Pyra audio - needs to be fantastic

[bzar]

Sounds like a way of working around compression algorithm flaws.

Since the amount of original information is the same, the only way to make the final output closer to the original (in an ideal, theoretical world) is to throw away less or different information during the processing stages.

I'm guessing the "different" option. I'd be curious to know exactly why this method of compression has become so popular in the anime crowd, but I'm guessing the material they're concerned with benefits from this more than most. Some examples list banding in subtle clean gradients, which for one wouldn't be as prevalent in, say, real-life footage.

5-20% lower bitrate.

Lower than what?

Lower than the (IIRC measurably) same-quality video stored in 8-bit colors. As far as I could gather from the presentation behind the link.

 

[comradekingu]

Where do you even find source material in 10 bit color depth? All DVDs and BDs are 8 bit, and I don't think there are many cameras that record at that depth. And where do you find a screen that displays at that depth?

Or are those two extra bits just always two zeros?

All the anime stuff is 10bit, has been for quite a while.
http://en.community.dell.com/support-forums/peripherals/f/3529/t/19488832.aspxThis is the monitor i got.

I think this works on a best effort basis to saturate the colour space / profile of the display.  http://www.tftcentral.co.uk/articles/content/ips_technologies.htm

 

[bzar]

^ Your link has some extra characters in the end and leads to a 404 page.

 

[Binky]

Link fixed now?

 

[bzar]

Seems so

 

[klapse]

For hacking fun and music recording, a line-in would be nice. EH.

 

[Exophase]

In our ideal world, you would expect the (x-bit raw data)compressed output) size ratio to go down by 20% since you've artificially inflated the raw data from 8-bits to 10-bits without adding any information.

This isn't about taking an 8-bit data source, adding two additional bits, then getting better compression. The benefit doesn't come from padding the source material but from having more precision when the data is in different spatial and temporal representations. This results in less quantization noise and therefore smaller errors in the compression. It just so happens that the improvement in quality generally outweighs the cost of those extra bits. Better quality at the same bit-rate means lower bit-rate at the same quality, or reduced file size.

 

[_wb_]

Yes, all lossy compression adds artifacts, and of course you can reduce the artifacts by working in a bigger space and that can make sense. Transforming RGB to YUV and back already introduces error even if the YUV is encoded losslessly. But I wouldn't call this 10 bit color depth unless it really is, it's usually just 8 bit color depth with a fancy codec that happens to have poor hardware acceleration because it is working in a somewhat nonstandard color space most of the time.

 

[Exophase]

It really does work with 10-bit color components and produce a 10-bit result though. If you have a 10-bit monitor you can feed the output directly. I don't really know exactly but I think these transformations work best when you don't try to change the precision between steps.

 

[levi]

If we can piss off the weaboos by not supporting 10-bit colour, then that's all to the good


Anyway, not sure what this stuff is doing in an audio thread.

 

[WizardStan]

I originally assumed they were talking about 10bit audio, and then something went visual. :wacko:  

 

[_wb_]

It really does work with 10-bit color components and produce a 10-bit result though. If you have a 10-bit monitor you can feed the output directly. I don't really know exactly but I think these transformations work best when you don't try to change the precision between steps.

OK, but if it's encoded starting from an 8-bit source, I still don't see the point. Maybe some kind of benign compression artifacts can result in slightly smoother gradients in some cases, but I doubt you would see the difference unless you're post-processing stuff or use really weird gamma settings.

 

[God Ginrai]

Anyway, not sure what this stuff is doing in an audio thread.

 _wb_ posted this:

The Pyra will probably be able to play any full HD movie just fine, even if the display is only 720p.

So I pointed out that it was highly doubtful that the Pyra would be able to play pretty much any 10-bit encode easily.

-God Ginrai

 

[klapse]

Im pretty doubtful normal ears can hear more than 16bit 44.1khz lossless.

The first sampling keyboard I used played 8-bit samples Ensoniq Mirage. Sounded like fuzz.