can ffdshow audio do this dream resample to 192000hz ?

[pandy]

Yes, i agree - famous "loudness war" kill high quality sound reproduction - all technological advances was destroyed with mastering targeted to portable mp3 players - mass market doesn't care about quality...

[vola]

What with older Creative cards with good driver like KX Drivers?

http://kxproject.lugosoft.com/

IMO Creative make quite good audio cards.

NO older Creative cards

Creative DON'T MAKE audio cards
NOT FOR MOVIES OR Music
Do not know if I working through SPIDF
Why do I need a better card???
No CARD processor information through SPIDF

Do not see why not trying to build Agurittim for 192000HZ
Through SPIDF> PASS THROUGH

Audio card did not analyze the sound Through SPIDF
MY Amplifier gets up 48 BIT 192000HZ

All what I read here is not practical ONLY theories
If the software will BE Supports in 192000
Hardware build By manufacturers .

Just do it
Just do it
Just do it

[Groucho2004]

Creative DON'T MAKE audio cards
NOT FOR MOVIES OR Music
Do not know if I working through SPIDF
Why do I need a better card???
No CARD processor information through SPIDF

Do not see why not trying to build Agurittim for 192000HZ
Through SPIDF> PASS THROUGH

Audio card did not analyze the sound Through SPIDF
MY Amplifier gets up 48 BIT 192000HZ

All what I read here is not practical ONLY theories
If the software will BE Supports in 192000
Hardware build By manufacturers .

Just do it
Just do it
Just do it

Did you run this through some sort of word randomizer before posting?

[vola]

English not MY native language

[pandy]

English not MY native language

Mine to but im trying to improve them and avoiding to make mess on public forums...

[vola]

Mine to but im trying to improve them and avoiding to make mess on public forums...

I do not understand what you want?

[Groucho2004]

Just do it
Just do it
Just do it

This is still about your desire to upsample to 192Kbps in ffdshow, right?

As far as I know, there is no limit in libsamplerate which is used in ffdshow, it just applies a multiplier. I have not looked at the ffdshow source code but I imagine that this would be easy to implement.

Still, lamenting about it in this thread will hardly attract the attention of the ffdshow developers.

[pandy]

NO

Creative DON'T MAKE audio cards
NOT FOR MOVIES OR Music


Creative made a lot of good sound cards, with good drivers they are much better than many similar competitors cards.
But Creative made few, not very good marketing moves and it have bad opinion - but technically they products are above average in terms quality and functionality.



Do not know if I working through SPIDF
Why do I need a better card???
No CARD processor information through SPIDF



So not use card processing - with KX drivers it can be disabled.


Do not see why not trying to build Agurittim for 192000HZ
Through SPIDF> PASS THROUGH

Audio card did not analyze the sound Through SPIDF
MY Amplifier gets up 48 BIT 192000HZ

All what I read here is not practical ONLY theories
If the software will BE Supports in 192000
Hardware build By manufacturers .

Just do it
Just do it
Just do it

At first - S/PDIF may not support 192/24 - it was made for 44.1/16 and work with 48/16 but not many devices support S/PDIF 96/24 and i don't know any device that will support 192/24 (over 9 Mbits), optical S/PDIF (TOS Link) is even more limited and most of optical connection will refuse to work with 96/24 (typical TOS Link transmiter/receiver is approx 4Mbps) - also this cheap plastic connection introduce more problems with quality of the audio (jitter).

For 192/24 probably You must go for HDMI connection - so instead searching for sound card - search for graphics card that have HDMI with embedded Audio on it.

[pandy]

I do not understand what you want?

English is also NOT my native language but im trying to improve my English and im avoiding to make mess on public forums...

Behave instead being like this child http://www.youtube.com/watch?v=HUs_4Aue-fw

[TinTime]

At first - S/PDIF may not support 192/24 - it was made for 44.1/16 and work with 48/16 but not many devices support S/PDIF 96/24 and i don't know any device that will support 192/24 (over 9 Mbits), optical S/PDIF (TOS Link) is even more limited and most of optical connection will refuse to work with 96/24 (typical TOS Link transmiter/receiver is approx 4Mbps) - also this cheap plastic connection introduce more problems with quality of the audio (jitter).

For 192/24 probably You must go for HDMI connection - so instead searching for sound card - search for graphics card that have HDMI with embedded Audio on it.

This Asus (http://uk.asus.com/product.aspx?P_ID=QsEKBPr6ko9pFF2D&templete=2) will do 192/24 over S/PDIF. I'm sure there are other devices and cards as well.

[pandy]

This Asus (http://uk.asus.com/product.aspx?P_ID=QsEKBPr6ko9pFF2D&templete=2) will do 192/24 over S/PDIF. I'm sure there are other devices and cards as well.


So im not so sure - not many S/PDIF receivers are capable to work with 192ksps - issue is not in source but in receiver.
You can transmit data but You must have device that is capable to receive them correctly (conversion from digital to analog is not important at this moment).

[vola]

With all due respect

This is still about your desire to upsample to 192Kbps in ffdshow, right?

YES

As far as I know, there is no limit in libsamplerate which is used in ffdshow,

Please see the picture is limited to 96000HZ

http://img403.imageshack.us/img403/2821/21255289.png

IT'S LIMIT

not many devices support S/PDIF 96/24 and i don't know any device that will support 192/24

All SPIDF support today on 24/96 by manufacturers
Today ASUS + GIGABYTE now support 24 / 192 according to their answers

http://img714.imageshack.us/img714/2664/70734797.png

IF NOT SPIDF
THe last two years all ATI video cards
4000HD +5000 HD +6000HD support through
HDMI in - 24/192


So there is no hardware limit:)

AND If the software will BE Supports 192000
Hardware build By manufacturers going Supports Also 18MBPS
If necessary.

[TinTime]

So im not so sure - not many S/PDIF receivers are capable to work with 192ksps - issue is not in source but in receiver.
You can transmit data but You must have device that is capable to receive them correctly (conversion from digital to analog is not important at this moment).

My fairly elderly Yamaha amp will play it, although all it disables all DSP. I don't know if it truncates to 20bits or not.

[Ghitulescu]

Do not see why not trying to build Agurittim for 192000HZ
Through SPIDF> PASS THROUGH

If you want to say that you would prefer to use your embedded soundcard to send 192kHz/24b over S/P-DIF, well there's a problem as S/P-DIF cannot carry this information, well it can carry 24b but chances are that the receiving end discards the extra 8 bits. 48kHz is the max sampling frequency. More info in IEC 60958.

Should one transport AC-3 signals, well, they may go a bit higher but all the advantage an algorithm in 192kHz would have is lost due to compression. More in IEC 61937.

[Groucho2004]

Please see the picture is limited to 96000HZ

http://img403.imageshack.us/img403/2821/21255289.png

IT'S LIMIT


You've taken my statement out of context and you obviously don't have a clue what I'm talking about.

I give up.

[pandy]

But Yamaha is very good manufacturer - they made good audio from ear and scope point of view - after fast search seems that Only ONE company make S/PDIF receivers capable to 192ksps - Cirrus (former Crystal) - IC's from TI and AD probably not supporting 192ksps on S/PDIF

[vola]

if you want to say that you would prefer to use your embedded soundcard to send 192khz/24b ober s/p-dif, well there's a problem as s/p-dif cannot carry this information
so hdmi 1.3a or 1.4 in hd 6000 ati

[pandy]

With all due respect



YES



Please see the picture is limited to 96000HZ

http://img403.imageshack.us/img403/2821/21255289.png

IT'S LIMIT


NO, put 192000 manually and FFDSHOW will mark output as 192000 but there is something wrong and when i tried to play ac3 file it was played twice slower (maybe issue with drivers on my notebook)


All SPIDF support today on 24/96 by manufacturers
Today ASUS + GIGABYTE now support 24 / 192 according to their answers

http://img714.imageshack.us/img714/2664/70734797.png

IF NOT SPIDF
THe last two years all ATI video cards
4000HD +5000 HD +6000HD support through
HDMI in - 24/192


So there is no hardware limit:)

AND If the software will BE Supports 192000
Hardware build By manufacturers going Supports Also 18MBPS
If necessary.

LOL - i dont believe that someone use Realtek chips as a good sound source example...

[vola]

NO, put 192000 manually and FFDSHOW will mark output as 192000 but there is something wrong and when i tried to play ac3 file it was played twice slower (maybe issue with drivers on my notebook)

my notebook
A full model sound card?
What format do you want to play it? 192 000?
Which player?
What speakers?
What Amplifier ?

i dont believe that someone use Realtek chips as a good sound source example

I HAVE Asus Xonar D2X 7.1
AND PRODIGY LT 7.1

AND I LOVE SPIDF HD
ON Motherboard

Sound Cards sound AS Polished (I DON'T LIKE AC97 ON PCI )

[pandy]

A full model sound card?
What format do you want to play it? 192 000?
Which player?
What speakers?
What Amplifier ?


hp6930p
ad1984a?
directx*
B&W802**
TACT T2**


I HAVE Asus Xonar D2X 7.1
AND PRODIGY LT 7.1

AND I LOVE SPIDF HD
ON Motherboard

Sound Cards sound AS Polished (I DON'T LIKE AC97 ON PCI )

You know that for S/PIDF such audio card act like UART...


*PC and windows software players are not good environment for anything in real time.

**joke - I use headphones - Sennheiser HD215 with my notebook

[vola]

hp6930p
ad1984a?
directx*
B&W802**
TACT T2**

you have
Beautiful parts
I can not know their ability to play your file
Because I do not know the FILE AND YOUR SPIDF
Day-Old in production
Therefore possibility the playr can't play 192 000 Through SPIDF

But the point if you do not play SPIDF So HDMI will play

The question who will put 192 000 algorithm IN FFDSHOW
" resample " ???

Who Can I Turn???

It could be a hit or flop
But then we know

[Ghitulescu]

Please use google translate - do us a favour, pliz pliz pliz

[ramicio]

When did I say that I don't downsample after I mix the audio? In my head I think of upsampling say from 48khz to 192khz almost like what a DAC does. So if there so much error and loss with upsampling, then the same will happen within a DAC when making the signal analog.

And what is so wrong with Sound Blaster cards? Yea their drivers suck anymore, but their hardware is amazing. I tried onboard sound ONCE and will never go back to any C-media or any generic garbage again. I have had zero problems with my PCI-express X-fi card.

People seem to think S/PDIF is an actual interface, which it is not at all. Yes S/PDIF has standards for hardware, but it is not hardware itself, just a way to code the audio. TOSLINK now supports really high bandwidth (125 mbps), so we only all need to wait for sound card makers and receiver makers to allow raw PCM to be transferred optically. That will never happen because of the music and movie industries. You could use TOSLINK for a network if you wanted to. Coax, TOSLINK, and HDMI are only interconnects, they can transfer any data you want to.

[TinTime]

TOSLINK now supports really high bandwidth (125 mbps), so we only all need to wait for sound card makers and receiver makers to allow raw PCM to be transferred optically.

And coax doesn't support high bitrates?

[ramicio]

I never said it physically can't, but TOSLINK already has it implemented, we just need to wait for audio hardware makers to take advantage of it. HDMI is still the best choice but it is too new and so restricted by big corporations.

[nurbs]

When did I say that I don't downsample after I mix the audio? In my head I think of upsampling say from 48khz to 192khz almost like what a DAC does. So if there so much error and loss with upsampling, then the same will happen within a DAC when making the signal analog.
So what you do is upsample, downmix, downsample. There won't be any problems with that, at least in the 48 kHz to 192 kHz case, could be a little more tricky with a 44.1 kHz source, but probably no real trouble anyway. Still two of those three steps are useless. There is no reason to upsample to begin with and the downsampling is only there because you want the original sampling rate back.
There are reasons why DACs upsample. It allows the use of a lower resolution DAC, like for instance the 1bit DACs that are standard (I think) in portable players, to convert higher resolution input for example. It also helps with quantization noise. The point is they do it because it helps with the actual digital to analog conversion. That has nothing to do with mixing a couple of channels together.
So if you want to do a good quality downmix leave the sampling rate alone and use a higher bitdepth.

[TinTime]

I never said it physically can't, but TOSLINK already has it implemented, we just need to wait for audio hardware makers to take advantage of it. HDMI is still the best choice but it is too new and so restricted by big corporations.

High bitrates over coax are also already implemented.

[ramicio]

Yeah that's exactly what I do for movie tracks, which are 5.1 24/48 lossless, and I just downsample to 24/96 then just because I don't care about space. I'm done beating my head into a wall here. Everyone tries to clutch onto something they like and invest so much into rather than accepting change when there is something much better available.

[ramicio]

High bitrates over coax are also already implemented.

In what aspect? Industry? Definitely not for home use. I know of all the crazy coax BNC crap out there, and industrial stuff is not much better than home stuff, just different so they can charge more for it. Anyone can have any technology that isn't classified by a military, if they have the money for it.

[nurbs]

OK so your workflow is this:
5.1 24/48 -> 5.1 24/192 -> 2.0 24/192 -> 2.0 24/96

A better workflow would be:
5.1 24/48 -> 2.0 24/48

No difference in quality, but the latter would be half the size.

If the audiophile magic in your workflow makes you happier that's fine, but it doesn't make the result any better, it just makes it bigger.

[ramicio]

As I have said before, no.

5.1 24/48 -> wavs

I mix each channel independently where they belong into a stereo 24/192, then go to 24/96. As I said before I don't care about space, and I use FLAC as the track that actually gets muxed into the movie itself.

[TinTime]

In what aspect? Industry? Definitely not for home use. I know of all the crazy coax BNC crap out there, and industrial stuff is not much better than home stuff, just different so they can charge more for it. Anyone can have any technology that isn't classified by a military, if they have the money for it.

I'm not sure what your point is because 125Mbps TOSLINK isn't supported for home use either.

[nurbs]

Yeah, because separating the source into mono wavs makes all the difference. :rolleyes:

It may well be necessary for your workflow, but it doesn't make what you do different from my description.

I did a short test. Your workflow resulted in a 40% bigger file. That might not be representative because I only tested one movie, but it goes to show that even with lossless compression there is still a significant overhead.
You might not care about filesize, but most people wouldn't waste bits on something that doesn't offer any benefit at all.

[Ghitulescu]

And what is so wrong with Sound Blaster cards? Yea their drivers suck anymore, but their hardware is amazing. I tried onboard sound ONCE and will never go back to any C-media or any generic garbage again. I have had zero problems with my PCI-express X-fi card.

The wrong is, apart from its catastrophic drivers, that it requires 3 IRQs (erh, not shared ones!). Other drawbacks can be derived from the drivers and the software it comes with.

For the same price one can get a decent semipro card (like M-Audio or Echo, they are many) will all the advantages such a card might bring. This depends however also on your priorities, because a semipro card doesn't care too much about 3D gaming acceleration and stuff.

Onboard soundcards suck, so better a SB than the onboard one.

[Ghitulescu]

Anyone can have any technology that isn't classified by a military, if they have the money for it.

For some gear they have to invest part of this money into a one-man-company, as they are not sold to individuals. :)

ANd if we're discussing about "the money", the military technology is also available.

[Ghitulescu]

You might not care about filesize, but most people wouldn't waste bits on something that doesn't offer any benefit at all.

Those people won't feel then any difference between 44.1 and 48 or between 16 and 24b, not to mention MP3 vs. WAV/FLAC. Because if they do, then they would care about.

[nurbs]

Those situations aren't the same at all. There might not be a perceptible difference between MP3 and WAV, but there is an actual difference so depending on what you want to do keeping the lossless track can make perfect sense.
Here the source is 24 bit 48 kHz, the operation is downmixing from 5.1 to stereo and the output is 24 bit 96 kHz. Upsampling doesn't provide any benefit in that case. It makes about as much sense as taking a 128 kbps MP3 and reencoding it at 192 kbps to improve the quality, except that the only drawback here is higher filesize since the resampling won't reduce the quality, unlike the reencoding would.

[vola]

Please use google translate - do us a favour, pliz pliz pliz

If you think I write wrong writing to tell me exactly what?
I have no idea what you want from me?



pliz pliz pliz

[Ghitulescu]

you have
Beautiful parts
I can not know their ability to play your file
Because I do not know the FILE AND YOUR SPIDF
Day-Old in production
Therefore possibility the playr can't play 192 000 Through SPIDF

But the point if you do not play SPIDF So HDMI will play

The question who will put 192 000 algorithm IN FFDSHOW
" resample " ???

Who Can I Turn???

It could be a hit or flop
But then we know

I don't understand what is written here, well, yes I know the words.

It was an advice, I know that many people around here, including me, are not native English speakers, but at least they try to make them understood. It wasn't disrespectful.

Translate.google.com can translate from Hebrew into English, how well, sorry, I can't tell, I don't speak Hebrew. But translations from/to FR, IT, DE, ES (some of the languages I know) are surprisingly good.

[ramicio]

People make CDs from mp3, then a friend rips that CD to mp3, and it goes on and on and I have heard how songs sound after a while. I might just try out your theory that just mixing them all together @48khz doesn't make a difference. And I don't like your comment about separating the WAVs making a difference. What the hell does that even mean? I do it so I can mix them in another program. I don't like the downmix scheme that eac3to applies so I do it manually. As I said before a GOOD SOUNDING and properly mastered CD isn't mixed in 16/44.1. You keep believing that. Maybe some cheap indie crap or any ignorant pop star CD.

And what programs do people need for a sound card? It's a device that plays sound. I don't recall of any sound card that ever came with a proprietary program that played music and became popular. For someone listening to music EAX and other DSP crap stay off. Drivers work fine. I view semipro or pro cards as what people need to get the best in recording, not so much playback. Who cares about IRQ anymore? Things are managed no so you don't have to mess with things. I remember trying to get my Sound Blaster 16 trying to work because of IRQ stuff, and even as recently as Live!. The Live! era was definitely the worst in Creative history, they are getting better.

[Ghitulescu]

Who cares about IRQ anymore? Things are managed no so you don't have to mess with things. I remember trying to get my Sound Blaster 16 trying to work because of IRQ stuff, and even as recently as Live!. The Live! era was definitely the worst in Creative history, they are getting better.

I do.

Besides, a card that needs 2-3 IRQs in an exclusive mode, while others can share 3 cards an IRQ, makes me think about how that card was designed in other places. Was it still "Who cares?" approach used or not? Remember that it was the Creative Live! the card that destroyed the fame of the emu10k.

[pandy]

As I said before a GOOD SOUNDING and properly mastered CD isn't mixed in 16/44.1. You keep believing that. Maybe some cheap indie crap or any ignorant pop star CD.

it is not mixed but it is rendered to 44.1/16 bit - in many cases at the maximum efficient way - it is like turning picture with 256 optimal color pallete and FS dither back to the 24 bit... how to recover 24 bit from 8?

[pandy]

Remember that it was the Creative Live! the card that destroyed the fame of the emu10k.

But EMU10K was designed with 48ksps on mind and for completely different purposes than PC audio cards. What is really funny - this card is still used by many musicians but with KX or EMU drivers - not Creative ones.

[Ghitulescu]

...Exactly :)

[vola]

I don't understand what is written here, well, yes I know the words.

It was an advice, I know that many people around here, including me, are not native English speakers, but at least they try to make them understood. It wasn't disrespectful.

Translate.google.com can translate from Hebrew into English, how well, sorry, I can't tell, I don't speak Hebrew. But translations from/to FR, IT, DE, ES (some of the languages I know) are surprisingly good.

I'm sorry I did not understand what you want from me I checked the inscription
Everything correctly, what do you want?

How can I understand what you write if you does not explain what the problem?
Do you see spelling Problems ?
and Where ?
Do words not understood to you?
and Where ?

[vola]

If you want to say that you would prefer to use your embedded soundcard to send 192kHz/24b over S/P-DIF, well there's a problem as S/P-DIF cannot carry this information, well it can carry 24b but chances are that the receiving end discards the extra 8 bits. 48kHz is the max sampling frequency. More info in IEC 60958.

Should one transport AC-3 signals, well, they may go a bit higher but all the advantage an algorithm in 192kHz would have is lost due to compression. More in IEC 61937.

http://en.wikipedia.org/wiki/S/PDIF
soundcard to send 192kHz/24b over S/P-DIF, well there's a problem as S/P-DIF cannot carry this information

As I wrote back and saying ALC 892 (realtek)
They say they found a way to deliver 24/192
I attach again the Document:
http://img714.imageshack.us/img714/2664/70734797.png
AND
Optical fiber again as in the picture:
There are regular SPIDF and SPIDF Optical
Again only to GIGABYTE ALC 892
http://img10.imageshack.us/img10/8862/88088124.png

Again regardless of the document or SPIDF
I wrote if you do not User SPIDF use HDMI :

Originally Posted by vola View Post
All SPIDF support today on 24/96 by manufacturers
Today ASUS + GIGABYTE now support 24 / 192 according to their answers

http://img714.imageshack.us/img714/2664/70734797.png

IF NOT SPIDF
THe last two years all ATI video cards
4000HD +5000 HD +6000HD support through
HDMI in - 24/192


So there is no hardware limit

AND If the software will BE Supports 192000
Hardware build By manufacturers going Supports Also 18MBPS
If necessary.


1. Ghitulescu- Do you understand the words literally ?
2.Ghitulescu- Do you not agree WITH something
3. I do not understand what all the debate?
There is now an appropriate hardware

AND If there is limited The manufacturers will find a way
Produce an appropriate hardware .

Now we have to make Alagoritm of 192000HZ
Again in "resample" !

[Ghitulescu]

Where does it say that your S/P-DIF transport more than 48kHz/24b in LPCM mode? First this is the highest format the S/P-DIF is allowed by standard to carry*. Secondly, any extensions to this must however either be standardized or are proprietary techniques (like ADAT).
You may wanna say it's DTS or DD, well, they are transported as user data format, and the sender and the receiver must agree on the format (channel status byte 1). You can send 96kHz or even 192kHz but in this case one must use lower bitrate. Like for MP3, if you want 20 songs instead of 15, then you have to pick a lower bitrate. Which makes ridiculous the whole idea of upsampling 48 to 192 for quality reasons if one then use aggressive compression upon it.

For your information: wikipedia is not authoritative, it's mostly written by regular people. I have an account there and if I want to change the article about S/P-DIF to write there that S/P-DIF can carry 24 channels of 384kHz and 48b in LPCM, I can do it. And until this error will be corrected by other people, some readers may believe me, because "it's wiki".

*For LPCM there are only 2 (eventual 4) channels that have a max. sampling of 48kHz. While the standard defines other sampling rates, up to 192 and low as 22.05, these are used in junction with compressed data (like DTS or DD). An indication of multichannel was also provided, in junction with the 4chn mode (which is not used any longer, I think) or it may be used with DTS or DD or any other "user data format".

As I said before, should the Gigabyte of yours deviate from the standard, then it's proprietary, and proprietary solutions means often incompatibilities with other manufacturers.

[Groucho2004]

You can send 96kHz or even 192kHz but in this case one must use lower bitrate.

This doesn't make sense. The bitrate is the product of sampling frequency and sample width (16, 24 bit, or whatever).

*For LPCM there are only 2 (eventual 4) channels that have a max. sampling of 48kHz.
Eventual? What does that mean? It starts with 2 channels and then turns into 4????

[Ghitulescu]

This doesn't make sense. The bitrate is the product of sampling frequency and sample width (16, 24 bit, or whatever).


Eventual? What does that mean? It starts with 2 channels and then turns into 4????

There's a bandwidth limitation, for the first point. To transport more one needs compression, which transform a higher bitrate into a lower one. Legal compression algorithms are MPEG-2, DTS, DD and the new AAC and another new one I forgot its name.

There were trials on 4 channels long time ago, this modus corresponds 1:1 to one of the 32kHz modi of a pro DAT (as the consumer ones cannot record/play more than 2 channels; it's called LP on these). AFAIK, the 4 channels was used on some SAT transmissions, in some TV stations, and for this reason was implemented in DATs and therefore in S/P-DIF.

If you want more details just read the standard.

Eventual NOT EQUAL eventually.

[Groucho2004]

Eventual NOT EQUAL eventually.

Seriously? You are giving me a lecture in English?