About recording audio...

[jFarley]

Interested in getting others takes on audio recording of radio receptions and what formats - not necessarily programs, but that may also be of interest - work best for you and why.

When I started recording audio for overnight receptions, I was recording directly into mp3 files, which I would later edit in Audacity for show archiving.  After a while, I switched over to recording in wav format for a few reasons, even though the file sizes were much bigger.  First of all - and I can't prove it because I have never simultaneously recorded in mp3 and wav and compared the two - I felt that I could hear very weak details better in the wav files (I tend to record at 11025).  Secondly, the wav files load a lot faster into Audacity, which can be an issue  when you have a 12 hour long file staring you in the face in the morning.  Thirdly, it seems to me that in the process of loading a compressed mp3 into an editor, selecting out a show, and then saving that show as an mp3 that the audio would be compressed twice.  True?  No?

Anybody see a diff in recordings between these formats?   I would appreciate any comments concerning your experiences or knowledge.

For recording live, I always use the mp3 format (generally 64kbps) since I can control the file start and stop manually, and the recording will not get any editing before archiving.  Audio recording is built into the SDR Console app I use, and really works well for me.  When recording in mp3, the audio seems to be streamed through the mp3 codec, and there is no save-to-file latency when the file is closed.   You click one button to close the current recording file and immediately open another;  great when one show tails another!

What do you use, and why?

[ChrisSmolinski]

mp3 is not a lossless compression format, so it is indeed possible that the compression is masking weaker signals, possibly by reducing the dynamic range of the audio?  I suspect mp3 was optimized for music, not static filled shortwave signals, which probably doesn't help.

I have learned firsthand that the mp3 encoding process often does mess up SSTV audio.

My workflow is probably different than most others. I typically record the I/Q stream from my SDR, then use a custom written playback app to go through it and demodulate signals of interest. The app displays a waterfall of each I/Q recording file, I can drag select a signal of interest and demodulate it into a WAVE file, that I later listen to. Or I can listen to it directly in the playback app.

I will usually convert a WAVE file to mp3 if I am sending it to an operator, to reduce the size of the file. The WAVE files produced by my playback app are 44.1 kHz 16 bit, so they run about 168 MB per half hour. A little too large to email :-)

[redhat]

Internally for quality reasons, I've always used PCM/WAV, for many of the reasons you mentioned.  Its much faster to edit, and you do not have to worry about generational losses and artifacting that occur with compressed formats.  That being said, when I record with one of the WinRadio's, it will create a wave file, I'll then go and edit the parts I want to save, and then archive it in MP3, usually at 128 mono.  It will then play on just about every device at my disposal, even the cars.

On a similar note, about 70 percent of the current X-FM music library is PCM.  When source material exists, the rest will be converted to PCM too.  Its much easier to process clean source material than lossy material, and its one of the reasons the station sounds as good as it does.  I also archive all my board feed recordings in WAV as well.

In regards to codecs screwing up weak signal details, that's what codecs do.  The algorithm in the codec, through a process known as "perceptual coding" has to determine if parts of the spectrum are audible or not.  In the presence of noise, the codec begins "guessing" what's perceptible and what's not.  That's one of the reasons cymbal crashes and white noise sounds so strange on MP3 and other lossy codecs.  Only so much bandwidth is available, and the codec has to figure out what to keep, and what to throw out.

+-RH

[rdla4]

You should be able to hear more detail in a WAV file, as others have mentioned. If disk space is an issue, you might try something higher than 64kBps for your MP3 files. I find that once above about 96kBps most (but not all) <some audio is simply NOT MP3 friendly> will sound fairly good. Higher rates create better fidelity. But, MP3 will never be perfect. I save mine in 128kBps MP3. At this rate the static and noise sound like the the static and noise i hear on the radio.

Raw recording, however is always in WAV format.

I have never tried to decode data from an MP3 file. But that makes sense....The perceptual coding throws away things the human ear does not always need. Software is not a human ear.  If you use VOIP for phone service things like fax machines and modems generally will have a hard time due to the lossy nature of VOIP. Same applies to MP3.

In commercial broadcast, (At least at my place)we generally will convert MP3 into a modified WAV format with 44.1/16, even though it may not sound tremendously better as a WAV. In most cases, it is easier to play a WAV file without  skips, pops, gurgles etc. We also generally try to get agencies etc to send files as WAV files if possible, either using You Send It or FTP. Low bit rate files are generally rejected due to quality.

When you smash audio (AKA Process) you can smash it better if its all there to start with.

[redhat]

In commercial broadcast, (At least at my place)we generally will convert MP3 into a modified WAV format with 44.1/16, even though it may not sound tremendously better as a WAV. In most cases, it is easier to play a WAV file without  skips, pops, gurgles etc. We also generally try to get agencies etc to send files as WAV files if possible, either using You Send It or FTP. Low bit rate files are generally rejected due to quality.

This is becoming more and more common.  In the past, most broadcast audio material was stored on disk as MP2.  With the advent of streaming, HD radio, et al, and the advent of cheaper large volume disk storage, lossless audio is becoming more common in commercial broadcast audio.  People finally learned that stacking codecs made for bad sound.  I was quite happy to retire a pair of Moseley DSP-6000's codecs with a PCM Starlink.  The difference was night and day.  It was finally possible to get the 'snark' out of the sound of the station.  We also started converting the music library to PCM by re-ripping all the music.  HD never arrived for us, but the stations sound much better in analog, and a quick spin around the dial reveals who's running compressed source material and who isn't.  It's pretty easy to spot when you know what to listen for.

+-RH

P.S. Another difficulty that arises with processing compressed audio is that it will make the compression artifacts more noticable.  The holes in the spectrum become deeper.  You can process clean audio much more with fewer sonic penalties than audio with holes in it.

Sorry for the thread hijack, Joe.

[ka1iic]

I am sure an mp3 does lose something but never tested it...

When I post to my domain I post MP3's but the original recording, generally, are in W64, or flac...   depends on what is open on the machine at the time.  These files are large because I use a rate of 48,000 and not the CD standard of 44,100... I save space by using a mono recording and not stereo.

I only use stereo when I am recording material with a 'time critical element' and then in that case the station being monitored is on the right channel and WWV is on the left channel...  I used that often when I was still active in the Army MARS program...  seldom today tho... in those days I used both computer and reel to reel tape... Teac and Sony reel to reels (7 & 1/2ips)... all VOX controlled.

The best recording via MP3 is to use the highest bit rate your sound set up has or perhaps using a variable bit rate... so I am told but I can't swear to this as true.

BTW I am using Slackware and or Ubuntu Studio OS on the el-cheapo HP quad core laptop with 4 terabytes of storage (external)... 16 gigs of RAM...

I do use the guitar audio filters to dig down into the dirt of shortwave signals ;-)

but that's just me....

[Sealord]

I use an old M-Audio Delta 66 card @ 48khz 24 bit running Audacity (in mono) at the same rate and then convert my edited wav files down to 128kbps mp3's.  Not a lot of science behind the settings other than trying to use a bit rate that provides decent audio playback for finding weak signals and doesn't eat tons of hard drive space.  Actually my headphones (AKG K701) have been the best tool for hearing the really weak stuff.

[Tom S]

I use an old program called RecAll which allows me to record constant audio.  I also record at 11025 and without compression.  The program allows me to split up the files into hour long recordings.  The hour long files are 76MB, which aren't too unwieldy to open or edit.  And then Audacity allows me to convert them to mp3's if I desire.

I've come a long way from using a small cassette recorder back when I first started listening, and then later using a VCR on ELP with a T-160 tape giving me 8 hours of record time.  Now I can record nonstop for a month and still have plenty of HD space.

[NJQA]

Compressing a MP3 300 times:

https://youtu.be/3nw_6BURm0k

[skeezix]

Depending on what is being recorded, if its something of sufficient signal strength on MW or SW, then its MP3 at 64 kbps at 44.1 kHz. I may record at a higher bitrate if I decide its necessary (nice strong, wide signal).

For weak signals, try to record the I/Q from the SDR when possible, then when finding the signal(s) of interest to be saved, convert to an MP3 at 44.1 kHz at some useful bitrate, no lower than 64 kbps. I also try to record & save the I/Q when I can as then can later switch to different audio formats with no loss. But those files are big, even at minimum bandwidth recorded.