#timvideos

xfxf: did you see what CarlFK did in the logs?

Hi cyrozap! Are you new?

xfxf: I'll get you recording sheets asap

CarlFK: i'm working around the need for them for now (conference already has started)

more getting an instance actually working i care about

mithro: read the logs, not sure what part you mean - you mean the fact carl recorded a talk with it? if so, woo!

as soon as the device works for me i'll be using it plenty too

xfxf: yeah

@#$@#$ "start": "10:30:00", that is not a datetime :p

why does everyone want to make their own export?!!!

rant rant rant.. I am going to start offering $1000 discount for using mainline symposium

and another discount if they supply a patch

xfxf: http://veyepar.nextdayvideo.com/main/recording_...

mithro: I'm here because of the Opsis. It looks pretty interesting!

cyrozap: Cool! How did you find out about the board?

Crowd Supply

I'm really glad that the Opsis now exists as a fully-Free video switching platform, but I question some of the design decisions that were made

cyrozap: I'm Tim Ansell and designed the board, so feel free to ask!

CarlFK: ta! if you can get your Vagrant file working so 'vagrant up' works would be appreciated

at least remove the 'rm' line, shouldn't ever be needed (it's part of provisioning, you only ever provision a specific instance once, removing existing first doesn't make sense)

xfxf: do you plan on encoding in the VM?

yes

but i can handle compiling up a new ffmpeg/melt fine

as well as manually creating + mounting another disk instance in the virtualbox if required

mithro: Oh, ok, well, for starters, why use a Spartan-6 when the Artix-7 uses less power, is less expensive and more powerful, and doesn't require an outdated IDE (ISE)?

cyrozap: Couple of reasons, when we started the project (~2 years ago) the Artix-7 wasn't readily available and our existing prototyping platform was a Spartan 6 and we didn't want to change up to much for our first attempt at doing hardware.

cyrozap: If the Opsis is successful there is no doubt we will look at doing an Artix-7 based board in the future.

mithro: Ah, I see, that might also answer my next question. Also, I apologize if that question came across as a little accusatory.

(veyepar)juser@negk:~$ dvsource-v4l2-other

usage: dvsource-v4l2-other [-d DEVICE] [-c CAPS] [-s {ntsc,pal}]

works! woo, first working package!

cyrozap: Nah - If we were doing the design from scratch today we would probably be using the Artix-7

cyrozap: I actually dislike Vivado more than ISE - both are a pile of crap, atleast ISE has been around for long enough that most of the major bugs have answers on stackexchange.com :P

mithro: I almost never use the IDEs since you can run all the steps with a Makefile, so I care more about the synthesis, mapping, and P&R bugs getting fixed than any of the IDE's features

cyrozap: ISE also does Verilog+VHDL together easily which I believe Vivado doesn't yet do

mithro: Anyways, here's my next question: Why use a Realtek Ethernet PHY whose documentation can only be acquired through an NDA or piracy when companies like Micrel produce PHYs with freely-available documentation (you don't even need to register if you know the direct URL)?

cyrozap: Ability of my manufacture to source them them at the right price and the fact they are a simple PHY that doesn't really "do" much.

cyrozap: And the fact they already use them in other boards they do means they have that supply chain set up

xfxf: I am having other problems - you may want to follow along /j #vagrant

mithro: I kinda figured it was a supply chain/price thing, like how Numato also seems to like to use PICs instead of a microcontroller with a fully-free toolchain and supporting libraries (PIC requires some proprietary Microchip libraries).

cyrozap: I've always used PICs with FOSS tools

cyrozap: It's always annoyed me the bad rap that Microchip gets compared to Atmel. Microchip was significantly more supportive of the hobbyist for a long time. They had much better ICs for significantly cheaper and having a program which would ship anyone in the world free samples. They where also happy to deal with someone who only wanted to order 20-30 of

a part.

mithro: I know you can use SDCC, but don't you need at least one binary blob for initialization?

cyrozap: Not for the PICs I was using 10 years ago - dunno how things are these days

mithro: https://web.archive.org/web/20130303092950/http...

Title: SDCC Library Licenses - SDCC wiki (at web.archive.org)

mithro: "Microchip requires that 'The header files should state that they are only to be used with authentic Microchip devices' which makes them incompatible with the GPL."

mithro: So I guess it's technically open source, but non-free

cyrozap: IIRC There are header files you can use which are not derived from Microchip's code but they are less complete

cyrozap: anyway - No PICs in this design, a previous design did have a PIC for doing power management and add a bunch of LEDs and stuff - but we ditched that

mithro: Yeah, I was really happy about the choice to use the FX2

cyrozap: I'm not, its expensive and only USB2.0 - but again we didn't want to change too much on our first board

mithro: Well, I meant I was excited since it can run free firmware, but I'm not aware of any USB 3.0 controllers that will do that

cyrozap: See the Daisho project! https://github.com/mossmann/daisho/tree/master/...

Title: daisho/sw/fpga/common/usb3 at master · mossmann/daisho · GitHub (at github.com)

mithro: Oh, neat! I was hoping you'd prove me wrong with something for the FX3 since my BladeRF uses that, but this is good for new designs!

cyrozap: FX3 seems to have a smattering of FOSS firmware stuff

cyrozap: I'm interested in trying to get someone to port the daisho firmware to use high speed transceivers rather then needing the external TI phys

cyrozap: But I only have so much disposable income to fund developers

cyrozap: FYI - The Opsis and TimVideos projects are my hobby not my day job :)

cyrozap: Are you "Nuand" then?

mithro: No, I just backed their kickstarter

cyrozap: ahh - I was hoping that "my BladeRF" was "my design for the BladeRF" :)

mithro: Haha, I wish! Unfortunately, my PCB design skills aren't _that_ good.

cyrozap: so what is your interest in the Opsis board?

mithro: Well, when I first saw it, I thought it would be great as a replacement for the BlackMagic rack-mounted HDMI switches that are used at the company I've been interning at

mithro: They're $1000, and yet they're pretty buggy

the atem TV studio?

xfxf: Yeah, that's the one

yeah, i use them currently for HD capture, not ideal for a number of reasons

xfxf: What do you mean? I haven't used one myself, so I don't know exactly what issues it has.

proprietary/non-extensible, require either dedicated HW to control them or OS X / windows software (not linux), can often get into strange states where they freeze + need power cycling, don't do pass through (so need extra hardware to split out HDMI signals), all inputs need to be exactly the same resolution as it's not a scan converter (so you need to use

HDMI scalers), etc

for the cost, they're the best option currently, but certainly can be improved on a ton

Well I'm going to get food - will be back in an hour

xfxf: what are you using for kiwi?

guess ;)

the AV here is from like 1999 and because it's a weekend the AV team has gone AWOL

and the emergency number i got isn't being answered by everybody

i've had to creatively work around a few issues

but... uni av... unfortunately normal

s/everybody/anybody/

I have food now!

And about enough energy to eat it, send one more email and then sleep for like 12 hours

back in oz yet?

CarlFK: any luck w/ Vagrant? I'm going to start debugging myself - noticed a few issues just then

xfxf: it barfs right away: vagrant up; /usr/lib/ruby/2.1.0/rubygems/version.rb:202:in `initialize': Malformed version number string virtualbox (ArgumentError)

doesn't for me

i'll try and get it working for me and submit you a PR

xfxf: Vagrant file doesn't do much for veyepar. can't you just bring up a vm?

there's all sorts of things I *can* do, i'm just trying to get your way of installing it working

the reason others are writing veyepar clones is "veyepar is hard to install"

i'm trying to help you fix this :)

my assumption was the vagrant file was Working For You (tm)

but i think i've spotted at least two issues which will prevent it from properly working period

testing now

probably be good to move a bunch of your stuff in INSTALL.sh into an ansible thing too

yep, got through tests

i'll send a PR

I think the reason people are writing clones is because veyepar does too much and you get overwhelmed with all the options

2 years in a row DC has stuffed up the image at the end of the video because they forgot about it

i still think there's a way to achieve both newbie friendly but featureful but that requires me writing code which i don't have any time to write currently

this year they had one that wasn't right, got it fixed but forgot to copy the file. so it used the 'bad' one

you might be right, but I don't think it is worth the effort

i think it is

then write the code :p

it'll stop this silly fragmentation of people expending energy solving the same issue

ya i know

no it wont

this is why i'm not complaining too loudly

it will, there's a point where if it's convenient enough to install and 'just work' then it'll overthrow the whole not-invented-here-syndrome thing somewhat

but it'll involve a bit of refactoring of the UI

somewhat.. maybe.

which is not a trivial effort

all i know is after you i'm the next best person on the planet who knows the codebase/product and even i'm hitting bumps

and that's definitely an issue

but i'll just shut up and fix it and submit PR's :P

rm veyepar/dj/scripts/tests.txt - does that abort if the file doesn't exist ?

yes

#!/bin/bash -ex

that shouldn't be in a vagrant file

imho

as per above ideally we'd remove the $script part entirely and use ansible

which i can do, just can't now

I can see that. tests.txt should be deleted if it exists.

it shouldn't if you're provisioning a new VM

config.ssh.forward_x11 = true I know what that does, but why is it needed?

because i might want to do something silly like run a local web browser inside of the instance

lol

you can't easily do things like 'ssh -X' with a vagrant instance as you use 'vagrant ssh' to access it

i've made a couple of other changes now i'm testing with a new provision that sets up a private network + port forwards the django testing server port

I think you have a different use for Vagrant file than I did

if we wanted to get fancy we'd stick in wsgi/nginx and just have it 'work'

I made that for GsoC students to show that INSTALL.sh can work.

ah

i'm trying to use it so i can easily get a instance running on <random machine?

>

if we include all of the encoding tools it means we can use random people's laptops as part of the render farm at conferences

right now using it to provison an instance on my OS X laptop