT their lab, an independent one.

[wesley.wittich]

+1 for posting the whole email conversation w/ BLA

I'm finding this thread extremely interesting...I'm not buying BLA's assertions. It would seem to me that the greatest imperfection with digital is the ADC, since it is essentially a computer estimating a sound wave (albeit quite accurately). Once it is converted and saved as binary information, I can't imagine too many issues from that point, especially audible ones. From this perspective, recording to digital and then summing through analog equipment would add more 'issues' as far as digital artifacts because it would go through two more steps of conversion, one from D to A for summing, and then back from A to D for its final format (CD/back to PT/etc). As someone else pointed out, no matter what, the computer is involved in the summing in some way, so there is no way to avoid these "clock-based errors" if they do actually exist (which I doubt).

[The Scum]

So if this mixer magically fixes "DSP clock errors," it's of no benefit to use it to sum analog source material, right?

It seems like they're trying to say something like "mixing ITB is too perfect, and some of the atrifacts of analog may actually be desirable," but it's been pretty solidly wrapped up with undecipherable cybercrud.

Computers, by definition, are linear processes: given the same set of inputs, they'll generate the same set of outputs. A + B + C today equals the same A + B + C as it will tomorrow.

Though there are times that this isn't true.

If computers didn't reliably produce the same calculated results all the time, they'd be nearly useless. Clocks would run backwards, planes would drop form the sky, the trains wouldn't run on time, and cars would mysteriously accelerate.

And this is really only in reference to the quality of individual calculations the computer is performing. Those calculations are done as requested by the software running on the computer, and getting software to use the CPU effectively is an artform in itself.

And to confirm some prior comments:

There are indeed to clocks at play - the sample work clock, and the CPU (or DSP) clock. As long as the CPU is sufficiently faster than the word clock, it can complete the computation of samples before the next samples arrive (or are needed by the DAC), and things work OK in real time.

If the DSP being performed can't be done in real time, it could always be done in non-real-time. I remember the old days when we'd run Csound overnight to get the results calculated on a 66MHz 486. I'd imagine my new i5 could do it in real time...but the results would be the same.

[wesley.wittich]

So if this mixer magically fixes "DSP clock errors," it's of no benefit to use it to sum analog source material, right?

This isn't what I gathered from reading their website...I understood them to be saying that the mixer simply avoids the DSP clock errors by summing in the analog domain, but the consensus of the board is that these DSP clock errors are either non-existent or completely insignificant, so it's only real use would be to "sum analog source material."

It seems like they're trying to say something like "mixing ITB is too perfect, and some of the atrifacts of analog may actually be desirable,"

I hope this is what they might be trying to say, but if so, their English skills blow, because they are saying the exact opposite...

[/i]

[Scodiddly]

Audio software programmers use the same languages and libraries as spreadsheet programmers. If audio was getting goofed up by clocking issues, spreadsheets would get munged up too. Yet nobody seems to be complaining about that happening in spreadsheet world.

If it's an ADC or DAC, then clocking matters. If it's DSP calculations, clocking does not matter.

[Snarl 12/8]

That one paragraph (and the gist of the rest of their position) sounds like they are conflating the attributes of digital and analog summing AND the attributes of computers and analog circuitry in an almost completely random way. The problem is that all of these systems have to obey the laws of physics and are therefore almost completely non-random, isolated from each other and understood.

One thing came up on another forum (related to IT, not audio) that I participate in. Apparently cosmic rays can have seemingly random effects on computer processes. They can hit RAM or CPU's and generate random and uncorrectable errors. IBM investigated this in the 70's and started heavily shielding their servers afterwards. I only mention this because I could see someone observing this phenomenon without understanding the cause and coming up with the sophistry above to try to explain how their snake oil helps alleviate it. Does BLA add extra shielding in their mods?

Edit: I really wish BLA had said "Oh yeah, you're right, we better fix that" when first contacted. This is sortof a weird, lengthy discussion in response to what, ultimately, is mainly a slight wording mishap on a web page. The cover up is always worse than the crime. I always tell people (and web is my day job) that they simply can't believe everything they read on a web page. I've seen, over and over that institutions have a hard time getting what they really know to be true up on their web sites. It's like a huge game of "telephone" within an organization to get the institutional knowledge through the web team intact. I mean BLA clearly knows how this stuff works, or their products wouldn't be so [generally] respected around here.

[KennyLusk]

Just to play devil's advocate a bit, and with no intention of making any claims, I want to throw something into the conversation:

Like many AMD fans, I have always overclocked my AMD CPU chips, and have usually recognized AMD math co-processors as being superior in performance and reliability to Intel chips. The gap has closed between the two I know and there have been a few AMD chips that didn't quite make the grade, I know. But putting that aside, I have noticed that sometimes my settings will revert back to the slower default clockspeed, and as a result I experience errors in video and audio processing when this happens. I simply bounce back into Bios and overclock the CPU again and voila, everything returns to normal. There seems to be a connection at those times between optimal clock speed, FSB speed and memory (RAM). Bottlenecking at the FSB? Perhaps.

Anyway, like most of you I'm a little disappointed BLA hasn't bounced in here to address this issue. Maybe they feel they don't have to.

[leigh]

Edit: I really wish BLA had said "Oh yeah, you're right, we better fix that" when first contacted. This is sortof a weird, lengthy discussion in response to what, ultimately, is mainly a slight wording mishap on a web page.

Yeah, me too.

I've given BLA a heads-up on this thread now, so maybe they will chime in.

I had initially assumed that it was just a wording mishap, or a copywriter gone wrong. However, in the email responses I've gotten from BLA, they are very much standing behind their claim. They have thrown a lot of jargon around, guessed that my "real-world experience with this subject is limited to textbooks", claimed that Bob Katz agrees with their position, and have not answered a couple very basic questions that I posed to them. Not what I expected from a company with DIY grassroots.

Leigh

[jgimbel]

Yeah..I was kind of blown away to see that response from them. I've never dealt with them personally, but from everything I've heard about them and all the praise they get, I was expecting this to be a quick "Oh wow, our mistake! These product blurbs go through the filters of a number of different people, not all of whom are connected to audio recording, so things can get off here and there". Never thought this is the road it would go down!

[leigh]

The latest email from BLA said that they are aware of this thread, and that they are "not interested in continuing the discussion".

Not a satisfactory conclusion. I am thinking of how to proceed.

Leigh

[armanbohn]

Hey after we're done with BLA can we head over and bash on FMR for a while? Their webpage for the RNC states that

SuperNice mode provides near-invisible signal compression

and I suggest that is categorically incorrect. It is entirely invisible. It is audible.

After that, let's all shamble over to the site for Cascade Mics and nuke it from orbit for saying that the Fat Head ribbon mic is

99% Pure aluminum

I challenge them to verify that the ribbon in my mic is actually composed this exact percentage. A class action suit will follow if they can't immediately respond and send someone to my door to verify this. Once the content of their website copy has been verified/disproved, I'm sure that the performance of my mic will change based on the posted specs.

What I would really like to see is the complete ellimination of all small American companies that pretend that they can compete with huge faceless corporations that never directly communicate with their customers. Maybe if we can get these American losers to waste more of their time answering emails from people who aren't even remotely interested in buying their products, we can help help drive 'em out.

AND FURTHERMORE (in regards to 'puter errors):

Everybody knows computers DON'T have issues with calculations. Just ask my P4 Denormal bug. Shure he's old, but he's reeeeal cute.

[Snarl 12/8]

The latest email from BLA said that they are aware of this thread, and that they are "not interested in continuing the discussion".

Not a satisfactory conclusion. I am thinking of how to proceed.

Leigh

Why can't you let it go? This thread exists for googlers curious to dig deeper into their claims. I'm not saying you're wrong, but wondering why you're compelled to "solve" this somehow.

[leigh]

Hey after we're done with BLA can we head over and bash on FMR for a while...

Whatever size company I'm dealing with, I find it profoundly irritating when someone tries to smokescreen me with a bunch of irrelevant jargon. Or, rather than directly addressing a question, brings up a bunch of tangential facts. It's like listening to a politician field a skewering question - rather than answer it, they work to quickly re-frame the issue.

I have had nothing but positive dealings with FMR and Cascade. I don't go around bashing "small American companies", and you won't make me feel guilty about following up on this issue.

And I do even appreciate the fact that BLA took the time to reply to my questions, at length. I simply do not understand why they have seemingly chosen to confuse and obscure with their answers.

Leigh

[leigh]

Why can't you let it go? This thread exists for googlers curious to dig deeper into their claims. I'm not saying you're wrong, but wondering why you're compelled to "solve" this somehow.

See above - having been irritated by the smokescreen, I find it hard to put the issue down. And I feel like other pro audio people with less technical knowledge would just go cross-eyed from all the jargon, and simply accept the answers from BLA out of confusion. And that would be sad and wrong.

That said, I don't have time to make this my life's crusade. Perhaps I should just post my email exchange with BLA, and be done with it. Then you, and future Googlers, can choose to believe BLA or not on their own terms.

Leigh

[Snarl 12/8]

I didn't mean to pile on. I wrote my question while that other dude was writing his thing and cross posted. Like I said, I don't think you're wrong. Just wondering what your reasons were for really pursuing this. I always say "you can't understand 'crazy'". Sometimes it can be detrimental to yourself to try. And, for the record, I totally think you're right and they're wrong here. That's not my point.