|
|
Joined: May 2000
Posts: 8,364 Likes: 14
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: May 2000
Posts: 8,364 Likes: 14 |
I thought it was in convert.conf (but I did not look for it) you can specify conversion based on devices.
- Fiddler (IMHO)
If you think listening to music is fun, try learning an instrument.
HT: G61RSL, 4x5K, 5.5KHC, 2xSC6000, Transporter, HD621, Philips BDP7501 4K, Tivo, 65" LG OLED 4K C6
Computer: Foobar2000, 5K, Mojo, HiFiMan HE1000 V2
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
Hmm. -b24 seems to encode a 24-bit FLAC, but only with 16-bit precision...
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
Ok. The -s "steep filter" option increases the bandwidth from 95% (44.1 / 2 * 0.95 = 20.95kHz) to 99% (44.1 / 2 * 0.99 = 21.83kHz). Given that steep filters cause more ringing, I'm not going to use the -s option. Edit: The sox(1) man page says as much. In fact, there is an argument for using the -b option to specify a bandwidth less than 95%, i.e. a more gentle roll-off in the filter. A bandwidth of 90.7% is equivalent to 20kHz for 44.1kHz sampled audio. A resampler's band-width setting determines how much of the frequency content of the original signal (w.r.t. the original sample rate when up-sampling, or the new sample rate when down-sampling) is preserved during conversion. The term `pass-band' is used to refer to all frequencies up to the band-width point (e.g. for 44.1kHz sampling rate, and a resampling band-width of 95%, the pass-band represents frequencies from 0Hz (D.C.) to circa 21kHz). Increasing the resampler's band-width results in a slower conversion and can increase transient echo artefacts (and vice versa).
Last edited by VirusKiller; 2010-09-05 13:55.
|
|
|
|
|
Joined: Feb 2002
Posts: 8,811 Likes: 85
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: Feb 2002
Posts: 8,811 Likes: 85 |
VK,
Hope the cold gets better asap. Thanks for the run through of the codes.
I've just put the -vMa filter in. Immediate thought is that it is better. I wasn't clear that -M did anything positive at all.
Am mainly focussing on vocal and acoustic tracks so far.
Not sure about the -D. Dither is required isn't it ? Is this switching it off ? If we are to switch back on is it just a removal of the -D in the code?
What's the best way to run this code over a track to create a copy so they can be compared A vs B? (apologies...it's probably obvious)
cheers
Nick
Rgds,
Nick
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
Cheers Nick. -D turns off dither! -M sets minimum phase and gets rid of all pre-echos! See also my previous post which I've just edited. I'm going to see what manual dither options would be best - it is possible to specify noise-shaped dither. Still trying to figure out why, apparently, the 24-bit output FLAC is getting encoded at 16-bit precision: sox INFO sox: effects chain: input 44100Hz 2 channels 16 bits (multi)
sox INFO sox: effects chain: rate 88200Hz 2 channels 32 bits
sox INFO sox: effects chain: dither 88200Hz 2 channels 32 bits
sox INFO sox: effects chain: output 88200Hz 2 channels 16 bits (multi)
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
I thought it was in convert.conf (but I did not look for it) you can specify conversion based on devices. Yep. I need to add my Transporter's MAC address to my custom-convert.conf. Not done any listening yet, mind. -v 0.9 : I don't like this volume factor. SoX FAQ #5 states that both resampling and dither require headroom, but I don't see why this should be necessary when going from 16 to 24-bits. Hmm. -b24 seems to encode a 24-bit flac, but only with 16-bit precision... I think I may have stumbled on something... If the encoding is only being done with 16-bit precision, that could be why the guys on the SB forum got clipping.
|
|
|
|
|
Joined: Feb 2002
Posts: 8,811 Likes: 85
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: Feb 2002
Posts: 8,811 Likes: 85 |
...and hence why they went for 90% volume ?
Rgds,
Nick
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
Yes, that's a working hypothesis. I'm checking one of my 16-bit 24-bit flacs now.
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
No longer a working hypothesis I think. Here's some hexdump from a WAV of Dancing Queen which has supposedly been upsampled to a 24-bit file: 00d1f60 e000 00fa f792 7e00 00fa f744 4100 00fa
00d1f70 f73f 2b00 00fa f78e 2f00 00fa f82b 3b00
00d1f80 00fa f8f3 4800 00fa f9cc 5f00 00fa fabd
00d1f90 9000 00fa fbce e700 00fa fcee 6000 00fb
00d1fa0 fdf6 e700 00fb fecf 5700 00fc ff7c 8900
00d1fb0 00fc 0001 7200 00fc 0047 1d00 00fc 0035
00d1fc0 9200 00fb ffb9 d000 00fa fed4 e200 00f9
A 24-bit file uses three bytes per sample, but note that of the three bytes, one is always 00 00.
|
|
|
|
|
Joined: Feb 2002
Posts: 8,811 Likes: 85
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: Feb 2002
Posts: 8,811 Likes: 85 |
So funny...have just been converting Chiquitita Don't tell anyone 
Rgds,
Nick
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
So you were going alphabetically and I was going by track number?  On topic, I think I may have stumbled on a howler of a bug. I'm doing a bit more reading before I contact the SoX developer DL. Edit: Email sent to SoX DL.
Last edited by VirusKiller; 2010-09-05 15:27.
|
|
|
|
|
Joined: Apr 2004
Posts: 12,733 Likes: 88
Don't Panic!
|
|
Don't Panic!
Joined: Apr 2004
Posts: 12,733 Likes: 88 |
Hmm. It turns out that in this thread (pages 4 and 5), tingtong and michael have been aware of the SoX bug I discovered and even come up with a source code fix. I'm not sure it's been reported to the SoX developers though :rolleyes: I'm also trying to find out why they want to disable dithering, because I think that the effects chain with -D will have truncation errors when going from 32 bits to 16 bits: sox INFO sox: effects chain: input 44100Hz 2 channels 16 bits (multi)
sox INFO sox: effects chain: rate 88200Hz 2 channels 32 bits
sox INFO sox: effects chain: output 88200Hz 2 channels 16 bits (multi) I also want to find out exactly how applying dither prevents truncation. My gut feeling is that the following effects chain is the correct way to reduce the word size: calculations are done at 32-bit precision, dither is added (at 32-bit precision), and then the bottom 16 bits are then (presumably) chopped. Because dither has been added, this is (I think) ok. sox INFO sox: effects chain: input 44100Hz 2 channels 16 bits (multi)
sox INFO sox: effects chain: rate 88200Hz 2 channels 32 bits
sox INFO sox: effects chain: dither 88200Hz 2 channels 32 bits
sox INFO sox: effects chain: output 88200Hz 2 channels 16 bits (multi)
|
|
|
|
0 members (),
108
guests, and
3
robots. |
|
Key:
Admin,
Global Mod,
Mod
|
|
|
Forums18
Topics31,152
Posts296,952
Members5,503
| |
Most Online4,788
May 22nd, 2026
|
|
|
|
Google Search
|
Meridian Systems
|
Meridian Streaming
|
Restaurant
|
WiKi
|
Duncan's Meridian Info
|
Board Rules
|
Restaurant Rules
