numberOfConsecutiveZeros > 0 if block has multiple channels all starting with a zero

[chrhaase]

Hi @sudara !

Really digging your audio block test helpers repo. Thanks for that!
I've run into a bug when filling a multichannel block with a sine. Both channels will be starting at zero and this is causing a REQUIRE_THAT (blockIsFilled (block)) check to fail.

I believe it could be fixed like this (simply skipping the first sample):

    template <typename SampleType>
    static inline size_t numberOfConsecutiveZeros (const AudioBlock<SampleType>& block)
    {
        size_t consecutiveZeros = 0;
        for (int c = 0; c < (int) block.getNumChannels(); ++c)
        {
            for (int i = 1; i < (int) block.getNumSamples(); ++i)
            {
                auto lastSample = block.getSample (c, i - 1);
                auto currentSample = block.getSample (c, i);
                if (currentSample == 0 && lastSample == 0)
                    consecutiveZeros++;
            }
        }
        return consecutiveZeros;
    }

[sudara]

Hi @chrhaase,

Thanks for filing an issue! Ah, I forgot that the test helpers depend on this sparklines repo! Yikes!

I definitely wrote that lil helper to allow something like a sine wave work, which is why at the end it only reports zeros if there were > 1 of them in a row.

But I'm betting you are running a multi channel audio block?

Edit: Ah, didn't read clearly enough. I think skipping the first one... might report different numbers, maybe we'd just have to track them on a per-channel basis vs. globally...

Sounds like we need tests for the test helpers 🤣

[chrhaase]

Yeah, but isn't the intention a bit more clear when skipping the first? Then zeros are only counted if they are actually next to another zero. Or maybe I misunderstood? :)

What do you mean by reporting different numbers? Maybe it's actually ambiguous to ask the question of consecutive zeros for a multi-channel block - are zeros only counted if all channels are zero for a given sample or are each channel counted individually. Is this what you're getting at? Maybe I'm overthinking it... :)

For isFilled, though.. I guess the most intuitive to count each channel individually? So that a block is only filled if every channel block is filled. That's at least what I think I'd expect.

[sudara]

What do you mean by reporting different numbers?

I guess in the sparkline context, an empty block would report as something like [0(64)] — so in this context, we'd need to prioritize that count being correct.

For isFilled, though.. I guess the most intuitive to count each channel individually? So that a block is only filled if every channel block is filled. That's at least what I think I'd expect.

Agreed!

I think what makes most sense to me is to break this method apart into the different use cases its serving right now... I'll take a look soon!