All guys are interest in video decoding, any one look into audio decoding?

[nijiko]

Is there a Hi-Fi res audio decoder helping improve audio quality?
Which player can output audio with high performance?
Any audio renderer can use excluive mode for best audio playing?

*Here is not discussed about such as foorbar2000, only MULTIMEDIA player.*

[huhn]

of cause the is not a discussion.
lossless decoded audio can be outputted as it is so no precessing is needed unlike video output.

[nijiko]

of cause the is not a discussion.
lossless decoded audio can be outputted as it is so no precessing is needed unlike video output.

But DirectSound is not the best way to "transfer" audio to speakers.
And, lossless decoders (filters) are not able to ensure what transfer to speakers is the same as SOURCE.
Such as audio players, foobar2000, WinAMP use many filters to ensure the PURE music or more CLEAR music.
Why there isn't the same things for MPC?

[huhn]

because you only get 100% accurate audio with 100% volume.
with my headphones that would be 116 DB no thank you.

MPC-BE has wasapi exclusive and with everything at 100% that'S all you need to get bit perfect output. you can use bitstreaming too does the same it's totally impossible to touch a none decode audio stream.

there si simply no magic needed.

[nijiko]

because you only get 100% accurate audio with 100% volume.
with my headphones that would be 116 DB no thank you.

MPC-BE has wasapi exclusive and with everything at 100% that'S all you need to get bit perfect output. you can use bitstreaming too does the same it's totally impossible to touch a none decode audio stream.

there si simply no magic needed.

Also,
because the video formats are LOSS formats now, we NEED to improve them.
There are many improve filters for them, such as madVR.
Many audio formats are LOSS formats, too. Such as MP3, AAC, and etc.
Why not to improve them?
Why there is no some soft AMP or DSP? Or other tools look like DSEE?

[huhn]

Also,
because the video formats are LOSS formats now, we NEED to improve them.
There are many improve filters for them, such as madVR.
Many audio formats are LOSS formats, too. Such as MP3, AAC, and etc.
Why not to improve them?
Why there is no some soft AMP or DSP? Or other tools look like DSEE?

because the soundcards are limited to 24 bit usually and lossy audio is bitstream(so bit perfect) or 32 bit float that dithered or rounded to 24 bit. what do you want do more?

[nijiko]

because the soundcards are limited to 24 bit usually and lossy audio is bitstream(so bit perfect) or 32 bit float that dithered or rounded to 24 bit. what do you want do more?

What I mean is upsampling some audio, better interpolation to improve bitrate, and fixing the high and low frequency in lossy audio.

[nevcairiel]

You need video processing because displays can only show RGB in one specific resolution, but video is not in RGB and comes in all sorts of resolutions.
You need chroma upscaling, you need image scaling. On top of that, you may need to compensate for your display, so you need a calibration using 3DLUT or similar approaches.

Audio comes in all sorts of "resolutions" as well (sample rate), but the DACs are usually equipped to handle all the common sample rates just fine, so in the majority of all cases, you don't "need" resampling.
So all you need is bit-perfect audio transport to the DAC, and thats what WASAPI excl. will do for you.

The only thing that would apply to audio here is the calibration part. For a perfect experience you need to compensate for your speakers and your room. There are projects for that, like REW (http://www.roomeqwizard.com/), but I don't think there is a convolution filter for DirectShow thats freely available, unless ffdshow audio can do that.

PS:
"Improving" lossy audio is a bit of a myth. I wouldn't put too much into that.

[nijiko]

You need video processing because displays can only show RGB in one specific resolution, but video is not in RGB and comes in all sorts of resolutions.
You need chroma upscaling, you need image scaling. On top of that, you may need to compensate for your display, so you need a calibration using 3DLUT or similar approaches.

Audio comes in all sorts of "resolutions" as well (sample rate), but the DACs are usually equipped to handle all the common sample rates just fine, so in the majority of all cases, you don't "need" resampling.
So all you need is bit-perfect audio transport to the DAC, and thats what WASAPI excl. will do for you.

The only thing that would apply to audio here is the calibration part. For a perfect experience you need to compensate for your speakers and your room. There are projects for that, like REW (http://www.roomeqwizard.com/), but I don't think there is a convolution filter for DirectShow thats freely available, unless ffdshow audio can do that.

PS:
"Improving" lossy audio is a bit of a myth. I wouldn't put too much into that.

For me, what I need madVR to do, is improving my old films which contains low resolution display finer in a high resolution displayer. (image upscaling)
And I think I need a tool to upscale the lossy AC3 or MP3 to let me hear more clearly.
Some old VCDs contain low bitrate MP2/3 audio, the audio is terrible now. Nothing to improve them?

[huhn]

For me, what I need madVR to do, is improving my old films which contains low resolution display finer in a high resolution displayer. (image upscaling)
And I think I need a tool to upscale the lossy AC3 or MP3 to let me hear more clearly.
Some old VCDs contain low bitrate MP2/3 audio, the audio is terrible now. Nothing to improve them?

you can play around with ffdshow audio or soundcard crystalizer. but bad audio is bad audio.

it's not like madVR can save S/VCD. it simply does the necessary step in high quality.

[edigee]

Is there a Hi-Fi res audio decoder helping improve audio quality?
Which player can output audio with high performance?
Any audio renderer can use excluive mode for best audio playing?

*Here is not discussed about such as foorbar2000, only MULTIMEDIA player.*
No! HI Res parameters aplied to a CDA standard audio material is a total waste. In general speaking ,encoding a wav 16/44,1 audio track to high res doesn't improve anything. All the hi res hype is placebo stuff. Any serious blind test can proove that.
Although hi res audio IS mandatory on studio processing ,in audition doesen't make much difference.
A high quality DAC (and the "transport bit" too) makes the real difference. And the whole audio gear.

[detmek]

Upsampling audio does not improve quality.

madVR does not improve quality, too. It just handles conversion process better then other renders. For example, we can compare EVR and madVR. EVR uses default resizer for installed graphic card and driver. It is usually Bilinear which produces blurry picture. In madVR you can use Lanczos that will not blur picture but will produce ringing. You can use Lanczos with antiringing that will not create blurry picture and you will not see much of ringing. Or you can use Jinc to get relatively sharp picture without aliasing. So, madVR does not improve anything, ie. it does not make perfect video out of crappy source. It just does not create a lot of damage.

With audio things advanced more then in video world. We deal with lossless audio on daily basis. If you want to preserve source quality you can bitstream audio to high quality DAC or decode in software and send bit-perfect PCM to DAC using WASAPI Exclusive, Kernel Streaming (Windows XP) or ASIO. And do not use volume control in player or in OS. Use speaker volume control.

High quality resampling does not increase quality, it just does not produce audible artifacts. If you can hear a difference between original and upsampled audio in player it means that you have a problem in playback chain or that your resampler is bad.

If you want to "enhance" audio you can use equalizer or some DSP/VST filter that basicly acts as custom preset equalizer that amplifies some frequencies.

[YxP]

Human hearing is pretty bad compared to eyesight. When audio sounds so bad you actually hear it, it's probably so bad that it can't be just "upscaled". Even MadVR can't make shitty noisy I love Lucy VHS rips look like HD.

[nijiko]

you can play around with ffdshow audio or soundcard crystalizer. but bad audio is bad audio.

it's not like madVR can save S/VCD. it simply does the necessary step in high quality.

Sadly ffdshow is dead.

[nijiko]

Upsampling audio does not improve quality.

madVR does not improve quality, too. It just handles conversion process better then other renders. For example, we can compare EVR and madVR. EVR uses default resizer for installed graphic card and driver. It is usually Bilinear which produces blurry picture. In madVR you can use Lanczos that will not blur picture but will produce ringing. You can use Lanczos with antiringing that will not create blurry picture and you will not see much of ringing. Or you can use Jinc to get relatively sharp picture without aliasing. So, madVR does not improve anything, ie. it does not make perfect video out of crappy source. It just does not create a lot of damage.

With audio things advanced more then in video world. We deal with lossless audio on daily basis. If you want to preserve source quality you can bitstream audio to high quality DAC or decode in software and send bit-perfect PCM to DAC using WASAPI Exclusive, Kernel Streaming (Windows XP) or ASIO. And do not use volume control in player or in OS. Use speaker volume control.

High quality resampling does not increase quality, it just does not produce audible artifacts. If you can hear a difference between original and upsampled audio in player it means that you have a problem in playback chain or that your resampler is bad.

If you want to "enhance" audio you can use equalizer or some DSP/VST filter that basicly acts as custom preset equalizer that amplifies some frequencies.

MadVR use better interpolation algorithm.

Human hearing is pretty bad compared to eyesight. When audio sounds so bad you actually hear it, it's probably so bad that it can't be just "upscaled". Even MadVR can't make shitty noisy I love Lucy VHS rips look like HD.

Human eyesight is the same.
Between diff. interpolation algorithms, the feeling is MORE SOFT or MORE HARD. Some like soft image, and some like hard image.

[huhn]

MadVR use better interpolation algorithm.



Human eyesight is the same.
Between diff. interpolation algorithms, the feeling is MORE SOFT or MORE HARD. Some like soft image, and some like hard image.

and the best interpolation algorithms is no algorithms so audio is fine as it is.

ffdshow still works fine doesn't matter if no ones is updates it anymore

[nijiko]

and the best interpolation algorithms is no algorithms so audio is fine as it is.

ffdshow still works fine doesn't matter if no ones is updates it anymore

You mean raw input? Or the codecs no update anymore is no problem?

[huhn]

You mean raw input? Or the codecs no update anymore is no problem?

what do you need more than ffdshow audio raw?

[nijiko]

what do you need more than ffdshow audio raw?

Dolby, IDCT, crystallization, and ETC?

[leeperry]

uLilith runs a full 64fp pipeline, from lossy audio decoding to VST plugins processing to bit-perfect output(WASAPI/ASIO). Of course it sounds better to have it rounding in the end to 24int instead of relying on Windows to do it.

It also supports automatic VST latency allowing you to use Linear phase EQ VST plugins glitch-free, no other audio player does that AFAIK and yeah Linear Phase sounds a lot better.

[nijiko]

uLilith runs a full 64fp pipeline, from lossy audio decoding to VST plugins processing to bit-perfect output(WASAPI/ASIO). Of course it sounds better to have it rounding in the end to 24int instead of relying on Windows to do it.

It also supports automatic VST latency allowing you to use Linear phase EQ VST plugins glitch-free, no other audio player does that AFAIK and yeah Linear Phase sounds a lot better.

Sound Player Lilith?