HELIX ADAC Converters?

[bartnettle1]

Would like to know what are the current converter chips in the the Helix and also any data concerning the circuitry

[duncann]

Don't know about the requested data, but there's this taken from TGP from a post by Digital_Igloo:
 
Here are Helix's, if you want to compare:
 
A/D 8 Channels: https://www.cirrus.com/en/pubs/proDatasheet/CS5368_F5.pdf
D/A 8 Channels: https://www.cirrus.com/en/pubs/proDatasheet/CS4385_F2.pdf

But again, we do some patented trickery to get 123dB of dynamic range on the Guitar Input, even though the converters are rated at 114dB.

[bartnettle1]

Thanks dunncan 

 

That is strange the Axe FX8 is using the same converters as the HD500x (CS 4272)

 

The different converters are undoubtedly for the ins and outs tally but rated the same in dynamic range. 114db

 

The question is; whose patented trickery gets an extra 9 db out of them?

[Digital_Igloo]

The question is; whose patented trickery gets and extra 9 db out of them?

 

Line 6's. That line was from the quote. We're able to get 123dB of dynamic range from 114dB converters.

[wenpaul]

Line 6's. That line was from the quote. We're able to get 123dB of dynamic range from 114dB converters.

Is the use of techniques something like "dithering"?To get more dynamic range?

[mdmayfield]

Is the use of techniques something like "dithering"?To get more dynamic range?

 

Whatever it is, I bet it delves many layers into the math and science and is probably not something "ELI5" or even "ELI15"-able. I know a lot of audio stuff in a reasonable amount of technical detail, but my knowledge only goes a few layers beneath the surface; not into much of the actual math and physics.

 

I wouldn't be surprised, though, if dither enters into it somewhere - for example, when you crank the gain WAY up, other processors tend to exhibit "spitty" quantization distortion as the notes die out. The Helix does not, and instead gracefully decays into a constant noise floor. That's similar to dithered vs. undithered audio, but of course it's only conjecture.

 

They might be using part of U.S. Pat. #8,098,830, which looks like it was originally developed for the Line6 wireless systems. It describes using a stereo ADC to digitize a mono analog signal, by sending the mono signal at a different level to each side of the stereo ADC, then summing them in the digital domain. On the DAC, both sides are output, summed within the analog domain, then turned down to compensate for the doubling. Pretty clever! That would reduce noise and therefore increase dynamic range. I could be wrong about the reference of course.

[SWoRd]

Is the use of techniques something like "dithering"?To get more dynamic range?

You can also oversample and then downsample digitally to get more bits. The chip supports up to 216KHz sampling, so if you're running 48KHz internally you could sample the ADC's at 192KHz and downsample to 48KHz in software giving 12dB extra in theory, but often a higher samplerate comes with higher noise (depending on the design of the chip that is used) so it might be possible that you only get 9dB in total (pure guesswork from my side).

 

The comment about the patent pretty much rules out this solution though, because digital down sampling is not a new thing and not something you could patent.

[bartnettle1]

Great explanations and sounds on the right track.

May just be how they are doing it though I'd say the processing might be 32bit oversampled at 192khz and the noise filter niquist would be way above anything audible and how most converters work, but oversampled and then down samples makes sense with aliasing above the audible range.

I suspect what gets regarded as fizz could very well be an artifact of the filter curve.It is not uncommon for

 

On the Helix I think I have seen a reference of DSD but I cannot find any reference of it now in the manual and this might just be my mis perception.

Nevertheless, a possibility, is the converters are DSD which is a much more linear natural sounding quality then Pulse Code Modulation. 

It is done by sampling 1 bit lengths at very high frequencies like 2.28224MHz or 5.6448MHz Whereas PCM works at sampling blocks of bits each from zero db to the full dynamic range of the the medium; DSD does it compared to the last bit much faster but results in a linear response much likened to how tape worked where PCM is a stepped non linear way of sampling.

PCM is editable. DSD cannot but can be cut in a gap of silence. Thats it. Like tape. Korg and Tascam have produced DSD recorders. Not that Yamaha have not adopted their own.

It may or not be how the Helix is working. If it was in there it would be talked up as selling point advantage. 

So far we are only supposing about the internal processing of the modeling.

Of course the USB and s/pdif interfaces are PCM

[TTRESHH]

Just wondering, because in my studio Helix is the main audio interface and there's lots of equipment im feeding from it...

Is there analog difference (circuit or DAC) and subsequent difference in output quality between Main 1/4" Outs and Sends 1-2-3-4?
Is the XLR out delivering the same quality like 1/4" (not counting difference between balanced and non-balanced connections)?