Taperssection.com
Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: dmonterisi on May 06, 2005, 01:24:04 PM
-
so i was working with some of dave friedman's 24/96 recordings from panic this spring. they were recorded 4022>v2>2k>722 @ 24/96. the files are listed as 24/96 in the properties, etc and wavelab recognizes them as such. however, i opened the bit meter in wavelab and it seems to indicate these are only 20 bit files (and 1 of them seems to be 16bit). is this bit meter thing in wavelab accurate? the floor of the meter is at 20bits and it jumps up to 2, 1 and 0 throughout the music. the levels typically peak at -6, i don't know if he needed to open this stuff up a little more or what. not sure if it matters or not, but curious if this meter thing is accurate. thanks
-
I have always thought they were accurate, and my limited testing with it seemed to indicate so... ex: when I first got my MiniMe I would record in waveLab as a 24 bit/48 KHZ file... get home, open up the bit meter and what do you know? 16 bits were lit up... WTF? I then learned that I had to go into the ASIO driver UI of the WaveLab -> Preferences -> Sound Card and select the 24 bit radio button. It was set at 16bit my first few outings, thus only 16 bits being lit up in the Bit Meter. When I switched the ASIO UI to the 24 bit radio button I would get 24 bits lit up. Don't know if this helps or not, but it was at least consistent for me and did indeed indicate the appropriate bits.
In your case, you say your peaks are around -6. As each bit can store 6db per bit, it would make sense that when you are peaking at -6 that the 0th bit is lit up. Since you say that you observed 0th, 1st and 2nd bits fluctuating I'm going to guess that the lowest level sounds are around 18 db below over. Does that make sense? if so, then I'd say your bit meter is accurate and the 722 only gave you a 20bit file. oops.
Rusty
-
I only used to the 722 Digital In from the 2k and the 722 said it was writing 24/96 digital in so im not sure what the problem is. Windows says the files are 24/96 even Wavelab says there are.
-
it's strange...here's what adobe audition says about one of the tracks in the statistics summary:
Left Right
Min Sample Value: -26702.56 -26281.94
Max Sample Value: 28136.69 26788.94
Peak Amplitude: -1.32 dB -1.75 dB
Possibly Clipped: 0 0
DC Offset: 0 0
Minimum RMS Power: -36.37 dB -37.56 dB
Maximum RMS Power: -10.98 dB -11.9 dB
Average RMS Power: -19.67 dB -19.86 dB
Total RMS Power: -19.24 dB -19.46 dB
Actual Bit Depth: 20 Bits 20 Bits
Using RMS Window of 50 ms
but here's what shntool says about that file:
-------------------------------------------------------------------------------
file name: wsp2005-04-19track15.wav
handled by: wav format module
length: 14:00.581
WAVE format: 0x0001 (Microsoft PCM)
channels: 2
bits/sample: 24
samples/sec: 96000
average bytes/sec: 576000
rate (calculated): 576000
block align: 6
header size: 12288 bytes
data size: 484174866 bytes
chunk size: 484187146 bytes
total size (chunk size + 8): 484187154 bytes
actual file size: 484187154
file is compressed: no
compression ratio: 1.0000
CD-quality properties:
CD quality: no
cut on sector boundary: n/a
sector misalignment: n/a
long enough to be burned: n/a
WAVE properties:
non-canonical header: yes
extra RIFF chunks: no
Possible problems:
file contains ID3v2 tag: no
data chunk block-aligned: yes
inconsistent header: no
file probably truncated: no
junk appended to file: no
odd data size has pad byte: n/a
-
That ain't no 24 bit signal. When the levels are lower, it's the first few bits that don't light up. Somewhere along the way this got chopped down to 20 (or it was only 20 bit to begin with).
If you want to check it another way, pop the file into CEP/Audition and go Analyze, Statistics and check 'actual bit depth' at the bottom.
whoop...we were typing at the same time
-
The couple times I used it I found it to be accurate, the LED's should go all the way down to 24 bit if the file is 24 bits, kinda funny what you are seeing. Not sure why that would happen. Do you have any other 24 bit sourced material you can run your bit meter with? Maybe pull something off archive to check.
Take care,
Ben
-
it's strange...here's what adobe audition says about one of the tracks in the statistics summary...<snip>
That's just telling you about the file format - for every sample the last 4 bits are just padded with zeroes.
-
i just opened another 24bit source i had on my HD and the meter reads it as full 24bit in wavelab. i don't know where the error is here. Dave says the 722 was reading as saying it was recording 24/96 on the digital input. strange, eh?
-
blame it on the ad2k+
-
Does the AD2K have an option to dither and output 20 bits? if so, you could record a "24 bit" file on the 722 via digi in, but there would only be 20 bits of actual data. I'm not too familiar with the ad2k, but that's the only thing I can think of.
-
well, dave said the 2k doesn't have an option to output 20bits, but i've never run it so i have no idea of how it works. is there a function in the 722 that will dither an incoming digi-signal? somewhere buried in that 200 layer menu??
-
you must have had the 2k set for 20 bit
output mode (clockwise from the 7 o' clock positions:
16TPDF
16NN1
16NN2
16NN3
16NS1
16NS2
16NS3
20TPDF
20NN1
20NS1
20NS2
24 bit
i'm guessing you switched to 24 bit all the way down on the right, and actually were on the next to last switch 20NS2
-
http://www.sonicsense.com/images/2kpiclg.jpg
-
Nope, it was on 24 Bit the whole time, just checked the 2k and it is still on 24 bit.
-
it defintiely can't turn 1 more time?
-
Nope, it was on 24 Bit the whole time, just checked the 2k and it is still on 24 bit.
any chance you have a 24/96 capable soundcard in your computer (and wavelab?) at home? You could monitor the digi out of your 2K at the current settings and see real-time what it is outputting.
Just a thought.
PS: Nice rig!
-
I guess it could have slid over but the 722 was locked in at 24/96. Who knows, lol
-
My guess is that it's user error related. :P
-
>>I guess it could have slid over but the 722 was locked in at 24/96. Who knows, lol<<
the 722 doesn't have a 20 bit mode. i think it'll either say 24 or 16, depending on what you choose
-
so what's the best way to seed the 20bit sources? is there a way to truncate the padded zeros? there's no reason to seed it as 24bit and the extra size when it is 20 bit, but i don't know the best way to get it done. thanks.
-
so what's the best way to seed the 20bit sources? is there a way to truncate the padded zeros? there's no reason to seed it as 24bit and the extra size when it is 20 bit, but i don't know the best way to get it done. thanks.
just dither down to 16. don't truncate ever!
-
so what's the best way to seed the 20bit sources? is there a way to truncate the padded zeros? there's no reason to seed it as 24bit and the extra size when it is 20 bit, but i don't know the best way to get it done. thanks.
just dither down to 16. don't truncate ever!
i did that for the 16bit source, i'm talking about seeding the 20bit source which is masquerading as a 24bit source.
-
ok.
dither it to 20bits
-
but then isn't the dither algorithm just trying to approximate the sound of the padded 24bit source? there's no way to just lop off the extra zero's??
-
if you dither down to 20 bits i'm guessing that's what will happen anyways since wavelab will see the extra bits as all zeros. maybe not though. you can just change the bit depth without adding a dither. that will truncate the waveform.
try both and analyze the waveforms. see what happens.
edit: spelling
-
I would just leave it as a 24 bit source with 20 bit resolution. The standards are either 16 or 24 - some folks may have difficulty with 20 bit. The size difference is minor and flac only works on 16 or 24 bit.
-
I would just leave it as a 24 bit source with 20 bit resolution. The standards are either 16 or 24 - some folks may have difficulty with 20 bit. The size difference is minor and flac only works on 16 or 24 bit.
a very good point to keep in mind!
i was just thinking not too long after this original discussion that it would be pretty worthless to have a 20bit copy other than for wave data archival purposes.
-
so what did you end up doing?
i would just keep the dithered down 16 bit copy and call it a day ;)
-
nothing yet...i just got back from delaware/philly and haven't done anything with it yet. i didn't realize flac didn't work with 20bit files, i was hoping to cut like 25% of the file size off by seeding 20bit sources, but it looks like i'll leave them as 24bit and inform people they are really only 20bit files.