This site is intended to become a complete archive of the audio messages of the late Hans R. Waldvogel.  Therefore, we present not only the edited recordings, but make the original media and project data available with the hope that over time, new and more effective noise reduction and audio enhancement techniques will become available.  You are invited to download the original files if you wish, and attempt to create a clearer representation of these audio messages.

Perhaps, in time to come, we will have the opportunity to document on this site other aspects of Hans Waldvogel’s life, and the ministries with which he was associated.

These recordings were of course originally on reel-to-reel tapes.  For many years, Hans Waldvogel’s nephew, Edwin H. Waldvogel, recorded his messages and produced 30-minute tapes for radio broadcast.  These master tapes, as well as tapes from regular church services, and talks at the faith home, accumulated over the years in Bro. Edwin’s basement, until he began a project to transfer the reel-to-reel recordings to cassette tape, evidently discarding the originals.

Many thanks to Paul Munsinger for tracking down these cassette originals, and granting the rights for them to freely appear publicly here.  And please note that the rights to these recordings are reserved insofar as while they may be freely copied and distributed, they may not be sold.

Thanks also to Gudrun Reichle for her invaluable contribution in providing German transcriptions, translations, research, and insight.

The goal has been to preserve as accurately as possible the spirit of these messages, and to make the words as clearly audible as possible.  Absolutely nothing has been removed for doctrinal, cultural, political, or reasons related to any other editorial sensitivity.

Some small edits have been made to remove what might be called distractions, such as long pauses, coughs, stutters, and background noises.  In many cases, it would be more distracting to remove the section containing a noise than to leave it, because the natural timing or inflection is altered, and so many of these rough spots can be found in the final processed recordings.  The most radical editing that has been done is occasionally to smooth over a misstatement.  For example, if the original says, “…he had to allow Hezekiah, …or Sennacherib… to blaspheme Jehovah…”, the edited version may have, “…he had to allow Sennacherib to blaspheme Jehovah…” 

A summary of the kinds of edits that have been made is presented in the project notes section for each recording, and of course, the original unedited file is always available as well.

The processed audio is presented in MP3 format.  The goal is to create audio suitable not only for internet streaming, but for download, distribution on CD, listening on higher quality equipment, etc.  For this reason, the processed audio uses bitrates somewhat higher than would normally be expected for internet distribution, and somewhat lower than would normally be expected for the other forms of distribution. 

So, then, the tuning of the data compression was done so as to be just at the threshold of perceptible compression artifacts.

Data compression was done with the LAME MP3 encoder, version 3.97 {->http://lame.sourceforge.net/}.  Most vocal material was compressed in this way:

lame -m m -V 3 --resample 11025

That is, monaural, vocal encoding of quality 3, (0 is the highest data rate, 9 is the lowest), and resampled to 11025Hz.  If a particular vocal recording was of exceptionally high quality it may be resampled to 22050 Hz instead.

The music on these recordings often has little more dynamic range than the voice recordings, however, they were processed using

lame --preset 112

for stereo music files (112kbps 44100Hz), and

lame -m m --preset 96

for monaural music files (96kpbs 44100Hz).

The need for multiple distribution targets is also the reason why each track of processed audio contains a few seconds of trailing silence.  This way, the tracks can be burned to an audio CD without needing to add short tracks of silence.

MP3 audio files, even at 256kbps, would not be considered sufficient archive media by any respectable archivist.  Yet, the generally poor quality of the originals, and the exceptionally low signal to noise ratio, means that realistically, digitizing at greater bit depths, sample rates, and with less data compression would preserve nothing beyond what is present in the noisy originals.

This, along with the desire to make this entire project publicly available at a manageable data size, means that the original masters are indeed MP3 files at 256kbps.

Processing of the audio was done using an older version of Sony Vegas, namely 5.0b (Build 160) from Sony Pictures Digital, Inc.  The Windows project files, having the extension “.veg” are included in the archive.  In addition, Vegas can output its project format to XML, which perhaps can be more easily interpreted in the event it becomes difficult to find an application that can open these archived .veg files directly.

In many cases, intermediate rendered files were generated from the originals, for example if click/pop removal was run, or a pitch shift was needed.  The Vegas project files, then, often reference these intermediate media files, rather than the original digitized masters.  It would consume far too much space to make these intermediate files available as well, so instead, detailed instructions are provided in each case for the creation of these intermediate renders.  Of course, if you want to work directly with the original audio, you are always free to ignore the Vegas project files altogether, and start from scratch using your own methods.

A rough target for volume level for the music was between -9 and -6dB(fs), with peaks above that, and -6 to -3dB(fs) for voice.  No hard limiting was done, so levels up to 0 dB(fs) can and do occur.

Audio Compression (that is, the processing of the dynamic range of the audio, as opposed to the data compression of the audio using MP3) is a topic that has entered somewhat into the public consciousness of late.  Be assured, that many experiments were done to see if it was feasible to reduce or eliminate the use of audio compression.  Certainly, for some talks, it could be done, but the long and the short of it is that considering the quality and character of the source material, as well as the ultimate goal—to make the words as clearly audible as possible—the use of a great deal of compression was unavoidable.

Our hope is that when those who were familiar with the previously available cassette duplicates hear these recordings, they will be pleasantly surprised how much more intelligible the words are, as well as relieved that constant volume adjustments are no longer necessary.

Experiments were done with various noise reduction techniques.  Unfortunately, most of the techniques that operate in the frequency domain, or are based off of a captured noise print, were found either to introduce very noticeable artifacts, or reduce the noise by such a small amount as not to be worthwhile.  Again, the masters are so noisy that these methods, which are so helpful in other situations, cannot successfully be employed here.

So then, most often simple EQ was used to mask 60Hz hum (with its harmonics), tape hiss, etc.  This is one area in which substantial improvement could be made in the future over what has been done here, although on many recordings it seems the higher frequencies contain nothing at all of the original signal, but only tape hiss and noise.

Starting with message 7A, some successful experiments using hiss reduction in Adobe Audition 3.0 were done.  The standard setting produced too many spectral artifacts, but the “Light Hiss Reduction” preset setting (noise floor 5dB, FFT 4096 points, precision 5, transition 5dB, decay 50%, 15dB hiss reduction) dealt with the hiss in a fairly clean and effective way.  The decision to use this hiss reduction was difficult, because when using it, some of the subtler cues of the noise presence of the room seemed to be eliminated, and the voice sounded just a shade more muffled (if that is possible given how muffled some of the originals are).  Nevertheless, there is no question that the level of hiss on many of the recordings—especially the sermons where audio compression raises the level of hiss—counts as a distraction that ought to be dealt with.

In some cases, noise that was more localized was removed, or rather attenuated, using the spectral frequency editing tool of Adobe Audition 3.0.

Most clicks, crackles, and pops that sound like static discharge can be effectively removed by using Sony’s Click and Crackle Removal plug-in, setting Sensitivity to 10, Click shape to 8, and Max click size to 1.0 ms.

Clicks and pops from the music that originated on vinyl records, can, of course, tolerate more drastic settings.  Settings of 17, 5, and 2.0 respectively seemed to be a good compromise in such cases.  See the project notes of the individual recordings for more details.

One advantage of the almost ubiquitous 60Hz background noise on the source tapes is that it makes it easy to detect and correct pitch shift problems caused by incorrect speed settings, or variations when copying the original reel to reel tapes.  Extreme and variable pitch changes caused by the original tape machine starting and stopping were deemed too difficult to remove—they don’t occur very often, and perhaps only an additional syllable of intelligible speech could be recovered in this way.

Pitch shifts were processed with Adobe Audition 3.0.

This site is intended to become a complete archive of the audio messages of the late Hans R. Waldvogel.  Therefore, we present not only the edited recordings, but make the original media and project data available with the hope that over time, new and more effective noise reduction and audio enhancement techniques will become available.  You are invited to download the original files if you wish, and attempt to create a clearer representation of these audio messages.

Perhaps, in time to come, we will have the opportunity to document on this site other aspects of Hans Waldvogel’s life, and the ministries with which he was associated.

These recordings were of course originally on reel-to-reel tapes.  For many years, Hans Waldvogel’s nephew, Edwin H. Waldvogel, recorded his messages and produced 30-minute tapes for radio broadcast.  These master tapes, as well as tapes from regular church services, and talks at the faith home, accumulated over the years in Bro. Edwin’s basement, until he began a project to transfer the reel-to-reel recordings to cassette tape, evidently discarding the originals.

Many thanks to Paul Munsinger for tracking down these cassette originals, and granting the rights for them to freely appear publicly here.  And please note that the rights to these recordings are reserved insofar as while they may be freely copied and distributed, they may not be sold.

Thanks also to Gudrun Reichle for her invaluable contribution in providing German transcriptions, translations, research, and insight.

The goal has been to preserve as accurately as possible the spirit of these messages, and to make the words as clearly audible as possible.  Absolutely nothing has been removed for doctrinal, cultural, political, or reasons related to any other editorial sensitivity.

Some small edits have been made to remove what might be called distractions, such as long pauses, coughs, stutters, and background noises.  In many cases, it would be more distracting to remove the section containing a noise than to leave it, because the natural timing or inflection is altered, and so many of these rough spots can be found in the final processed recordings.  The most radical editing that has been done is occasionally to smooth over a misstatement.  For example, if the original says, “…he had to allow Hezekiah, …or Sennacherib… to blaspheme Jehovah…”, the edited version may have, “…he had to allow Sennacherib to blaspheme Jehovah…” 

A summary of the kinds of edits that have been made is presented in the project notes section for each recording, and of course, the original unedited file is always available as well.

The processed audio is presented in MP3 format.  The goal is to create audio suitable not only for internet streaming, but for download, distribution on CD, listening on higher quality equipment, etc.  For this reason, the processed audio uses bitrates somewhat higher than would normally be expected for internet distribution, and somewhat lower than would normally be expected for the other forms of distribution. 

So, then, the tuning of the data compression was done so as to be just at the threshold of perceptible compression artifacts.

Data compression was done with the LAME MP3 encoder, version 3.97 {->http://lame.sourceforge.net/}.  Most vocal material was compressed in this way:

lame -m m -V 3 --resample 11025

That is, monaural, vocal encoding of quality 3, (0 is the highest data rate, 9 is the lowest), and resampled to 11025Hz.  If a particular vocal recording was of exceptionally high quality it may be resampled to 22050 Hz instead.

The music on these recordings often has little more dynamic range than the voice recordings, however, they were processed using

lame --preset 112

for stereo music files (112kbps 44100Hz), and

lame -m m --preset 96

for monaural music files (96kpbs 44100Hz).

The need for multiple distribution targets is also the reason why each track of processed audio contains a few seconds of trailing silence.  This way, the tracks can be burned to an audio CD without needing to add short tracks of silence.

MP3 audio files, even at 256kbps, would not be considered sufficient archive media by any respectable archivist.  Yet, the generally poor quality of the originals, and the exceptionally low signal to noise ratio, means that realistically, digitizing at greater bit depths, sample rates, and with less data compression would preserve nothing beyond what is present in the noisy originals.

This, along with the desire to make this entire project publicly available at a manageable data size, means that the original masters are indeed MP3 files at 256kbps.

Processing of the audio was done using an older version of Sony Vegas, namely 5.0b (Build 160) from Sony Pictures Digital, Inc.  The Windows project files, having the extension “.veg” are included in the archive.  In addition, Vegas can output its project format to XML, which perhaps can be more easily interpreted in the event it becomes difficult to find an application that can open these archived .veg files directly.

In many cases, intermediate rendered files were generated from the originals, for example if click/pop removal was run, or a pitch shift was needed.  The Vegas project files, then, often reference these intermediate media files, rather than the original digitized masters.  It would consume far too much space to make these intermediate files available as well, so instead, detailed instructions are provided in each case for the creation of these intermediate renders.  Of course, if you want to work directly with the original audio, you are always free to ignore the Vegas project files altogether, and start from scratch using your own methods.

A rough target for volume level for the music was between -9 and -6dB(fs), with peaks above that, and -6 to -3dB(fs) for voice.  No hard limiting was done, so levels up to 0 dB(fs) can and do occur.

Audio Compression (that is, the processing of the dynamic range of the audio, as opposed to the data compression of the audio using MP3) is a topic that has entered somewhat into the public consciousness of late.  Be assured, that many experiments were done to see if it was feasible to reduce or eliminate the use of audio compression.  Certainly, for some talks, it could be done, but the long and the short of it is that considering the quality and character of the source material, as well as the ultimate goal—to make the words as clearly audible as possible—the use of a great deal of compression was unavoidable.

Our hope is that when those who were familiar with the previously available cassette duplicates hear these recordings, they will be pleasantly surprised how much more intelligible the words are, as well as relieved that constant volume adjustments are no longer necessary.

Experiments were done with various noise reduction techniques.  Unfortunately, most of the techniques that operate in the frequency domain, or are based off of a captured noise print, were found either to introduce very noticeable artifacts, or reduce the noise by such a small amount as not to be worthwhile.  Again, the masters are so noisy that these methods, which are so helpful in other situations, cannot successfully be employed here.

So then, most often simple EQ was used to mask 60Hz hum (with its harmonics), tape hiss, etc.  This is one area in which substantial improvement could be made in the future over what has been done here, although on many recordings it seems the higher frequencies contain nothing at all of the original signal, but only tape hiss and noise.

Starting with message 7A, some successful experiments using hiss reduction in Adobe Audition 3.0 were done.  The standard setting produced too many spectral artifacts, but the “Light Hiss Reduction” preset setting (noise floor 5dB, FFT 4096 points, precision 5, transition 5dB, decay 50%, 15dB hiss reduction) dealt with the hiss in a fairly clean and effective way.  The decision to use this hiss reduction was difficult, because when using it, some of the subtler cues of the noise presence of the room seemed to be eliminated, and the voice sounded just a shade more muffled (if that is possible given how muffled some of the originals are).  Nevertheless, there is no question that the level of hiss on many of the recordings—especially the sermons where audio compression raises the level of hiss—counts as a distraction that ought to be dealt with.

In some cases, noise that was more localized was removed, or rather attenuated, using the spectral frequency editing tool of Adobe Audition 3.0.

Most clicks, crackles, and pops that sound like static discharge can be effectively removed by using Sony’s Click and Crackle Removal plug-in, setting Sensitivity to 10, Click shape to 8, and Max click size to 1.0 ms.

Clicks and pops from the music that originated on vinyl records, can, of course, tolerate more drastic settings.  Settings of 17, 5, and 2.0 respectively seemed to be a good compromise in such cases.  See the project notes of the individual recordings for more details.

One advantage of the almost ubiquitous 60Hz background noise on the source tapes is that it makes it easy to detect and correct pitch shift problems caused by incorrect speed settings, or variations when copying the original reel to reel tapes.  Extreme and variable pitch changes caused by the original tape machine starting and stopping were deemed too difficult to remove—they don’t occur very often, and perhaps only an additional syllable of intelligible speech could be recovered in this way.

Pitch shifts were processed with Adobe Audition 3.0.