The Importance of the "Original-Speed" Audio

[Moshrom]

I don't think so. At least, not in the way you probably mean. In some cases I've synced DVD mono tracks to blu-rays that only offer surround remixes. But in most cases, they're just the same mixes but with better fidelity.

For example, right now I'm doing the mono mix of A Clockwork Orange from a source that sounds noticeably better than the LaserDisc.

I also recently did the entire World at War series, syncing the mono tracks from the R2 DVDs to the new 4:3 Network blu-rays.

[Chewtobacca]

I also recently did the entire World at War series, syncing the mono tracks from the R2 DVDs to the new 4:3 Network blu-rays.

That's great. I'd love to hear the result. Apart from the audio, the new BDs are a big improvement.

[Stamper]

Awesome Moshrom, concerning Decalog, I had no idea, what a travesty!

[jerryshadoe]

Yeah, in my original audio, I was showing a lossless audio source to demonstrate the differences VERY clearly in the different alterations of the audio. Of course, these differences occur when dealing with lossy audio too, but an example wouldn't show it that defined. However, taking that into consideration, if you are already dealing with a lossy source and making these alteration, the fidelity/dynamic range is slightly compromised. Granted, most people will NOT hear the differences, but we are kinda "purists" here (or audiophiles) and a lot of us DO notice, especially when dealing with heavily compressed audio.

On note of heavily compressed audio, people don't realize that in a lot of cases 2 CHANNEL (I can't stress that part enough) 192kbps AC3 audio will sound BETTER than a 5.1 384kbps audio track (which is very commonly used in HDTV broadcasts) and the reason comes down to the math here of the bitrate...

Let's see...

192kbps divided by 2 (channels) equals 96kbps PER channel...

on the other hand...

384kbps divided by 6 (5.1 is technically 6 channel audio) equals 64kbps PER channel...

so, for 5.1 AC3 audio to be equal in quality to a 2.0 AC3 track, it would have to be encoded at 576kbps...

There are numerous HDTV broadcasts that have 2.0 AC3 at 384kbps and when they are not a fold-down of the 5.1, these are excellent sources for some original audio.


I'm not surprised that there are 2.0 stereo/dual mono/mono AC3 tracks at 192kbps that can sound better than even LD PCM. A lot of all of this is dependent on the original source of the transfer. I was just showing the difference on the same transfer to illustrate it easily...

[Moshrom]

Essentially, yes.

But my point is that fussing over the specs of a lossy encode for an audio track is usually the most menial factor at play. I'll always level-match and A/B any two tracks I'm comparing, then look at their spectrals to validate my findings, and then, if the two tracks are audibly similar, I'll choose the one with the most robust encode.

Of the >100 films I've done this for, it very rarely came down to that last bit.

Typically, the source elements from which the track was transferred is the most important determinant of sound quality. Then comes the mastering/processing (NR, compression, etc.). After that, the encode.

And I'll just add that I am a very critical listener and do all my comparisons with a pair of HD600s playing from an ODAC and O2 amp. I'm in my twenties and can hear to 20 kHz. 96 kbps/channel for an AC3-encoded file is usually transparent enough -- if a more robust encode exists, I'll take it, but it isn't really a priority for me.

[jerryshadoe]

We are reaching the same destination, but taking different roads to get there. I can usually spot the difference visually, just by looking at the spectral analysis, before I even listen to the track/compare waveforms/ or do A/B tests and have been right in nearly every case except when the two look nearly identical.

Of course, it always comes down to source of the original audio, then the processing, then the compression - in that order. I agree with you there completely. However, just because a track at 96kbps per channel sounds "ok" doesn't mean it IS "ok" IF it were to be compared to a lossless encode sourced from the SAME master. This is why a lot of us here prefer audio from laserdiscs because it's PCM. Granted, it's NOT always better because, as we already both acknowledged, it comes down to the source

I'm VERY happy to see someone, besides myself, that looks at these things. I've been matching the "best" audio tracks with the "best" video tracks for years, because official releases usually include only one or the other

[Moshrom]

IF it were to be compared to a lossless encode sourced from the SAME master.

If they're from the same master, absolutely!

I can usually spot the difference visually, just by looking at the spectral analysis, before I even listen to the track/compare waveforms/ or do A/B tests and have been right in nearly every case except when the two look nearly identical.

As can I, but I try not to when comparing tracks from different masters/restorations, to minimise whatever bias I can. ...Especially since I post my findings publicly and there's always a faction of sceptics out there who are quick to jump on the dude pointing at spectrals and waveforms and condemning the disc they just paid for.

[jerryshadoe]

I see the point you are making when it comes to the "nay-say'ers" and having to justify yourself when making a comparison. And, I understand how some people could look at it as "a bias," but I must say that I categorically disagree with that viewpoint personally. If I was a "noob" to the "art of audio" then maybe I could claim "bias" but I have been editing audio since the late 90s in the digital realm and even longer in the analog realm. I believe that looking at a spectral analysis can tell me EVERYTHING I need to know about an audio track BEFORE I ever listen to it. This is NOT "bias," but EXPERIENCE. About two decades of experience, combined with excellent and very acute hearing (I have been truly "blessed" here and verify this with actual hearing tests every 6 months since I was 16) and the proper equipment has helped me get to where I now have "opinions" about "things."

Not sure about your amount of experience in the field, although you sound like you have at least a basic grasp on things, but here (in this forum) if you claim "spectral analysis," you won't get your head chewed off

HOWEVER, just to play my own "Devil's advocate" (LOL) - "spectral" isn't always everything and the best example of this are some of the Vinyl LPs that I have captured (I actually have a rather big collection there) where it looks like the Vinyl is "better" than it's CD counterpart, but this isn't the case. The vinyl sounds like crap because of the source it was transferred from or how it was transferred or just the pressing... And, those differences can only be heard when doing an A/B comparison.

This is why I NEVER really on just one method or the other (spectral vs listening) and use a combo of both... However, in most cases - I just happen to be right based off the spectral alone

Also, I NEVER judge ANY track based of the waveform, because those are very hard to use to gauge anything other than volume level in db...

[Moshrom]

Haha, I agree.

Cases where spectrals can be misleading: when a track transferred from a higher generation source has received less destructive processing than a track transferred from a lower generation source but subsequently no-noised (for example).

The spectrals of vinyl pressings originally mastered from nth generation sources, as you say, would demonstrate this too. I have a lot of experience with that as well :-)

[ilovewaterslides]

on the other hand...

384kbps divided by 6 (5.1 is technically 6 channel audio) equals 64kbps PER channel...

Except that I don't think each track is encoded with the same bitrate. I believe that the main stereo channels use a higher bitrate than the bass channel for instance.