AC3 is to DTS as MP3 is to AAC?

[Ditto666]

I know there's no direct relation but let me explain... I figure that largely, it's a matter of taste. From what I've heard through just casual listens, DTS vs. AC3 has similar qualities that differentiate them as AAC vs. MP3 does. DTS has a cleaner, crisper sound with less tonal accuracy while AC3 is more accurate but more muddy sounding (especially low frequency). DTS is tighter while AC3 is more loose which can be opinion but really, it's usually seen as the tighter the better. The reason the title has a question mark is because for DTS and AC3, I've heard it myself and then it was confirmed by others, while with AAC and MP3, it's basically been with DIFFERENT songs my listens, only occasionally same tracks that I've heard though after large gaps of time, and just from the general sound I picked up those same attributes. Literally if I were asked now I'd give that same description in relation between the two aside from the tightness aspect.

With these differences, for stereo often times I end up with a preference for MP3 over AAC for music as would make sense, though I enjoy AAC also as more of a gimmick in this case I guess, while DTS over AC3 for music in multichannel surround. For movies, I haven't heard too many stereo AAC encodes but I assume I'd prefer it, and DTS hands down for movies with multichannel.

Would this be a correct way of looking at it? When people ask the question, "which format is better", I'm pretty sure they're asking for a description of what separates the different existing formats. The only way something can be better is if it sound exactly the same or at least very similar to another format with a lower bitrate. How would you guys recommend this. I never really got the hang for encoding audio just because it's hard to tell what's "right" vs. simply something visual, easily comparable. Audio heavily relies on a chain of memory whilst creating opinionated differences. At one point I was using Vorbis for movie stereo because I felt it had the best bang for its buck so-to-speak, in this case in relation to bitrate, until I realized that the sound is REALLY displaced towards "clarity" and sharpness with large sacrifice to accuracy and fullness. On top of that, it's not very well supported. MP3 for these purposes is the exact opposite. AAC is a viable option but somehow nonsensical, maybe because of its rarity (at least in my experience). For surround audio, nothing really makes more sense than to leave the original from what I see around as options. So yeah.. thoughts? :)

[nurbs]

AC3 is better than DTS at the same bitrate. Blind test have shown that 1.5 Mbps DTS is about the same quality as 640 kbps AC3 while 768 kbps DTS came out worse.
When people hear a difference between high bitrate encodes in both formats it's often the Dialog normalization. AC3 has it activated by default and many decoders don't allow you to turn it off, so you end up comparing a normalized AC3 with an untouched DTS. Since Dialog normalization is done after decoding and the value is part of a tag in the AC3 file it's easy to get rid of it.

[Ditto666]

AC3 is better than DTS at the same bitrate. Blind test have shown that 1.5 Mbps DTS is about the same quality as 640 kbps AC3 while 768 kbps DTS came out worse.
When people hear a difference between high bitrate encodes in both formats it's often the Dialog normalization. AC3 has it activated by default and many decoders don't allow you to turn it off, so you end up comparing a normalized AC3 with an untouched DTS. Since Dialog normalization is done after decoding and the value is part of a tag in the AC3 file it's easy to get rid of it.

Yeah, I of course know that DTS requires a much higher bitrate. I was comparing 640KB/s AC3 with 1.5MB/s DTS. The description I gave in the sound was between those two. Sorry that I didn't mention that. I was comparing the formats at their "perfect" point.

What were you trying to say by the normalization? You're saying that's what usually makes it sound worse? You disagree with the differences I described? It's completely different engineering of the sound the two formats... If you're trying to tell me that with the normalization of AC3 removed that they sound very similar, then I would strongly have to disagree.

[nurbs]

The Dialog Normalization in AC3 causes a volume change. Difference in volume is easy to hear, so that makes direct comparison between DTS and AC3 a bit tricky if the DN hasn't been deactivated in the decoder.

Personally I don't find the way you described the differences you hear descriptive. From my experience there is generally no significant difference between high bitrate AC3 and DTS tracks apart from the size. Many on this forum agree and there are also listening tests (http://tech.ebu.ch/docs/tech/tech3324.pdf) that suggest the same thing.

Your original question was what the better (lossy) audio format is.
In my mind it comes down to quality, efficiency, compatibility and availability of tools to work with the format. In my opinion AC3 and DTS can deliver equivalent quality and for the other points AC3 wins, so that's what I'd personally use.
Still there are reasons to keep DTS tracks if you already have them, for instance to prevent losing quality by reencoding.

[tebasuna51]

... You disagree with the differences I described?
I disagree, especially with: "more muddy sounding (especially low frequency)"

Seems you have two different source mixes, or different encoding settings, or different decoder settings, because low frequency is very easy to encode and we can't expect differences at all.

We can respect all opinions, but this discussion was been done many times is this forum and I can't accept a new one because forum rule 12:
Do not ask "what's best" because this question cannot be answered objectively. Each and everyone has their own view about what's best in a certain area. The best is what works best for you!

Then please follow the rule and don't put subjective opinions. Read test, make test yourself, and select the best for you.

About multichannel test I only know:
http://tech.ebu.ch/docs/tech/tech3324.pdf

For stereo test you can see http://www.hydrogenaudio.org/

Thanks.

[Ditto666]

The Dialog Normalization in AC3 causes a volume change. Difference in volume is easy to hear, so that makes direct comparison between DTS and AC3 a bit tricky if the DN hasn't been deactivated in the decoder.

Personally I don't find the way you described the differences you hear descriptive. From my experience there is generally no significant difference between high bitrate AC3 and DTS tracks apart from the size. Many on this forum agree and there are also listening tests (http://tech.ebu.ch/docs/tech/tech3324.pdf) that suggest the same thing.

Your original question was what the better (lossy) audio format is.
In my mind it comes down to quality, efficiency, compatibility and availability of tools to work with the format. In my opinion AC3 and DTS can deliver equivalent quality and for the other points AC3 wins, so that's what I'd personally use.
Still there are reasons to keep DTS tracks if you already have them, for instance to prevent losing quality by reencoding.
It's not volume that I mentioned. I am aware of such differences and I try not to factor them in.
I disagree, especially with: "more muddy sounding (especially low frequency)"

Seems you have two different source mixes, or different encoding settings, or different decoder settings, because low frequency is very easy to encode and we can't expect differences at all.

We can respect all opinions, but this discussion was been done many times is this forum and I can't accept a new one because forum rule 12:


Then please follow the rule and don't put subjective opinions. Read test, make test yourself, and select the best for you.

About multichannel test I only know:
http://tech.ebu.ch/docs/tech/tech3324.pdf

For stereo test you can see http://www.hydrogenaudio.org/

Thanks.

I wasn't asking for opinion. I know about this rule and I specifically mentioned that rather than "what's better", I'm asking for the differences in the sound. What's better is, well yeah, it's opinion. The differences in tone or any other element as such is actual fact.

In terms of the sound of DTS vs. AC3, it can't be coincidence that such a large number of people have heard exactly the same differences before hearing of them. Anyway, it's aside the point. I was actually ASKING if I'm correct or not. I didn't declare anything. If incorrect, I was asking for a description describing otherwise.

Just like different video codecs, if not even more so, different LOSSY audio codecs sound different, and their flaws are more easily heard as you lower the bitrate, obviously. That gradual deterioration of the sound is quite different going from format to format, and those differences in compression of sound still apply at the formats' peak.

[nurbs]

I was actually ASKING if I'm correct or not. I didn't declare anything. If incorrect, I was asking for a description describing otherwise.
I told you that many think you are incorrect and both tebasuna51 and I linked a study of the European Broadcast Union that puts DTS at 1.5 Mbps and AC3 at 448 kbps in the same quality range.

[Ditto666]

I told you that many think you are incorrect and both tebasuna51 and I linked a study of the European Broadcast Union that puts DTS at 1.5 Mbps and AC3 at 448 kbps in the same quality range.

Ok then, exactly. I didn't say you did anything wrong. I was saying that I haven't done anything wrong either.

Regardless though, in terms of the audio, while it may put it in the same range in its own way, 448KB/s AC3 would not SOUND the same as 1.5MB/s DTS. What's "better quality" is debatable, or in terms of this analysis, equivalent, but either way, they would have their own differing attributes.

[GodofaGap]

Regardless though, in terms of the audio, while it may put it in the same range in its own way, 448KB/s AC3 would not SOUND the same as 1.5MB/s DTS.
This is not an assumption you can make upfront when talking about perceptual coding. Normally the intent of perceptual coding is to sound the same as the source (while not being the same), so ideally both WILL sound the same.

If there are differences it is not in the terminology you use, because words like "cleaner", "crisper" and "tonal accuracy" are audiophile buzzwords and not descriptions of artifacts introduced by a perceptual coder (ringing, pre-echo, etc.).

[Ditto666]

This is not an assumption you can make upfront when talking about perceptual coding. Normally the intent of perceptual coding is to sound the same as the source (while not being the same), so ideally both WILL sound the same.

If there are differences it is not in the terminology you use, because words like "cleaner", "crisper" and "tonal accuracy" are audiophile buzzwords and not descriptions of artifacts introduced by a perceptual coder (ringing, pre-echo, etc.).

You are right. Those are words quite often used to describe and compare the sound of certain speakers/headphones vs. others. I was applying this terminology here. I wouldn't call it buzzwords. They are really just descriptions of sound and they could apply if you let them. I say this because yeah, you say the intent is for it to sound the same as the source, of course, but in the end that's never the case. The coding is different for each and it creates different forms of compression, creating differences in the sound. While one might sacrifice clarity, another might sacrifice accuracy.

Sigh... If only I could encode something with DTS, I could then make a comparison where short segments are played one after another, one encoded to AC3, the other to DTS. The only problem then would be people's listening equipment but they'd be asked to list what it is which would make thing clearer.

[GodofaGap]

You are right. Those are words quite often used to describe and compare the sound of certain speakers/headphones vs. others. I was applying this terminology here. I wouldn't call it buzzwords. They are really just descriptions of sound and they could apply if you let them.
I can't apply them because I have never heard or read a satisfactory definition of them. When sound is "crisp", what does that mean? And how can something be "accurate" and "muddy" at the same time?

I say this because yeah, you say the intent is for it to sound the same as the source, of course, but in the end that's never the case.
I think this is quite a bold statement to make. Especially for AAC and MP3.

I think you first need to do some learning in how to evaluate perceptual coding. The descriptions you have given so far don't make any sense, and neither do the conclusions you draw from them. Sorry if this sounds a bit hostile, but you have already been pointed in the right direction by tebasuna51. I suggest you stop thinking you know anything about compressed audio and relearn from there.

[Ditto666]

I can't apply them because I have never heard or read a satisfactory definition of them. When sound is "crisp", what does that mean? And how can something be "accurate" and "muddy" at the same time?


I think this is quite a bold statement to make. Especially for AAC and MP3.

I think you first need to do some learning in how to evaluate perceptual coding. The descriptions you have given so far don't make any sense, and neither do the conclusions you draw from them. Sorry if this sounds a bit hostile, but you have already been pointed in the right direction by tebasuna51. I suggest you stop thinking you know anything about compressed audio and relearn from there.

I don't act like I know anything past the basics with compressed audio. You don't sound hostile at all. We have understanding from opposite sides. While I don't know most of the technical elements of how the compression works exactly, you can't make anything out of those descriptions which tells me that you don't have any perspective on the actual sound itself. I guess you at least a "musical ear" to begin with. It's not all hocus pocus BS as that's actually what all the technical elements are based around. The best sound technology comes from people that hear how sound should be placed. The knowledge of sound engineering is not the same as being a sound engineer which has that to begin with as the basis for any action. The fact that you can't apply the words is actually lack of knowledge/understanding on your end. You likely never had much of an interest and that's fine, but my words are not the words of blind inception.

It's a bold statement to say that it's not the case that MP3 and/or AAC doesn't sound the same as the source, or even close for that matter?... Cowon, or a company you're probably more familiar with, Samsung must be spending millions of dollars on nothing. People must be crazy for utilizing FLAC - so much more space for insignificant differences. The fact of the matter is, the terminology I used would be understood and preferred on many other professional forums that deal more with playback like speakers, receivers, PMPs, etc...

These differences are what is for me and SHOULD BE the foundation for choosing one format over another. How well it works with whatever container for instance is a secondary concern. Of course that's not all a decision otherwise would be based around but you get the idea... Everything else after that is secondary.

Cowon: http://www.chattershop.com/showthread.php4?t=6684
Golden Ears (as I'm just applying towards Cowon; there's a more general notion behind this): http://en.wikipedia.org/wiki/Golden_ear

[GodofaGap]

It's not all hocus pocus BS as that's actually what all the technical elements are based around.
I think you'll find that there is a lot of BS in the audiophile segment.

The fact that you can't apply the words is actually lack of knowledge/understanding on your end. You likely never had much of an interest and that's fine, but my words are not the words of blind inception.
We can only use words if they have clear definitions (even sound engineers). If these words have an agreed upon definition used that I must be able to find it somewhere.


It's a bold statement to say that it's not the case that MP3 and/or AAC doesn't sound the same as the source, or even close for that matter?... Cowon, or a company you're probably more familiar with, Samsung must be spending millions of dollars on nothing.
Firstly these companies are not altruistic in nature. They spend millions of dollars to make millions of dollars more. If they'd have to spend 1 million to make 2 million out of selling you back your own farts, they'd do it.

Secondly, I don't know how your first sentence relates to second. It is actually quite uncommon to succesfully ABX MP3 and AAC on high quality settings.

People must be crazy for utilizing FLAC - so much more space for insignificant differences.
FLAC works really well as an archiving format. It's not crazy to have a lossless archive of your music. If you then encode to MP3 for your portable or just play the FLAC is a decision that depends on other factors.

The fact of the matter is, the terminology I used would be understood and preferred on many other professional forums that deal more with playback like speakers, receivers, PMPs, etc...
Yes I know these words exists and that you think there is a common agreed understanding of them. I've just been unable to find it.

These differences are what is for me and SHOULD BE the foundation for choosing one format over another.
I'm saying the differences you describe do not exist in the world of perceptual audio encoders. Encoders don't make sound crispy or muddy, they introduce ringing or pre-echo. Just like video encoders don't make the video less "vivid", but they do introduce blocking or blurring.

There is not much point in discussing this in your way, because a. your terms are poorly defined, and b. you can't really apply them to audio compression.

[tebasuna51]

This is a "Audio encoding" forum and "audiophile" discussions never end with a common opinion.

In terms of efficiency (quality/bitrate) is commonly accepted than (> mean "better than"):
AAC > MP3 > AC3 > DTS
Of course is better DTS 1536 Kb/s than AAC 192 Kb/s, but the exact point of equivalence can't be easily found, and can be different for a kind of source than another.

Most users aren't "Golden ears" and/or don't have the necesary audio equipment to listen the differences between lossless audio and high bitrate lossy encodes.

I think everybody have enough opinions and info (tests) to take their personal option.

Thread closed