Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Enhanced 20 Bits Surround sound? (Read 15594 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Enhanced 20 Bits Surround sound?

Reply #50
Quote
Yep, I'm planning to try exactly this at home to see what happens. I already just tried with Virtual Audio Cable, but it doesn't work with 24 bit data (I believe you found out same thing).

I got it to work with 24bit and 32bit, e.g. to record fb2k (before diskwriting was introduced) or XMplay high resolution output with CEP under Win98. But, as said on NTONYX hompage, WinXP that is needed for 24bit HDCD playback is not supported. (In fact VAC works in WinXP, but it's not able to make WMP9 and CEP like each other at 24bit).
Let's suppose that rain washes out a picnic. Who is feeling negative? The rain? Or YOU? What's causing the negative feeling? The rain or your reaction? - Anthony De Mello

Enhanced 20 Bits Surround sound?

Reply #51
Quote
I was wondering how this topic was getting so long... Hi, budgie

Hello, long time no see  Do you really think, that just I am a thread-stretcher on this forum? 

Enhanced 20 Bits Surround sound?

Reply #52
Quote
Use WMP9 with 24bit ouput for HDCD playback (WinXP, 24bit soundcard needed)

Uosdwis said, in here, that a soundcard could not be used to play HDCD, because it almost sounds like asking to simulate a DAC in software.

Enhanced 20 Bits Surround sound?

Reply #53
Well, yeah, the PMD100 chip (or substitute) is part of the HDCD decoding process, and while I don't have researched this particular bit very thouroughly it appears counter-intuitive to think that you can simulate the "complementary" filter and then hit another filter in a generic DAC, without missing out on some of the claimed benefits. This, of course, presupposes that the PMD100 filter does the job better than any (or most) generic filters out there, which I find reasonable in this context (why else bother to include it in the specification?). One could debate the quality of the PMD100 filter vs. other filters generally, though, and the result may end up being an argument against implementing the HDCD format. It appears this is the stance taken by most high-end manufacturers.

As for the test methods proposed by KikeG and tigre, they can be useful for evaluating some aspects of the decoding part of HDCD. But if you mean that it can be useful when trying to decided the benefits of HDCD over regular CDs, it appears you are forgetting one crucial aspect of the HDCD format: HDCDs can only be created one way: from passing the source through the PMI/Euphonix Model Two (or its predecessor Model One).

Now, this unit can accept digital as well as analog signals, but the general concensus (and this rings true with the information on the HDCD website) appears to be that to obtain the greatest benefits of a HDCD encoded CD, you should let the Model Two handle the analog to digital conversion. Not only because of the supposedly stellar conversion, but also because of the feature discussed above: switchable and complementary A to D and D to A filters. When creating HDCDs, this box also employs what they call "dynamic decimation" which can work in a similar fashion depending on the sample rate, and also sports the gain structure process, and additional dithering. You can however work the Model Two as a regular ADC or DAC as well, and this is not uncommon, from what I understand.

The fact that this one box is such an integral part of the format - in fact, it IS the format - makes HDCD vs. CD different from comparing DVD-A to CD, for instance. In the HDCD vs. CD debate, what you are in fact comparing, is the Model Two processor (with/without the decoding part, depending on which aspect you wish to test), to any combination of units/techniques used in creating (and playing back) a standard Red Book CD. Needless to say, it's a project that you have to devote some serious time and money to perform correctly. I doubt it is at all possible to reach one single and universal conclusion; it may well vary depending on the source material, not to mention the vast combinations possible with the traditional methods (the various combination of ADCs and DACs (including filters used by both), dithering algorithms, noise shaping curves and so on).

This is why you can't make any comparisons using only a HDCD encoded source; the fact that the audio has already passed through the Model Two can't be undone (and certainly not by decoding!).

Hope this clarified my point.

-- Uosdwis

Enhanced 20 Bits Surround sound?

Reply #54
Well, I'm sure that with enough processing power, you can duplicate via software and in real-time the HDCD decoding process, and generate an equivalent 24-bit signal. 32 and 64-bit floating point processing can surely overcome all resolution limitations in a proper software HDCD decoding process. A final 24-bit output data provides more resolution than any DAC or electronics can do in real world. At last, you would be limited by your actual 24-bit DAC, but I doubt HDCD DACs + HDCD dynamics processing can be much better than good 24-bit DACs available today (HDCD decoding is common in many inexpensive cd players now).

Still, I don't know if 24-bit output of Microsoft HDCD decoding engine would be a proper "decode".

Anyway, the method I proposed consisting in just recording the analog output of a HDCD player, downconverting it to 16 bit, and just compare this with that same HDCD analog output, surely would work.

Edit: whole post edited for a better explanation.

Enhanced 20 Bits Surround sound?

Reply #55
I have to agree with KikeG, though I am by no means very knowledgeable in the matter:

If you succeed in converting the output stream of an HDCD to the standard Red-Book CD Audio format in a non-distinguishable way, you have proven a lot:

To human ears, the theoretical improvements provided by the HDCD format represent NO actual improvement over standard Red-Book CD Audio.

Or some other way to look at it: you can produce Red-Book CD Audio that sounds just as good as HDCD, as long as you master carefully the sound.

<offtopic>Unlike what nowadays is performed to the commercial CDs, with all that clipping...</offtopic>

Enhanced 20 Bits Surround sound?

Reply #56
Of course HDCD can be processed in software, and after you can use a regular 24bits DAC.
Do you really think that HDCD needs a different DAC? The purpose of a DAC is to do the DA convertion, nothing else. If you have a DAC that is advertised as doing something else, I'd say that it is a DAC and a DSP inside the same chip.

If it was impossible to decode HDCD in software, how could some DSP decode HDCD in software?

Enhanced 20 Bits Surround sound?

Reply #57
Quote
Well, I'm sure that with enough processing power, you can duplicate via software and in real-time the HDCD decoding process, and generate an equivalent 24-bit signal. 32 and 64-bit floating point processing can surely overcome all resolution limitations in a proper software HDCD decoding process. A final 24-bit output data provides more resolution than any DAC or electronics can do in real world. At last, you would be limited by your actual 24-bit DAC, but I doubt HDCD DACs + HDCD dynamics processing can be much better than good 24-bit DACs available today (HDCD decoding is common in many inexpensive cd players now).

Still, I don't know if 24-bit output of Microsoft HDCD decoding engine would be a proper "decode".

Anyway, the method I proposed consisting in just recording the analog output of a HDCD player, downconverting it to 16 bit, and just compare this with that same HDCD analog output, surely would work.

Edit: whole post edited for a better explanation.

So what you're interested in is comparing a fully decoded "pure" HDCD signal to a fully decoded HDCD signal that has also passed through an ADC, been normalized (to compensate for the peak extend/ low-level processing) and then had dither applied to it a second time? To me this doesn't sound too useful when comparing HDCD as a format to CD as a format. If all you want to do is claim there's no real world advantage with 20 or 24 bit audio vs. 16, you're better off finding a source that isn't HDCD encoded (since that has already had its word length reduced). If you want to prove you can take all the so-called advantages of HDCD processing and cram that into a regular CD without you being able to tell the difference, I agree that your test method is somewhat useful. But it doesn't really deal with evaluating HDCD as a format.

Your remarks in the end of the first paragraph, regarding "good 24-bit DACs available today" and how they compare to the DACs found in HDCD equipped players, are perfectly valid. I find it interesting that so few professional or high-grade consumer DACs have implemented the PMD-200 chip [corrected; it's 200, not 100](although it may have to do with licesing issues etc.).

And in answer to Gabriel's post, you may very well be correct. I have not studied the patent application to see if it clarifies things, and I don't intend to. If someone is willing to invest the time, it would be interesting to hear what they come up with.

This is from the HDCD Technical Paper (PDF):

Quote
Finally, the signal is interpolated to twice the sampling frequency using a filter which is
complementary to the anti-alias filter used in the encoder. This signal is available as the
output of the process, or, in the IC, it can be further interpolated to a four or eight times
oversampled signal to drive common D-A converters. The IC also incorporates features
designed to improve the performance of multi-bit converters, such as selectable levels of
supersonic dither and output timing designed to reduce conversion timing jitter.


If this is correct and complete, and the software decoding process is as well, it means you should get an 88.2 kHz output from WMP9. If not, something is not happening as described above, or something in addition to what's described above is happening (such as downsampling again). The kind of upsampling mentioned later (four or eight times) is the kind that is done as sort of the next-to-last step in many DA converters, and not something that can be done inside Windows; no soundcard will accept a 705.6 kHz signal.

Based on what I can find on the HDCD website, it is my understanding that the PMD-200 chip is needed for the filtering part of the decode process, and that it needs to be coupled with an approptiate DAC (hardwired), as listed on the IC Application Chart. I'd be happy to be corrected.

-- Uosdwis

Enhanced 20 Bits Surround sound?

Reply #58
The pmd-200 itself is just a programmed dsp chip, nothing else:
http://www.hdcd.com/partners/tech_partners...200Overview.htm

By searching "hdcd dsp decoder" on google, you will find many other general purpose dsp which are able to decode hdcd with the proper decoding software.

Enhanced 20 Bits Surround sound?

Reply #59
Excellent! That particular page had indeed escaped me. So, it's THAT kind of filtering. Interesting. Can't say that it makes me more impressed with the format, though.

But do you also mean to tell me that WMP9 is indeed a full-fletched HDCD decoder? If so, how do you comment on the upsampling issue?

-- Uosdwis

Enhanced 20 Bits Surround sound?

Reply #60
According to tests done by other people here, it seems that the only thing that wmp9 is doing on hdcd is detection and displaying a logo. Only marketting and placebo in this case.

But it should have been possible for it to do hdcd decoding. There are hdcd decoding chips that are only supporting 44.1kHz:
http://www.hdcd.co.kr/chips/chips3c.htm

So 88.2kHz is no a requirement.

Enhanced 20 Bits Surround sound?

Reply #61
Quote
According to tests done by other people here, it seems that the only thing that wmp9 is doing on hdcd is detection and displaying a logo.

From the test I did there's no conclusion like this possible. It  just didn't work because WMP9 needs WinXP for 24bit output but the recent version of NTONYX Virtual Audio Cable that was supposed to capture WMP9's 24bit output doesn't work under WinXP properly. ATM the only possibility I know of to do this test would be to use a soundcard with 24bit digital out as I wrote before in this thread.
Let's suppose that rain washes out a picnic. Who is feeling negative? The rain? Or YOU? What's causing the negative feeling? The rain or your reaction? - Anthony De Mello

Enhanced 20 Bits Surround sound?

Reply #62
Well, I've been able to record 24-bit data at the digital output of my 8738 card, playing an HDCD disc with WMP9 in WinXP. Recorded data has resolution over the 16th bit level. Still, I'm not 100% sure that WMP9 + my card produces bit-perfect digital output at 24 bit mode. Quite likely it does, but I haven't had time to verify it.

As to the test consisting in recording the output of an HDCD player and convert it to 16 bit, if it sounds the same, it doesn't matter the process carried. It would mean that plain 16 bit can sound as good as HDCD.

Enhanced 20 Bits Surround sound?

Reply #63
Just a note.

I don't know if it is because my soundcard doesn't support 24bits or what, but the HDCD option in WMP9 for my configuration was disabled by default. (I have no HDCD here so I haven't even tested if it worked or not).

I just want you to be sure that you have the decoding enabled.

Go to Options, "Devices" tab, select Speakers, and press "Properties". There are two options there, being the 2nd one "Use 24bits for audio CD".

(Of course, it requires digitally extracting CD data, but I am sure you know it already :·P)

Enhanced 20 Bits Surround sound?

Reply #64
KikeG:

Not that it is too related to this topic, but does the 8738 (CMI, right?) resample the digital output to 48kHz like most other codecs?

I have one in box somewhere, and if it doesn't, I might put it to use.

-- Uosdwis

Enhanced 20 Bits Surround sound?

Reply #65
I've been using M-Audio Dio2448 WinXP drivers, which are better than generic CMI8738 drivers, and setting all wave mixer sliders at max, at 16 bit it has verified bit-perfect output. However, sometimes the card simply refuses to play at 44.1 until I reboot the computer. Could be my CMI8738 motherboard implementation, though.

With generic CMI8738 drivers, I was unable to get bit-perfect output. There was no resampling, but the output values were always scaled below 100% no matter the mixer sliders.

 

Enhanced 20 Bits Surround sound?

Reply #66
Interesting.

A local store here sells a 8738-based card with a full-size S/PDIF output (and screaming "24 bit" all over the box) for less than €20. If I can get it to work properly it would be a great (and certainly cheap!) solution for "jukeboxing" with an external DAC. Hell, I'll buy 3 of them, and put one in each computer I use.

Thanks for the information.

-- Uosdwis