Im looking for a way to run this Sony DSEE Algo within my daw or even export through the algo.. The closest thing I could find is this player that uses the algo (https://www.sony.com/electronics/support/articles/MC4PC020001)

What DSEE does is it enhances low bitrate audio. Maybe there is something better or a way to use it that im not aware of?

https://helpguide.sony.net/mig/Z002640611/EN/contents/TP0000221988.html

 

Reminds of the good old LAME vs OGG wars..

At the lowest Bitrates OGG tends to be better than LAME,
at full rates they tend to perform about the same..

This article is 20 years old tho, OGG has improved quite a bit since then..

 

This Sony thing seems to have some dubious claims tho..

You cannot really restore what's not there to begin with.

They throw the 32bit number around just for effect.
As we all know 32bit audio is just 24bit audio with a much higher dynamic range,
it has nothing to do with audio quality or resolution per se..

They should give more specific data/results, this graphs are pretty much fantasy

 

The best results I have heard on here of software for similar purposes, were all done by a user of Thimeo Stereo Tool. https://www.thimeo.com/stereo-tool/

 

What is this comparison? OGG is a container, and LAME is an .mp3 encoder.
Let's compare .flac to .zip while we're at it.

 

https://xiph.org/vorbis/

"Ogg Vorbis is a fully open, non-proprietary, patent-and-royalty-free, general-purpose compressed audio format for mid to high quality (8kHz-48.0kHz, 16+ bit, polyphonic) audio and music at fixed and variable bitrates from 16 to 128 kbps/channel. This places Vorbis in the same competitive class as audio representations such as MPEG-4 (AAC), and similar to, but higher performance than MPEG-1/2 audio layer 3, MPEG-4 audio (TwinVQ), WMA and PAC."

Here's the FAQ


*It has been improved a lot in the last decade tho,
now it supports up to 500KB/sec and seems to perform better than LAME in certain scenarios..

There's some fundamental differences in the way things are processed, and that makes it interesting too.

 

Last edited: Jul 24, 2024

Yep, @ᑕ⊕ֆᗰIᑢ is right. While the.ogg file format is also a container it's been used almost always for ogg vorbis codec. For a couple of years was used to store opus codec too. Before the .opus file extension was created.

 

There's quite a few algorithms for improving lossy-compressed audio playback. Haven't tested them much.

 

I remember reading here in Audiosex about an AI that restored content on lossy audio..
it gave pretty interesting results.

Can't remember the name atm, there's just too much AI nowadays lol
But yeah somebody will probably come with it.

 

If you find it please let me know!

 

Yes! I got this. It’s super hard to wrap my head around on how to use ir. But I’m getting there!

 

Meanwhile I found some interesting links..

PGGB Upsampler:
Article- https://audiobacon.net/2024/02/07/review-pggb-your-personal-music-upscaler/?amp=1
Discussion - https://www.audiosciencereview.com/...pggb-upsampler-deltawave-null-analysis.42315/

AudioSR Scaler:
https://audioldm.github.io/audiosr/
Usable Model - https://replicate.com/nateraw/audio-super-resolution?input=form&output=preview
Article- https://www.marktechpost.com/2023/0...upsampling-audio-to-incredible-48khz-quality/

Similar project:
https://kuleshov.github.io/audio-super-res/


Sony DSEE:
Basic info - https://www.sony.co.uk/electronics/support/articles/00230269

Samsung UHQ:
https://www.samsung.com/uk/support/...o-codes-and-ultra-high-quality-audio-options/


As far as tools go, there's things like Unchirp,
meant for cleaning lo-res compression artifacts...
https://www.zynaptiq.com/unchirp/


And ofc there's all those On-Line AI upscalers..
https://music.ai/modules/enhancement/
https://stockmusicgpt.com/features/upscale-music
https://hance.ai/

 

Not actually a plugin, but iZotope RX has an algo in the standalone version to restore high frequency content lost in lossy audio. It's awfully slow though.

 

This post is top notch! thank you!

 

oh yeah. ive used this replicate space a few times. It actually makes the sound worse IMO im assuming the github would do similar. Love that suggestion though!

DSEE combined with iZotope RX spectral recovery is sounding pretty good. Im still wrapping my head around stereo tool

 

Forget immediately everything where the term "upsampling" is involved. Upsampling does nothing. Upsampling can not restore anything where there is no information.

 

yeah maybe the term upsampling isnt correct. I saw a video a while ago about zynaptiq - unchirp talking about how a lot of the low end plugins that enhance bass use harmonics to generate harmonics in lower frequencies. They applied that same theory for upper harmonics.