This content has been marked as final. Show 34 replies
Do you need to? No one can hear that high anyway? And if you're not editing in a 96 KHz project with 96 KHz audio to begin with, that added sampling won't make things any better.
The thing is that, I have to make the videos for a live performance, and the Pro-Tools guy told me that he is giving me the audio in 96khz, so he said that I should keep it in 96khz.....because I have to output everything to a DVD.
When no video was used in their live shows, they used to output the audio through an M-box at 96khz, and he says it's better sound quality than 48khz......
what's your advice?
>Do you need to? No one can hear that high anyway?
Same as with HD video: The human eye simply can't handle the enormous amount of pixels. ;) :p
The human eye can see more than HD offers. That's one reason why film, with higher resolution still, looks better.
A 96KHz .wav file listened to by an audiophile through a set of $20,000 speakers in a professional recording studio just might sound better than a 48KHz file. But I'm guessing your average consumer won't notice by the time it gets to DVD.
Besides, the DVD spec only allows up to 48 KHz for 5.1 Dolby audio. 96 KHz requires a stereo PCM signal or DTS, which will require a separate encoder.
Regardless of the seemingly continual debate about the point of using 96KHz audio (I know it sounds far superior to 48KHz, for a number of reasons. The actual frequency it can reproduce is not the main reason for using 96KHz either - if you're doing Opera for dogs & bats then a top end of 48KHz is audible, but not for people. The reason is that it is far easier to manufacture a set of converters that sound good at 96KHz than it is at 48KHz, because of the aliasing issues amongst other things. So if your source is at 24/96, then this is a good thing).
Never let anybody tell you it is not worth the effort - it certainly is. Strange the people who say it's pointless. I wonder if they believe 10 bit video to be pointless. Almost certainly not. But this is all beside the point, as you cannot create a Dolby Digital stream with 96KHz audio. you are limited to 48KHz. So you will need to resample down to 48KHz using the best quality SRC available - R8Brain Pro. See www.voxengo.com for the download. Use Minimum Phase - not linear phase - and Ultra Steep Slope. This will give you the equivalent of running through seriously good hardware converters.
Also, see http://src.infinitewave.ca/ for detailed comparisons between SRC. R8Brain pro stands out head & shoulders above the rest.
Ignore anybody who tells you you cannot hear the difference. What you need though is an amp that can reproduce it, and decent monitors to match. but the difference is there, and it is very real.
You also have other problems.
24/96, although supported in the DVD-Video specifications, is optionally supported. NO player is required to decode it as is, and some will pass it intact, others will SRC to 48KHz (often in a very ugly manner), more will truncate the bit depth down to 16 bits, and yet others will both truncate & src as well.
To ensure compliance with all players, you need to remain at 16/48 for LPCM, or 24/48 should be okay in 95% of players.
If you want to remain at 24/96, then you must use LPCM or DTS. LPCM will carry a huge data rate implication for the video. You have a total of 9.8 MB/sec available, and a 24/96 stream will take up 4.7Mb/sec of this, leaving you with 5Mb/sec for your video component. If you want surround, then DTS is your sole option as DVD-Video can accept a DTS 96/24 stream with no problems. This is at a bitrate of either 1509Kb/sec (1.5Mb/sec).
Downside is that this cannot be the sole stream available to the user, or your DVD will be out of spec as DTS is also only optional.
You will need either the DTS-PSE encoder (obsolete) or the DTS-HD SAS (this will do the job) or the DTS-HD MAS.
I recommend this route:
Audio Stream #1 - 5.1 Dolby Digital from a 24/48 SRC.
Audio Stream #2 - 5.1 DTS 96/24 stream
Total load - 2Mb/sec of your 9.8, leaving you up to 7 for the video, unless using a replicated disc, in which case - depending on length of program & details, you can go all the way up to 9.8 - just do not try this on a written disc.
This is a music title.
Therefore the Audio must get the priority. You are correct to insist on this, and so is the audio mixer.
I author DVD-A/V titles all the time using both High Resolution Lossless components as well as Video options making the disc compliant with every DVD player sold - all 500,000,000 of them. It can be done.
Good Luck in your project.
>What you need though is an amp that can reproduce it, and decent monitors to match.
That is the point I was making. Most folks do not have the kind of hardware (or listening skills) required to notice the difference.
Excellent explaination Niel !
Just to clarify a couple of things (about what Neil wrote)
48 or 96 KHz are the sampling frequencies, and are NOT the upper frequency that can be reproduced. Theoretically the maximum frequency that can be sampled will be the half of the sampling frequency. Practically it will be a bit lower than the half because the filters that filter out the higher frequencies don't have indefinite steepness.
So, in order to avoid aliasing issues, the low pass filters will start "their job" at a couple of KHz's before the half of the sampling frequency. Practically, it will very seldom be aliasing issues because (at least in decent equipment) the low pass filters are good enough.
What Neil meant about this: "because of the aliasing issues", I'm not quite sure about, but he's free to explain if he wants.
-The human ear is capable to hear 20Hz to 20KHz (normal best case).
The low pass filter introduces a variable phase shift dependent on frequency. If that's a problem for the human ear or not is not up to me to say. But, 96 KHz sampling will make it easier to make a good filter with less phase shift in the hearable range.
For a 48KHz sampling frequency, a 20KHz tone (let's say a sine wave) will only have averagely 2,4 samples per cycle. To reconstruct a sine wave cycle with 2.4 samples can be somewhat iffy (subject for discussions), but again, not up to me to decide if the average Joe can hear that. Anyway, the 96KHz sampling will double the amount of information when it comes to number of samples per cycle.
16 or 24 bits per sample: To say it this way, every bit more will double the information (of level accuracy).
Oh, finally, for those who mean that one should read manuals and not ask questions about basic things, please read this very basic information:
There has actually never been one single scientific or double blind test proving anybody can hear the difference between 48khz and 96khz, and yes, even on extremely, extremely expensive setups. it is all them believing they are hearing better. when they are put in an actual test, blindfolded, and do not know which is which, they have always had more than around an 70% failure rate.
and btw, trying to compare this to the eye seeing hd is more than rediculous
Anybody ever take Calculus? Remember the integral? How it was easier to approximate the area under a curve by using progressively thinner rectangles lined up side-by-side under the curve?
A higher sampling frequency allows thinner rectangles to better simulate/reproduce the analog audio curve.
and if you cant tell the difference, then the reason for wasting storage and money is?
96khz is the same as wooden volume knobs, $1000 3 foot cables, marking the edge of a cd green, etc. show me a scientifically proven case of somebody being able to tell the difference, and Ill believe it. but so far all im hearing is "the numbers are better so it must be better, sure, theres no actual proof what so ever, but just look at the numbers!"
oh also, I don't know where you found your "pro tools guy", but any experienced audio engineer or mastering engineer will tell you there is absolutely not one reason to go about 48khz 24 bit.
>There has actually never been one single scientific or double blind test proving anybody can hear the difference between 48khz and 96khz
Although I do agree to that a blind test would be the correct thing to do, but, how on earth do you know that it never has been one?
>any experienced audio engineer or mastering engineer will tell you there is absolutely not one reason to go about 48khz 24 bit
Do you know that for sure, I mean, do you know them ALL?
I post and browse extensively on professional studio and recording engineer forums. (audio is my second passion, next to video) I can get you contact information for several highly successful recording engineers and mastering engineers if you'd like to ask them yourself. on every one of these forums people have asked or looked for any reason or proof to use 96khz, never finding any. no tests, no anything. it ussually gets brought up in the same threads that ask about $1000 audio cables, or how nice wooden volume knobs make your equipment sound. all of these threads turn into a laugh factory of people poking fun at audiophiles and their ridiculous theories.
I don't mean to sound harsh or anything. it just bugs me when people purposely mislead and waste peoples time, and I've seen this argued a million times over.
I actually agree with you when it comes to "audio is a religion" (more believes than facts), but, using words like everybody or nobody when there is no facts that can prove that is wrong.
Theory and numbers are exact! Meanings are not! (not less important though, but should not be mixed)
>...but any experienced audio engineer or mastering engineer will tell you there is absolutely not one reason to go about 48khz 24 bit.
One experienced audio engineer told you that there is - in post #5 . :)
...and didn't provide any proof or tests to back him up. Or credentials. if you'd like I can get you in contact with
Like I said, all you are doing is listing off numbers. sorry, but I don't hear numbers. do a blindfolded test or otherwise, and get back to me.
One word for you: harmonics. Ask your engineer buddies about that one.
>...and didn't provide any proof or tests to back him up.
Well, you did list two, and how many else are out there?
Are you going to answer my questions? (ALL of them, and how on earth...)
Jeff had an important question for the engineers: harmonics
Oh, my only proof to back me up is nothing more than being an electronic engineer that have been working with sound measuring/analyzing equipment for more than 18 years. And yes, calculus I have taken.
Point is, shall we discuss believes or facts?
>And yes, calculus I have taken.
>Point is, shall we discuss believes or facts?
I would refine that down even further to "beliefs or practicalities". Even if it were a proven "fact" that 96KHz audio sounded better on the right equipment to the well trained ear, is that the intended audience for memevertical's final DVD? If not, if average Joes will be any significant part of the intended audience, then it a useless "fact" at best.
Thats true.....my audience are just people at the concert, and I doubt that they can tell the difference....after all, after reading all your opinions, I dont think the difference between 48 and 96 is not that big (if there is any).....
thanks a lot for all the help.....
Let me just ask a little quick question.....what's the difference between AC3 and DTS-WAV files?
>Let me just ask a little quick question.....what's the difference between AC3 and DTS-WAV files?
Maybe the question was quick, but the answer is not.
Ok I just finished reading it.....I think I get the point.....what I saw is that DTS is much better than a AC3.....
I'm doing a 5.1 Dolby DVD, which one would be better? DTS-WAV? AC3? Both?
>I'm doing a 5.1 Dolby DVD, which one would be better? DTS-WAV? AC3? Both?
Do you have surround sound available when you watch Hollywood DVDs at home? Are you happy with the quality?
Most of those use DD 5.1 surround. There are a good number of DTS titles available as well. I generally prefer DTS, but not always.
The same audience that won't be able to tell the difference between 96 kHz audio and 48kHz audio also won't be able to tell the difference between surround implementations. Not to mention that support for DD in the amplification section is more ubiquitous than DTS (or at least it used to be).
Go with whatever format is easiest and most cost-effective for you to implement.
Dolby can be added to Premiere for about $300 with the SureCode plug-in. The cheapest DTS encoder I could find is about $1,200 as a stand alone application. That may affect your decision.
Regarding "double blind" tests, and "golden ear" mastering engineers...They absolutely DO exist!
Many, many years ago (before the earths crust had cooled), professional digital audio equipment was just begining to become available...The CD hadn't even been invented yet at the time this happened...
Anyway, I was doing a demo of an a first generation digital device (The Ampex ADD-1), that was intended to replace the "pre-read" head on analog audio mastering machines...The pre-read head is used to drive the mechanics of the disc cutting lathe to allow for proper groove nesting...
We had about 20 or 30 mastering enginers and members of the Detroit AES, and the "Southeastern Michigan Woofer & Tweeter Society" in attendance...The idea was to set the delay for no delay at all, passing the audio only through the A/D converters and codecs, and do a double blind A/B comparison to see if anyone could "catch" the converters.
The double blind system was set-up with a computer that randomly switched between sources, and kept a log of the switches...No one even knew which source was A and which source ws B.
The group of engineers in listenening, where to select which source was which on 30 some odd tests.
EVERY ONE OF THEM NAILED IT CORRECTLY...EVERY TIME!
I was stunned, as was the marketing manager for the product...Obviously, it was back to the drawing board for this device, but from that day forward, I never doubt mastering engineers when they say they hear something...There really are "golden ears" out there...
There are also "golden eyes" out there as well...Now, I can tell when the phase is a little off on video, having been in the business for over 35 years, but I had a boss (actually the owner of the TV station I was working at) that literally could visually tell the differnce of 1 degreee of color phase error...I tested him on it, and he was dead accurate to one degree...
Anyway, I thought I'd add this bit of trivia to the thread...
Hmm I see....ok, but let's forget about budget for a second. I have a Portable Media Center, with 6 RCA Analog outputs for the Dolby Surround. Can the device work with DTS-Wav files, or only with AC3? The device says it supports AC3 and "DTS Pass through".....
thanks for everything.....
>Can the device work with DTS-Wav files, or only with AC3? The device says it supports AC3 and "DTS Pass through".....
When you read the link I put up, you probably saw that is was a patented CODEC involved. My wild guess (without knowing details about the DTS codec), is that when it says "DTS Pass through" it is that the producer of the Media Center has tried to get the most less expencive way out of this. It's probably easier to make a "Pass through" for less money than it is to make a full support (probably meaning also converting and exporting) of the DTS format.
Do some tests and figure it out for yourself.
OOOOhh! Sound stuff! It comes down to the end-user. Back in the day, I used to mix live sound. One of the solo acts I worked for spent some time in the studio for her first tape (I told you it was way back - but after eight-tracks.) I helped her tweak the mix to balance the guitars and vocals (always a fight in the midrange for female voices). When we were happy, we ran off a dupe. We then went out to the parking lot to see how it sounded through the car stereo speakers. She wanted to hear what the audio was like for most of her listeners. She was absolutely correct in doing so. If it was good there...
Ok, I'll defenetly do my tests......
I'm sorry for all the questions, but I just have a final one. When I'm in Premiere, I have my 6 Mono Audio Channels, and in the MASTER channel I see 6 level meters. I've managed to have 5 channels sound separately, I mean, in the Master Meter, you can see each one play separately, but I cant manage to have a 6th channel sound separate from the others, even when the 6th level meter is THERE all by its self, I cant get a 6th audio channel to be separate.......or is that only for the sub-woofer?
All of this because I need each channel to be TOTALLY independent, and If i can have 6 independent channels.....it would be better.....
The 6th is the ".1" in 5.1, used for subwoofers.
More precisely the ".1" for Low Frequency Effects. It is a limited frequency bandwidth channel, covering only the 3-120 Hz range.