AC-3 vs CD

Hi. I was just listening to a dvd (AC-3 in stereo and LFE setup) and found that the audio had more detail than a cd. (first impression)

My question is seeing that DVD-A discs will include AC-3 will it be worth getting a DVD-A disc player.

According to dolby, AC-3 can have a resolution between 18 and 22 bit and can sound transparent from as low as 320kbps data rate.  

I found that AC-3 had more detail in high frequencies.
Is it true that high resolution audio has more audible detail especially in the high frequencies.

The way I know it is High bit depths replays the small signals more transparently ( the small waves (HF) that are superimposed on the big waves (LF)) which are then amplified through speaker crossovers which is what would happen if you done a high pass filter at 10khz on CE
the smaller signals would become larger.

But I guess my real question is do you think DVD-A (AC-3) is better than CD?

I wouldn't of thought AC3 was an improvement over CD audio as it has been lossily compressed.  320kb might have been deemed transparent for the vast majority of people, but if you have good hearing this might not be enough.

I remember from reading some old Widescreen Review magazines that DTS is highly preferred over AC3, but I would have to say their opinion sounded more than a little coloured.  DTS is a better format, more advanced algorithms and operates at a higher bitrate usually, but my guess is it's still below cd standard!

Could it be that my dvd player sounds better with 48khz material than 44.1khz material? (poor hardware design sounds better with higher sampling rates)
Is it possible for someone to do an ABX on this? (to show us some results)

Could it be that my dvd player sounds better with 48khz material than 44.1khz material? (poor hardware design sounds better with higher sampling rates)
Is it possible for someone to do an ABX on this? (to show us some results)

It's possible, but not the most likely cause, IMO.  This is the same issue that makes ABX difficult on this - you need identically mastered versions to compare, and I suspect there are very few of these out there (and how would you know if you had one?

You could start with a 24/96 wav file and encode yourself, but AC3 encoders are pricey.

This is on the right track.  Perhaps there is a hardware explanation for the apparent improvement of 48K material.  But I can speak from experience that stereo AC-3 is audibly inferior to the PCM source, at least in the case of one mix that I supplied to such an encoder, probably at 256kb/s, with a 2.0 config.  It definitely is lossy.

Tef, the best test (not that any of us can actually do this!) would be to compare the AC-3 tracks to the PCM equivalent on the same DVD-A disc.  I look forward to the day when a number of us can readily discuss that.

You could start with a 24/96 wav file and encode yourself, but AC3 encoders are pricey.

TMPGenc AC3 (2-channel) encoder: $29

http://www.pegasys-inc.com/en/product/tsp_ac3.html

Well, there ya go.  I wonder why I thought it was pricey...  Maybe it's only 5.1 that costs the big bucks?

Problem is I haven't got a DVD - burner, I'm waiting till dual layer DVD burner's become available here in Australia.
I'm looking forward to HD audio.

Well, there ya go.  I wonder why I thought it was pricey...  Maybe it's only 5.1 that costs the big bucks?

Exactly!  Precisely!

Well instead of encoding wav to AC-3 I have decoded AC-3 to wav and then burned a audio cd from it.

I compared the dvd and the cd (short segment).
The dvd definetely sounded better so it definetely is a sampling issue.

I think I might buy a DVD-A player just to listen to my upsampled cd's.

Well instead of encoding wav to AC-3 I have decoded AC-3 to wav and then burned a audio cd from it.

I compared the dvd and the cd (short segment).
The dvd definetely sounded better so it definetely is a sampling issue.

I think I might buy a DVD-A player just to listen to my upsampled cd's.

You can't rule out the effects of the 48k to 44.1k resampling.  That could theoretically be part of the difference in your current test.

I don't believe going from 48k to 44.1k is the problem, I used the highest
possible downsampling accuracy in CE.

But now that you have mentioned it, can you please explain what type of effects there is downsampling? I really like to know.

I don't believe going from 48k to 44.1k is the problem, I used the highest
possible downsampling accuracy in CE.

But now that you have mentioned it, can you please explain what type of effects there is downsampling? I really like to know.

Actually, that wasn't a great example.  In a cheesy resampler, added noise and possible artifacts at difference frequency and multiples might show up.  But I'll agree with you here that these aren't likely in CEP.

It IS possible that your specific hardware does better with 48k; maybe since it's primarily a dvd player it does some funky internal resampling for 44.1k sources.  But 48k is not really that big a difference over 44.1 - rip some DVD audio to WAV and frequency analyze it.  I'd be surprised if there's much at all above 20k.

I don't believe my dvd player resamples internally.(at least not 44.1k)
It's maximum digital out (coax and optical) sampling frequency is 48k.
So I think(strongly believe) it downsamples 96k to 48k for digital output etc....(only)
Yes I have noted the frequency anlaysis and there is no frequencies above (around) 17khz.

I thought the difference in sound quality was to due to the phase response of the recontruction filter..??

Thats why there are upsampling cd players? Isn't it?
Maybe steve would like to comment.

I don't believe my dvd player resamples internally.(at least not 44.1k)
It's maximum digital out (coax and optical) sampling frequency is 48k.
So I think(strongly believe) it downsamples 96k to 48k for digital output etc....(only)
Yes I have noted the frequency anlaysis and there is no frequencies above (around) 17khz.

I thought the difference in sound quality was to due to the phase response of the recontruction filter..??

Thats why there are upsampling cd players? Isn't it?
Maybe steve would like to comment.

I'll let Steve handle this if he wants; he's better equipped for it.  But you can't get anything for free.  Upsampling cannot improve the sound quality, it can ONLY worsen it (even if the effect is inaudible).  This is different from oversampling, which is more or less a technique of digital filtering.