16, 20, 24 bits-can you hear the difference?

[Park Seward]

This past weekend, I attended the 2002 Technology Retreat hosted by the Hollywood Post Alliance in Palm Springs. Mark Schubin put on the conference.

Dave Schnuelle of THX presented a paper that showed the results of experiments conducted at the Lucas Film Skywalker Sound facility. They had five buttons on the console that would playback a 5.1 soundtrack at 44.1, 20 24 bits as well as two Dolby encoded compressed soundtracks at different bit rates.

The listeners were all professional Lucas Film mixers and THX professionals. These people make their living with their ears.

They had material recorded specially for this test. It consisted of voice, music and a mix of both.

The used the Lucas Film Skywalker mixing theaters as well as other commercial mixing rooms.

Using the standard deviation, no difference of any significant measurement was found. Some 16 bit processing was actually found to be slightly better than the 24 bit mix.

All the standard deviations were overlapping.

While this test is not considered the final word, 24 bits was not considered to be any better than 16 bits in this test for motion picture sound. There was NO significant difference. Dave is going to do more tests.

I hope to post a web site son for you to get more information.

[Robotnik]

I attended a test a while back that compared a 16-bit/96k recording to a 24-bit/48k recording. Everybody in attendance agreed that the 16-bit/96k recording sounded better. The test was performed on a Sonic HD system and it was double-blind; the tester nor the listeners didn't know which was which at the start of the test. (Some people attended only for the free donuts & coffee...but, that's a different test altogether!)
Another test was for dithering. The conclusion: a properly dithering 16-bit system will sound far superior to a poorly dithering 24-bit system.
Did you know that the last Slayer album was done at 16/44.1? Can you hear the difference? It's rock and roll, baby!

[Doc]

Seems to me the humble CD player with decent converters SHOULD be with us for a while yet.
I say, "Focus on the filters, not the sample rate or the bit depth alone".
One of my specialties is mixing and encoding DD5.1 and to be honest, they sound pretty darn good when done properly. In fact, the final mix of a live performance I am writing will most likely be in 16 bit/48K DD5.1 format at 640K/s. Another encoding specifically for DVD release later will be at 448K/s.
A couple of hours of 5.1 channels on a single 650MB CD, with SMPTE code. Sounds good to me. [img]images/icons/cool.gif[/img]

[where02190]

I can hear the difference between 16 vs 24 bit, same as I hear it between 44.1k and 48k.

[where02190]

I can hear the difference between 16 vs 24 bit and 44.1 vs 48k. I'll stick to recording 24bit 48k.

[shaggy]

I think THX screwed up that test. That can't be right. Something must have truncated those bits.

(facts)

Digi's first Audiomedia cards were said to be 15 bit (1 bit lost for CPU noise)

Blackface Adats really do sound bad.

Sony Super Bitmapping DAT sounds great even at 22k.

I'd like to hear Motu's new Firewire box vs PTs HD at 24/88.2k just for comparison sake.

02 fleas

[jamlaw]

robotnik, how do you know that Slayer was done 16/44.1? That **** is off the rack!

[jamlaw]

whoa, this board self censors! Kick ass

[shaggy]

Careful.

Someone once hacked into the BBS censor and caused every t to spit out ****. And they thought it was us foul mouthed peons.

BTW- I typed four asterix above.

[Robotnik]

I've worked with the engineer/producer.