2.3.0 on Win10(64) In the UK
(this is a solo piano recording)
I’m not tech-savvy in audio nor do I have hi-fi ears so forgive me if I am naive or stupid and/or terminologically inexact. Or if this is the wrong forum to be in! I mostly just use Audacity to edit recordings into tracks.
Context: I occasionally record live concerts (classical) from BBC Radio to a Panasonic Blu-ray/HD disc recorder (DMR-BWT700), timer recordings as if they were TV programmes. This and the HDLED TV (Samsung UE32D-6530WK) are plumbed into my old analog stereo system via the headphone output from the TV. Playing the recording I could via the amp record it (in real time) to a Brennan JB7 hard disc music recorder which then compressed it to MP3. This produced files that sound good to me (and stand comparison to CDs ripped by the JB7).
I recently changed the (full) JB7 for a Brennan B2 which compresses to FLAC. The few recordings I have made since seem to my far from expert eye all to have a certain flattening of the negative signals below the 0.0 line on both channels. I’ve no idea what the waveform is actually measuring but my impression is that the JB7 MP3 files displayed on average roughly equal + or - peaks. I’m not sure my ears can hear anything wrong but does the waveform suggest something isn’t quite right? Or is it just a peculiarity of the types of music recorded (orchestral, string quartets, piano - solo & with strings)?
The B2 records in .WAV format before compressing. The .WAV files have the same symptoms so presumably it’s not the compressing.
I have taken an optical feed straight from the HD recorder and run this into the B2 through a basic DAC (FiiO D03K) and again the results are the same, so presumably it’s not the (no longer young) AV kit failing. So probably the B2’s own recording system then? (It’s a Raspberry Pi basically.)
I have also copied a recording to a DVD and extracted the audio file (AC3?) from the VOB file using the Pazera Extractor and output it FLAC. This seems to produce a more balanced waveform but also a much smaller one — sounds OK but needs louder amp setting. So they don’t sit well with the existing recordings on the B2 (all my CDs, LPs & cassettes!). Is there a better extractor? Time also to look at some of the many other Audacity functions, maybe?
Thanks! BJ
The few recordings I have made since seem to my far from expert eye all to have a certain flattening of the negative signals below the 0.0 line on both channels.
No that’s not normal. There is a sister problem called DC Offset where the blue waves do not settle in the middle (0.0) when the performance goes silent. Do you have that, too? Wave on the left.
Koz
Well… Your image didn’t show up… [u]How to attach a file[/u].
FLAC is lossless compression so that’s not the problem.
Some musical instruments do generate asymmetrical waveforms. I don’t know if it’s normal for piano* but I wouldn’t be that surprised since it is a percussion instrument and the hammer strikes in one direction. Commercial studio recordings go through a LOT of processing, and usually multiple instruments & voices are mixed so you don’t often see it.
If you are getting [u]clipping[/u] (squared-off waves) at less than +/- 1.0 that’s usually a hardware problem on the “analog side”. You can work-around that by reducing your recording level. (You can boost the volume digitally later.)
If you are getting a [u]DC offset[/u] where everything is shifted up or down, including silence, that’s also a problem on the hardware-side. The Normalize effect has an option for removing DC offset.
*** P.S.**
A quick Google says “loud piano” can be assymetrical.
VOB file using the Pazera Extractor and output it FLAC. This seems to produce a more balanced waveform but also a much smaller one — sounds OK but needs louder amp setting. So they don’t sit well with the existing recordings on the B2 (all my CDs, LPs & cassettes!).
You can use the Amplify effect. There is a loudness standard for movies/DVDs and it’s nominally at a lower level to leave headroom for extra-loud dynamics & effects.
Is there a better extractor? Time also to look at some of the many other Audacity functions, maybe
I don’t know anything abut the Pazera Extractor.
Audacity with the optional [u]FFmpeg Import/Export Library[/u] can open/decode AC3 from an audio/video file. However… It won’t read an encrypted commercial DVD and If there are multiple soundtracks (English & French or Stereo & 5.1 surround, etc.) you can’t choose. And, the audio/video on a DVD is “split” into 1GB VOB files and if you’ve got a concert DVD those splits will likely be in the middle of a song.
Thank you gentlemen. I hope the image works this time — inserted previously as instructed but forum does not seem to like Opera browser…
No sign that ‘DC offset’ is an issue. Or at least that I can see.
Would ‘clipping’ operate equally at + and - ends of the range, rather than just the negative end?
‘Amplify’ sounds distinctly useful… The Pazera extractor (http://www.pazera-software.com/products/audio-extractor/) has a ‘volume’ slider on the file conversion setup but no description that I could find of what it does (bar some contradictory notes in the comments!).
The recordings are not derived from commercial DVDs, they’re just copies (somewhat tortuously obtained) of my own recordings of digital transmissions (like taping a TV prog). As I understand it, Audacity is not able to read the audio track direct from a VOB file so some sort of extractor is needed in between. The VOB files are quite small in comparison with full AV recordings (typically <500Mb for an hour or so): because they are radio transmissions there’s very little actual video content.
If you can now see the image I’d be most grateful for your views.
Thanks again,
BJ
That looks normal! I’m pretty sure it’s the nature of the piano.
Would ‘clipping’ operate equally at + and - ends of the range, rather than just the negative end?
“Normally”, if you over-drive the analog-to-digital converter you get clipping at exactly +1.0 and -1.0 (which is 0dBFS). For example, with 16-bits you can “count” from −32,768 to 32,767*. You can’t go any higher without more bits so you get clipping. 0dBFS (zero decibels full scale) is defined as the maximum you can go with a given number of bits. And, everything is automatically scaled so a 0dB 8-bit file plays-back at the same loudness as a 0dB 24-bit file. Audacity uses floating-point for internal processing. 0dB in floating-point is 1.0. You can go way-way over 0dB and there is virtually no upper or lower limit.
Analog-to-digital converters, digital-to-analog converters, regular (integer) WAV files, and CDs are hard-limited to 0dB for the same-simple-mathematical reason and if you “try” to go over 0dB you’ll get clipping.
But some cheap computer soundcards will clip asymmetrically at lower levels. Especially the microphone input.
Audacity is not able to read the audio track direct from a VOB file so some sort of extractor is needed in between.
It should work if you install FFmpeg.
\
- The positive & negative limits are different by one count because one bit is used for the +/- sign, and the way signed binary integers are stored (two’s compliment) there is no negative-zero.
Thank you — I think I understand that!
In amongst my playing about I did load FFmpeg and thought I’d tried & failed to open a VOB file but it must have been before: so that most usefully cuts out a step. Taking the file straight from the AV recorder & converting it seems the way to go (‘Amplify’ will be my friend) rather than solemnly converting to analog and back again.
It’d still be reassuring to know the B2 line-in recording isn’t misbehaving: I’ll do some parallel recordings by both methods, and of different types of instrument groupings, and check if I can see or hear any differences…
Thanks again, BJ