**Simeon Brown removed Med Council wokesters**

https://www.stuff.co.nz/politics/360993290/simeon-brown-removes-medical-council-leaders-over-idealogical-agenda

**Wanganui hospital**

**T-A-S video expose**

https://x.com/The_Anglo_Saxon

**DEI industry - equality v equity**

https://www.telegraph.co.uk/news/2026/06/13/equity-equality-act-public-services-nhs-police/

**Simeon Brown on party vote**

https://www.rnz.co.nz/news/politics/610017/national-s-simeon-brown-discourages-members-from-strategic-voting

**Winston Peters’ response**

**Wayne Brown equality=racism**

https://www.rnz.co.nz/news/te-ao-maori/601621/new-zealand-first-s-campaign-to-scrap-independent-maori-statutory-board-just-dumb-racist-stuff-mayor

**Mitchell’s private India police**

https://awaaz.co.nz/new-zealand/concerning-police-step-in-after-tamaki-calls-for-purging-indians-over-modi-visit.html

**MoE officials encouraging thought stopping**

Get more from Simon Anderson in the Substack app

Available for iOS and Android

Get the app

The sweet spot for streaming audio is 160 kbps (Kilobits per second). Some platforms will accept higher bitrates but 160 kbps is currently, the common denominator with reasonable expectations of reliability.

Fortunately for creators the audio quality you broadcast can be different to the audio quality you record. Using the channels feature in tools such as OBS allows you to segment your audio and treat each channel differently.

In this article I’ll show you how I do it. Have a look at this table depicting my audio channel configuration and the sources that are included within them:

You’ll see I have six discreet channels. I’ve assigned the first two as broadcast channels at 160kbps and the remaining four as recording channels at 320 kbps. This enables me to broadcast livestreams at the thresholds set by the social media platforms while also facilitating edits of pristine audio which has been nicely segmented after the broadcast has concluded.

Here’s how this is achieved in OBS. First, set the quality of the six channels:

Next, assign audio sources to the appropriate channels:

Then adjust your streaming setting to only send out the main mix:

..and your recording to all of the channels you wish to capture for subsequent editing:

The end result is a recording of the livestream which includes six distinct channels along with the video channel.

My use-case is a little different because I broadcast in two video formats (landscape and portrait) each with its own main mix so I can play different music between them. Here’s how it all looks when I bring the recordings into my editing software, Davinci Resolve:

Recording the low quality Main Mix 1 in both portrait and landscape simply allows me to align the two video clips. Once the clips are aligned I dispose of those audio tracks because they’re lower quality. From there I can selectively mute the pristine audio channels to create a new, high definition main mix. If I forgot to turn the music off during the livestream before I started speaking for example, I can correct it here in the edit. Or if two guests spoke over one another I can mute one of them.

320 kbps is very high quality for a lossy format. Some would argue 256 kbps is ample as few listeners can distinguish a difference, and doing so would save on computation cycles and storage space. I argue why not? Most modern machines can keep up with audio recording at this level of detail comfortably and downscaling is a lot easier to achieve than upscaling.

In summary configuring channels in your broadcasting software facilitates quality improvements and greater flexibility in edit. While a lot of the recording ends up on the editing room floor it’s worth it for pristine final mixes.

-SRA. Auckland, 20/vi 2026.

A while back a friend persuaded me to shift from Kdenline to Davinci Resolve for video editing. I’m very pleased he did.

Resolve is the tool of choice for Hollywood films and has the steep learning curve one might expect from professional software. It has a vast array of features I find incredibly useful but the number one attraction is the time it saves me. Simply by switching from Kdenlive I’ve saved myself many hours of boring, repetitive work.

Use Case

Here’s the use case: each Sunday I host my weekly podcast SimonTV LIVE for two hours. Several thousand people tune in to watch the live show but tens of thousands more watch the highlight clips I release daily until the next live episode. I need to select seven clips from the recording of the two hour-long podcast, add a credit sequence, subtitles, fades and a bit of polish. Nothing too complicated, simple edits that are time consuming and repetitive.

The workflow in Kdenlive was laborious, I’d scrub through the recording -essentially watching a repeat of almost the entire show- then process each selected clip individually. This process could consume as much as five or even six hours of my time.

Thanks to Davinci Resolve I can now do it in one hour if I’m in a hurry and comfortably in 90 minutes. How, is almost entirely depicted in this screenshot:

Step 1 - Transcription

Resolve has an AI feature which transcribes the entire two hour recording into text. This is not the same as the AI auto-captioning which is another useful feature I use. The transcription takes about ten minutes on my little machine to complete which is worth it for the wonderful world of Text-Based Editing it opens.

Step 2 - Text-Based Editing

This is the greatest time saver. If you’ve ever used a video editor you’ll be familiar with the standard process of scrubbing backwards and forwards to perform your edits along a timeline. With Text-Based Editing I don’t need to do this, instead of watching and listening I can read the transcript -and I read a heck of a lot quicker than I can watch.

At the bottom of the floating transcription box in the picture is a Create Subclip button. All I have to do is highlight a section of the text I think might be worthy of becoming a clip, hit that button and Resolve creates the clip for me from the corresponding segment of the recording. This is a rough cut. I read through the transcript, highlight interesting bits (typically up to a dozen) then move onto the next step having pruned most of the recording.

Step 3 - Clip Selection

I now have 10 to 12 highlight clips of 30 seconds to three minutes in length to review, from which to select the final seven. At this stage a two hour podcast has been reduced to about 25 minutes of useful footage. I watch these to choose the seven finalists.

Step 4 - Fine Cut

The seven clips are still rough cuts needing to be trimmed and edited, and otherwise modified. At this stage I might snip some uhms and ahs and a sentence here or there where I’ve digressed. This is the stage to add polish but I don’t tend to do anything too sophisticated, not even zooms for emphasis (though I probably should.) If I do need to perform anything more complicated -such as audio clean up- this is the step at which to do it.

Step 5 - Final Cut

This is the next big time saver. I’ve created a template containing the generic edits applied to all my clips: fade in, fade out, add credit sequence, position word-highlighted subtitles. This exists as a separate timeline tab in Resolve to which I can switch. I select each of the seven fine cuts I created in the previous step then right-click the dummy clip on the timeline to perform a Conform Lock with Media Pool Clip operation. This replaces the placeholder clip with the fine cut and crucially, applies the edits from the template to the fine cut clip.

Then I just need to position the sequences correctly and I’m pretty much done.

Step 6 - Normalisation and Subtitling

I haven’t worked out a method for automating the AI Subtitle and Audio Normalisation operations. I typically normalise audio at -3 to -1 dB which is a simple two-click operation.

The I run Resolve’s AI to create the subtitles which I review -the AI is very very good but it still makes the odd mistake which needs to be manually corrected. Once again, this AI subtitling is a different feature to the AI transcribing. The template already has the word highlighting and positioning of the text defined, as soon as the AI has generated the subtitles and I’ve manually corrected any mistakes it’s applied.

Step 7 - Delivery

The clip is now done. I head over to Resolve’s Delivery tab to check the preset output parameters are correct and add it to the render queue.

Step 8 - Render

Whichever video editor you’re using rendering always takes a while, fortunately it’s a process which seldom needs to be supervised. I repeat steps 4 through 7 for each of the seven clips and when I’m done, render them all as a batch job which in Resolve is a one-click operation.

SimonTV LIVE concludes at 9.30pm on Sundays. I click that Render button sometime before 11pm then go to bed. Davinci Resolve probably renders all seven clips in half an hour but I don’t pick them up until the morning, which is when I queue them for publication on the various social media sites. Job done.

Conclusion

I am not an expert, Davinci Resolve improves its capabilities rapidly and there are numerous third party AI solutions which integrate with it. Doubtless there are methods to improve my workflow and I’d particularly like to discover ways to automate the application of subtitles and audio normalisation. These would merely be icing on the cake though: my time commitment has already reduced by 80% simply by switching across from Kdenlive and utilising the capabilities I’ve discovered for myself.

-SRA. Auckland, 16/vi 2026.

I’ve written a series of articles previously on how I produce live and recorded content. In this article I’m going to delve into audio, a key aspect I haven’t covered previously.

Good audio quality is more important than video in podcasting. Often the audience isn’t even watching, they’re listening while they’re doing something else. Poor video quality is forgivable but audiences tend to switch off when the audio is sub-optimal.

For conversations great audio is relatively easy to achieve - use quality microphones and an audio interface, and learn how to configure and apply filters such as Gain, Compressor, Expander, Limiter, Noise Gate and Noise Reduction. I’m not going to cover these subjects, instead I’m going to describe how I include music in my broadcasts as a demonstration of what not to do.

Let’s begin with some diagrams so I can walk you through how I got here. First is the Optimal State - this is how my audio configuration would look if I was designing it from scratch: simple and reliable.

Below is the Current State - how my audio set-up actually is. I arrived at this destination through a philosophical commitment to recycling components instead of replacement, amalgamating the modular solutions I had in place and augmenting their capabilities where necessary.

Evolution, in other words, rather than revolution.

These two diagrams are probably difficult to interpret because I haven’t labelled the components. Conceptually you’re seeing three discrete modules amalgamated so I can broadcast analogue and digital music, and mix it together.

I’ll show you below how I got here.

Starting with the Analogue Module, a standard stereo configuration everyone has always had. An amplifier and speakers connected to an audio source - in my case a couple of turntables and a mixer. With the addition of a microphone and headphones this is the standard configuration for every bedroom DJ (or at least it was until the advent of Digital DJing.) In my case I’ve owned most of this kit for three decades.

When I started podcasting I needed to present good vocal audio so I built myself an entirely separate Podcasting Module connected directly to my workstation.

This separation makes sense because the stereo (Analogue Module) is in the lounge whereas I podcast from my home office, a different room. The module comprises just a microphone and an audio interface for input and speakers and earpieces for monitoring. They’re directly connected to my workstation so that I can broadcast and record content in OBS.

So far so good. I wouldn’t do this any differently.

I’ve been holding out against digital music for a long time and finally caved. This is point at which I made the calamitous decision to integrate a discrete Digital Module. This comprises a digital controller and another workstation to run the DJing software, Mixxx. It has RCA outs for the main audio and a headphone jack. It’s also located in the lounge, where it can readily connect to the existing stereo equipment a.k.a the Analogue Module.

What I should have done is replace my existing analogue mixer with a digital mixer which has analogue inputs, the one I purchased does not. If I had I’d now be enjoying the simple and reliable elegance of the first diagram. In my defence the controller I have is a lot less expensive, few controllers with analogue inputs are compatible with Mixxx and it would have meant my existing analogue mixer was redundant.

This is the point at which complications began. Issues, for which I needed to find solutions.

The first was amplification and signal splitting. The output from the digital controller needs to be close to the output signal strength of the turntables so that they can be balanced when I’m mixing. The digital mixer also needs to be connected to both channels of the analogue mixer so I can mix either turntable with a digital track. These two issues necessitated the introduction of a small amplifier and a Y RCA splitter.

Next was audio output. My analogue mixer doesn’t have a USB interface so I tried converting the analogue signal from the amplifier into digital with a USB converter but the result was sub-optimal audio quality. The solution was a Behringer mixer with good processing a friend donated. It’s the grey box at the bottom right:

The next issue was audio monitoring, which is controlled by a little switcher (bottom left in the image.) Because I mixer tracks digitally on the top mixer and analogue with analogue/digital on the mixer at the bottom, and audio output is controlled by the Behringer, I need to monitor audio coming out of three separate components and switch my headphones between them.

The final aspect is the distribution, which occurs digitally between the two workstations. The music workstation processes all the audio generated by the components through Thimeo Stereotool. The music workstation then sends this over Ethernet using Icecast, which the broadcast workstation consumes and sends out into the world through OBS.

A problem with this I’m yet to resolve is the time it takes to process, which causes desynchronisation with the video when I’m filming. If you watch me perform right now you’ll see me cross fade from one track to another a full eight seconds before you hear it. (To be fair this is an issue which I would also need to solve in the Optimal State.)

[UPDATE 18/vi 2026: the solution for the synchronisation problem was to add a ~7 second delay with the OBS filter Video Delay (Async).]

The end result of all this is a solution which is unnecessarily complicated. I have to switch inputs and outputs when I’m performing, depending upon the sources. Every component in the chain tends to add noise so gain staging between them is crucial, and limits how much equalisation I can safely adjust when I’m live. Above all, a lot can go wrong and a failure point is difficult to identify.

Here’s the diagram again. It works and I successfully preserved existing components. Trust me when I say though, this is not the approach I recommend.

-SRA. Auckland, 13/vi 2026.

Full conversation: