Stop button should not be instantaneous

Aaron Newsome · October 28th, 2009, 11:05 PM

Hi Lance, knowing a little bit about software development, I can say that even the simplest feature you could think of *could* cost more to develop than the entire cost of a nano.

The testing matrix for the Nano is already ridiculously large I'm sure. Regression testing added features through iterations of Q/A and development is *really* expensive. It's very unlikely that the CD developers are working on the low end of the pay scale for that type of work.

Yes, sad to say that these features that everyone clamors for costs thousands of dollars. Without the economies of scale, they simply would not happen (just ask Mark how tough going it is for his box). Or at least, that's my thought.

A documented SDK for the hardware might not be bad though. Lance and other nano owners could hire other developers to program the features they want if that were available.

Dan Keaton · October 29th, 2009, 06:45 AM

Dear Friends,

I like your suggestions.

In this post I will add more insight into the nanoFlash than ever before.

The nanoFlash has a microprocessor and a Field Programmable Gate Array (FPGA).

The microprocessor uses standard programming techniques, i.e., one develops software one instruction at a time and one instruction is executed at a time.

With this microprocessor, once a software routine is developed and tested, then it can be generally expected to run flawlessly until it is changed, or another routine, or design change, interferes with the original software routine.

This is the standard world of programming.

But, microprocessors have major limitations when one needs to do a massive amount of work in a very short period of time. For one "core" or hyperthread, generally only one instruction at a time is executed. Of course, there may some situations where some microprocessors execute just a few instructions at a time.

The FPGA world is entirely different. One programs the FPGA by building up logic blocks and one can have many logic blocks, but there is a finite number allowed for a given FPGA.

Once a logic block is setup, for one cycle of the FPGA, all of the logic blocks operate in parallel. This would be like a microprocessor that executed all of the program's instructions at once, as opposed to "in sequence, one instruction at a time".

While, with a microprocesor, once a routine is programmed and tested, it can generally be assumed to work if a change to another routine is made, the opposite is true with the FPGA.

Once any change is made, to any part of the FPGA, such as adding a new, unrelated feature, one has to test everything! This explains why we are always testing!

To be clear, some feature requests go into the microprocessor, some go into the FPGA, and quite a few involve both.

A simple request, such as adding a delay to the Stop button, probably goes into the microprocessor alone, but I am quite frequently surprised when some feature requests unexpectedly involve the FPGA.

Due to the nature of the beast, I know of no company that offers a Software Development Kit for their FPGA; it is just too complex.

I hope this helps explain why we test so much and why such a small box is so powerful.

I completely omitted a discussion of the main codec engine, which also does a great deal of the very complex tasks in the nanoFlash.

Lance Librandi · October 29th, 2009, 07:12 AM

Thanks Dan and Aron great insight into the testing and program structure of the NanoFlash.

Mark Job · October 29th, 2009, 07:16 AM

Quote:

Originally Posted by Aaron Newsome

Yes, sad to say that these features that everyone clamors for costs thousands of dollars. Without the economies of scale, they simply would not happen (just ask Mark how tough going it is for his box). Or at least, that's my thought.

.......Right on brother ! I feel as though I've been forced into also adding the special feature of my nervous system to my box ! Man ! This project is not easy. First, it's freakin expensive. Secondly, it's incredibly complex. Thirdly, it's fun to work at, but really stressful. We keep intermittently running out of cash, then we start up again. There's so much about my project I want tio share with the people on this forum, but I am sworn to secrecy about allot of it at this point. However, I won't be silent on the details for ever to be sure.

....There's so many questions we need to answer, like, Do we add a complex MPEG 4 encoding engine or do we just make the box a full uncompressed recording device ?

.....Do we release a product to market which does a few basic things and slowly enable it bit by bit via firmware releases to the end users or wait a really long time and just drop it on the market in the end with everything it can possibly do and that's it ? Time is money.

.....How many bloody SD card slots ? 1, 2, 4, or 6 ? You wouldn't believe the panic of this question !

.....How many bits ? 8, 10, or 12 ? I say 12 bit to facilitate 4:4:4 cameras such as Sony's SRW 9000 and Vipers, and Red 1 and a half's.

Quote:

Originally Posted by Aaron Newsome

A documented SDK for the hardware might not be bad though. Lance and other nano owners could hire other developers to program the features they want if that were available.

...This is good for the end user but bad for the company selling it because then they lose control over design. Funky third party compny could design something which is not part of original designer's vision but would be forced into supporting.

Justin Benn · October 29th, 2009, 05:53 PM

Quote:

Originally Posted by Billy Steinberg

Do you even own a nano? How do you use yours?

Billy

Billy - apologies if I came on too strong. My own fault for posting when tired.

In fact I own an XDR and not a Nano, although I have played with a Nano. My XDR is toted about in a Portabrace bag. I've had an XDR from early on.

I don't have the same frequent (?) studio/signal requirements as you clearly do and can see why an instant stop might not work for you. My use is far more pedestrian and smaller in scope but I'm finding the output of the XDR excellent - so much so that like others I tend to rely on my EX3 SxS output as back-up.

You're completely entitled to find the decision re the 'stop action/button' a bad one but it's clear others don't. I feel sure Dan, Mike and the crew will come up with a solution for you.

Cheers,

Justin.

Aaron Newsome · October 29th, 2009, 06:12 PM

Sometimes when deciding which way to go with features, they may have to side with the masses. Which is scary though, since "the masses" are the ones who picked VHS over Beta,.. if anyone remembers that battle.

Dan Keaton · October 29th, 2009, 08:26 PM

Dear Aaron,

I remember the Betamax well.

I bought my first Betamax, the original Beta 1 Betamax, on November 16, 1976.

I always liked the quality of Betamax. Of course, it does not compare to what we can offer today, but for a home VCR it was just great.

October 29th, 2009, 06:45 AM	#17
Dan Keaton Inner Circle Join Date: Dec 2002 Location: Augusta Georgia Posts: 5,421	Dear Friends, I like your suggestions. In this post I will add more insight into the nanoFlash than ever before. The nanoFlash has a microprocessor and a Field Programmable Gate Array (FPGA). The microprocessor uses standard programming techniques, i.e., one develops software one instruction at a time and one instruction is executed at a time. With this microprocessor, once a software routine is developed and tested, then it can be generally expected to run flawlessly until it is changed, or another routine, or design change, interferes with the original software routine. This is the standard world of programming. But, microprocessors have major limitations when one needs to do a massive amount of work in a very short period of time. For one "core" or hyperthread, generally only one instruction at a time is executed. Of course, there may some situations where some microprocessors execute just a few instructions at a time. The FPGA world is entirely different. One programs the FPGA by building up logic blocks and one can have many logic blocks, but there is a finite number allowed for a given FPGA. Once a logic block is setup, for one cycle of the FPGA, all of the logic blocks operate in parallel. This would be like a microprocessor that executed all of the program's instructions at once, as opposed to "in sequence, one instruction at a time". While, with a microprocesor, once a routine is programmed and tested, it can generally be assumed to work if a change to another routine is made, the opposite is true with the FPGA. Once any change is made, to any part of the FPGA, such as adding a new, unrelated feature, one has to test everything! This explains why we are always testing! To be clear, some feature requests go into the microprocessor, some go into the FPGA, and quite a few involve both. A simple request, such as adding a delay to the Stop button, probably goes into the microprocessor alone, but I am quite frequently surprised when some feature requests unexpectedly involve the FPGA. Due to the nature of the beast, I know of no company that offers a Software Development Kit for their FPGA; it is just too complex. I hope this helps explain why we test so much and why such a small box is so powerful. I completely omitted a discussion of the main codec engine, which also does a great deal of the very complex tasks in the nanoFlash. __________________ Dan Keaton Augusta Georgia

October 29th, 2009, 08:26 PM	#22
Dan Keaton Inner Circle Join Date: Dec 2002 Location: Augusta Georgia Posts: 5,421	Dear Aaron, I remember the Betamax well. I bought my first Betamax, the original Beta 1 Betamax, on November 16, 1976. I always liked the quality of Betamax. Of course, it does not compare to what we can offer today, but for a home VCR it was just great. __________________ Dan Keaton Augusta Georgia

October 28th, 2009, 11:05 PM	#16
Aaron Newsome Major Player Join Date: Jul 2009 Location: Rancho Santa Margarita Posts: 590	Hi Lance, knowing a little bit about software development, I can say that even the simplest feature you could think of could cost more to develop than the entire cost of a nano. The testing matrix for the Nano is already ridiculously large I'm sure. Regression testing added features through iterations of Q/A and development is really expensive. It's very unlikely that the CD developers are working on the low end of the pay scale for that type of work. Yes, sad to say that these features that everyone clamors for costs thousands of dollars. Without the economies of scale, they simply would not happen (just ask Mark how tough going it is for his box). Or at least, that's my thought. A documented SDK for the hardware might not be bad though. Lance and other nano owners could hire other developers to program the features they want if that were available.

October 29th, 2009, 07:12 AM	#18
Lance Librandi Major Player Join Date: Aug 2009 Location: South Australia Posts: 374	Thanks Dan and Aron great insight into the testing and program structure of the NanoFlash.

October 29th, 2009, 06:12 PM	#21
Aaron Newsome Major Player Join Date: Jul 2009 Location: Rancho Santa Margarita Posts: 590	Sometimes when deciding which way to go with features, they may have to side with the masses. Which is scary though, since "the masses" are the ones who picked VHS over Beta,.. if anyone remembers that battle.