Jump to content
IGNORED

Offline Upsampling


Jud

Recommended Posts

For some time I've used and been happy with software that allows me to upsample in the computer, such as Audirvana Plus on the Mac, XXHighEnd on Windows, and HQPlayer on both, that allows me to avoid the upsampling filters in the DAC chip. They each offer slightly different "takes" on their upsampling filters. Audirvana Plus allows users to vary the parameters used by the bundled iZotope sample rate conversion software. HQPlayer provides a choice of several different types of filters created by the developer. XXHighEnd offers one filter created by the developer, plus numerous settings to modify the operation of the software and the computer on which it runs. These software players upsample inline, i.e., as part of the process of playing a music file.

 

For the past week or two I have been beta testing offline sample rate conversion software, Audiophile Inventory (AuI) ConverteR 48x44. AuI ConverteR had provided the ability to downsample files from higher sample rates and/or word lengths to lower ones - 24/88.2 to 16/44.1, or what have you. The developer, Yuri Korzunov, has now built in the ability to downsample or upsample between any word length and sample rate combination from 16/44.1 to 24/384. I have been using it to upsample lower resolution files to 24/352.8 or 24/384 files. Those are "native" resolutions for my DAC; it doesn't upsample those resolutions any further in PCM.

 

I compared offline upsampling to inline upsampling with iZotope, and I preferred offline. Not that the inline upsampling was bad at all, but music was clearer with offline upsampling. This wasn't shocking, because the inline upsampling uses the filter parameters I've tuned by ear, and someone who actually knows something about filters should be able to do better. Using a demo version of iZotope RX3, I compared the offline version of my filter parameters to AuI ConverteR 48x44. It came closer than the inline version (in other words the same filter sounded better offline than inline), but AuI ConverteR was stlll better. Again, not surprising. I then compared the very good new beta of HQPlayer for OS X. AuI ConverteR still sounded better - clearer, easier to follow individual instrumental and vocal lines. As Barry Diament has mentioned with regard to iZotope, and as I confirmed for myself by comparing my iZotope filter settings inline and offline, it appears offline filtering has a sound quality advantage. It would be interesting to listen to the HQPlayer filters applied offline, but as far as I know they're not available that way. I also do not have Windows on my MacBook Pro currently, so I wasn't able to compare XXHighEnd.

 

The "PROduceR" version of AuI ConverteR 48x44, the one with the ability to upsample or downsample to or from any rate/word length between 16/44.1 and 24/384, currently sells for $147.80 US for either Mac or Windows (or both at a discount for the bundle) at samplerateconverter.com. Either a linear phase or a minimum phase (no pre-ringing) filter can be used for conversion. I preferred the linear phase filter.

 

There are some disadvantages to offline oversampling. The resulting files are very large (near 1GB in some cases). It's a two-step process, first converting the file(s), then playing on other software, so it's not quite as convenient as all-in-one player software that upsamples inline. And there may be a potential disadvantage for developers, as I imagine it might be easier to reverse engineer an offline upsampling filter than an inline one. With regard specifically to AuI ConverteR, it doesn't offer the ability to upconvert to DSD right now, though that may be considered for addition in the future.

 

But overall, to me this is the best I can remember hearing my PCM files sound on the Mac, so I bought a license.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment

Jud: Your post makes complete sense, but it does raise the question of "how much more storage will I need?" Are you up-converting everything or just part of your library?

Synology NAS>i7-6700/32GB/NVIDIA QUADRO P4000 Win10>Qobuz+Tidal>Roon>HQPlayer>DSD512> Fiber Switch>Ultrarendu (NAA)>Holo Audio May KTE DAC> Bryston SP3 pre>Levinson No. 432 amps>Magnepan (MG20.1x2, CCR and MMC2x6)

Link to comment

Another option is to split out conversion and playback to separate computers while still doing it on the fly (or inline as you call it). That's what I did with the NAA thingie.

 

I believe main reason for any differences is due to isolation issues and that can be improved at the DAC side (such as exaSound) or to lesser extent using a separate isolator like iFi iUSB/iPurifier or NAA.

 

Main issue I have with offline conversion and one of the reasons for NAA is the hassle to maintain two sets of files and extra effort needed whenever there are advances/improvements to the filters/dithers/modulators. Also some of my DACs support DSD256 while others support only DSD128 or 384/32 PCM, so I would need multiple different conversions of the same content... So that would expand the required maintenance effort by several times.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

That is awkward ...there should be no differences whatsoever.

 

Different filters/upsamplers sound different anyway ... never experimented with "offline" upsampling but there shouldn't be differences unless the processing power is under capacity...

 

To me, HQPlayer sounds "smoother" when upsampling to DSD compared to Jriver on OSX and Mytek.

Link to comment
Have you compared this to Audiophile Engineering Triumph? Or Sample Manager?

 

Yes and no. :) Here's what I mean:

 

Both Triumph and Sample Manager use iZotope, but both, as far as I could tell, are limited to 24/192. My DAC will take input at 352.8/384 resolution, and that is the "native rate" in PCM to which my DAC chip interpolates (as it is for most). So I needed software that would get me to that resolution, which excluded Triumph and Sample Manager.

 

I did try the RX3 demo, which also uses iZotope and does upsample to 352.8/384kHz resolution. So the results with it are equivalent to the results I would have gotten with Triumph and Sample Manager if they were capable of the resolution I wanted. The sound quality was superior to inline conversion with iZotope as used in Audirvana Plus, but was not as good to my ears as the result with AuI ConverteR. That is reasonable because Yuri knows a great deal more about filters than I do.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Yes and no. :) Here's what I mean:

 

Both Triumph and Sample Manager use iZotope, but both, as far as I could tell, are limited to 24/192. My DAC will take input at 352.8/384 resolution, and that is the "native rate" in PCM to which my DAC chip interpolates (as it is for most). So I needed software that would get me to that resolution, which excluded Triumph and Sample Manager.

 

I did try the RX3 demo, which also uses iZotope and does upsample to 352.8/384kHz resolution. So the results with it are equivalent to the results I would have gotten with Triumph and Sample Manager if they were capable of the resolution I wanted. The sound quality was superior to inline conversion with iZotope as used in Audirvana Plus, but was not as good to my ears as the result with AuI ConverteR. That is reasonable because Yuri knows a great deal more about filters than I do.

 

Ah OK. I don't know why I thought AE did DSD. Is DSD in the works with AuI?

W10 NUC i7 (Gen 10) > Roon (Audiolense FIR) > Motu UltraLite mk5 > (4) Hypex NCore NC502MP > JBL M2 Master Reference +4 subs

 

Watch my Podcast https://www.youtube.com/channel/UCXMw_bZWBMtRWNJQfTJ38kA/videos

Link to comment
That is awkward ...there should be no differences whatsoever.

 

Not awkward at all, Leonardo, just different than perhaps you would expect. :) And of course this was not scientific, so I could easily be imagining things - I will never claim anything like infallibility, or "golden ears," or any of that.

 

I am sure you have access to RX3 or a demo, and the trial of Audirvana Plus is available if you don't already have it, so it is easy to try for yourself, though I imagine any differences would be most apparent if your DAC accepts 352.8/384kHz input resolution.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Ah OK. I don't know why I thought AE did DSD. Is DSD in the works with AuI?

 

As noted in the OP, "With regard specifically to AuI ConverteR, it doesn't offer the ability to upconvert to DSD right now, though that may be considered for addition in the future."

 

So let Yuri know in this thread if you want the ability to upconvert to DSD offline (I do), and perhaps if he sees there is a market he will consider doing the work necessary to add it sooner.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
As noted in the OP, "With regard specifically to AuI ConverteR, it doesn't offer the ability to upconvert to DSD right now, though that may be considered for addition in the future."

 

So let Yuri know in this thread if you want the ability to upconvert to DSD offline (I do), and perhaps if he sees there is a market he will consider doing the work necessary to add it sooner.

Yes, DSD128 would be nice. What is Yuri's moniker on CA?

W10 NUC i7 (Gen 10) > Roon (Audiolense FIR) > Motu UltraLite mk5 > (4) Hypex NCore NC502MP > JBL M2 Master Reference +4 subs

 

Watch my Podcast https://www.youtube.com/channel/UCXMw_bZWBMtRWNJQfTJ38kA/videos

Link to comment
Another option is to split out conversion and playback to separate computers while still doing it on the fly (or inline as you call it). That's what I did with the NAA thingie.

 

I believe main reason for any differences is due to isolation issues and that can be improved at the DAC side (such as exaSound) or to lesser extent using a separate isolator like iFi iUSB/iPurifier or NAA.

 

Main issue I have with offline conversion and one of the reasons for NAA is the hassle to maintain two sets of files and extra effort needed whenever there are advances/improvements to the filters/dithers/modulators. Also some of my DACs support DSD256 while others support only DSD128 or 384/32 PCM, so I would need multiple different conversions of the same content... So that would expand the required maintenance effort by several times.

 

Hi Miska, appreciate your input in the thread. Does that mean you have compared on the fly and offline upconversion, and feel NAA and/or a DAC like the exaSound gets you some or most of the advantage of offline conversion?

 

Don't you offer "multiple different [on the fly] conversions of the same content" now for different DACs? How would this be different with offline conversion?

 

With regard to two sets of files and changes when there are changes to filters, etc. - so it's not as simple as just plugging the same filters, etc., into an inline and an offline version, eh? :)

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Yes, DSD128 would be nice. What is Yuri's moniker on CA?

 

audiventory

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Jud: Your post makes complete sense, but it does raise the question of "how much more storage will I need?" Are you up-converting everything or just part of your library?

 

Yep, that's certainly the issue foremost in my mind. I'm not upconverting everything, and I'm also not keeping all the files I upconvert. I only retain a few. The rest I feel that I can convert again whenever I want, so why take up the space (until storage becomes even cheaper than it is now)?

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Hi Miska, appreciate your input in the thread. Does that mean you have compared on the fly and offline upconversion, and feel NAA and/or a DAC like the exaSound gets you some or most of the advantage of offline conversion?

 

Yes, that's the conclusion I have. Maybe someone who has tried this can also comment...

 

Don't you offer "multiple different [on the fly] conversions of the same content" now for different DACs? How would this be different with offline conversion?

 

Yes, but of course doing five different on the fly conversions of the same content consumes zero additional disk space and is not a maintenance burden. I only need to maintain five different settings, but not six different content sets. Making five different offline conversions of all the content I have is major disk space and maintenance burden, whenever I'd add more content or re-conversion due to some conversion changes...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

"I believe main reason for any differences is due to isolation issues and that can be improved at the DAC side (such as exaSound) or to lesser extent using a separate isolator like iFi iUSB/iPurifier or NAA."

 

Sorry if it's an obvious answer, but what is it about the Exasound which makes it preferable as regards isolation issues?

 

Thanks,

 

Joel

Link to comment

Hi Gents:

I am a little late to this thread but the topic is of great interest to me. Yuri has sent me links to a few versions over the past month (at first they did not work on my AIFF files and then did not preserve any metadata, but now it seems to work fine), but I have been very busy and my Mac mini has been in pieces (still) on my desk as I do fitment tests for our fan controller board.

A while back I did some experiments with Triumph and Sample Manager (one of the trials did give access to the Advanced iZotope settings), and when I was able to produce a filter close to my fine tuned A+ settings, I also found that the offline created files sounded a bit better. But for me it was not enough to be worth the trouble and storage space, though I did keep a few favorite test tracks I created. The trouble is, those files are of course a snapshot of settings preferred at that time.

 

I am looking forward to testing AuI ConverteR 48x44--especially now that Jud, who knows well what can be accomplished with A+ and iZotope, has reported such good results. I do think that Jud continues to sell short his own ears and ability to tune a good PCM SRC. Yet certainly I am curious about the quality of Yuri's filters --as well as about Miska's algorithms. I have long suspected that the magic in HQPlayer is very much about its high-rate SDMs, as well as about splitting the processing/rendering load with the player/NAA approach. I don't currently have a DSD-capable DAC, so I can't utilize the former, and I spent a bunch of time with Jussi's help trying to get Debian/NAA working on my DAc-attached Intel Mac--to no avail.

 

My abilities to try HQPlayer will soon change, both due to the beta version of OS X HQPlayer not requiring an NAA, and due to the arrival this next week of a new i7 Mac mini, capable of running his more processor-intensive algorithms.

 

I will try AuI ConverteR 48x44, but to properly compare its offline to A+/iZotope real-time--to get a sense if the benefit was due to the off-line processing or to the design of the filter--it would be nice to try to closely replicate Yuri's filter with iZotope. I guess I need to bone up again on the use of RX3 to run graphs and impulse response shots. I could first run a 16/44.1 impulse file through AuI ConverteR 48x44, and then the same through iZotope (with RX3 Advanced iZotope plug in, or with Sample Manager w/Action Pack) to get as close as possible. Did I mis-read your original post Jud, or is that essentially what you did to make your off-line/real-time comparison? Aside from just listening of course.

 

Anyway, great topic. I don't have much to contribute at this point, but I will be watching closely and hope to do some of my own SQ experiments soon.

 

Regards,

 

Alex C.

Link to comment

Hi Alex. I did not try to reverse engineer Yuri's filter. :) What I did was listen to A+ with my iZotope settings; iZotope RX3 playing the same track converted offline with those same settings; and the same track converted offline with AuI played back with RX3. With my settings I preferred offline to inline. I preferred AuI to both.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Hi Alex. I did not try to reverse engineer Yuri's filter. :) What I did was listen to A+ with my iZotope settings; iZotope RX3 playing the same track converted offline with those same settings; and the same track converted offline with AuI played back with RX3. With my settings I preferred offline to inline. I preferred AuI to both.

Would you be willing to share the AuI setting that you're using?

W10 NUC i7 (Gen 10) > Roon (Audiolense FIR) > Motu UltraLite mk5 > (4) Hypex NCore NC502MP > JBL M2 Master Reference +4 subs

 

Watch my Podcast https://www.youtube.com/channel/UCXMw_bZWBMtRWNJQfTJ38kA/videos

Link to comment
Hi Alex. I did not try to reverse engineer Yuri's filter. :) What I did was listen to A+ with my iZotope settings; iZotope RX3 playing the same track converted offline with those same settings; and the same track converted offline with AuI played back with RX3. With my settings I preferred offline to inline. I preferred AuI to both.

 

Thanks for the explanation Jud. But why did you play back your AuI-converter files with RX3 instead of with A+ (with SRC turned off of course)? Tha that should yield the best SQ overall.

 

Jason: AuI only gives a choice of two filters, and Jud said that he prefers the linear phase one.

Link to comment
Thanks for the explanation Jud. But why did you play back your AuI-converter files with RX3 instead of with A+ (with SRC turned off of course)? Tha that should yield the best SQ overall.

 

Jason: AuI only gives a choice of two filters, and Jud said that he prefers the linear phase one.

 

That's the way I play them now. :)

 

But I wanted a fair comparison between AuI and my iZotope settings first to tell what I preferred. The RX3 demo doesn't allow saving converted files, so I couldn't play both with A+.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment

Hey Jud, you mentioned you compared this method of offline up-sampling with HQPlayer and preferred the result of the offline samples.

 

Do you have a DSD capable DAC? If so did you compare the offline up-sampled PCM with the original PCM files converted to DSD realtime in HQPlayer?

 

I ask because I'm preferring my PCM converted to DSD right now.

 

I have tried offline up-sampling with AE's sample manager, which uses iZotope and the offline up-sampler supplied with Amarra. I got some good results.

 

I also did some offline up-sampling using the Japanese "Bug Headed Emperor" Although here I was limited to upping to 88.2. Bug Head produces raw files and I couldn't find a way of converting these to a playable format such as flac at higher sample rates.

 

For me the interesting aspect of this, is that you can take the resulting file and play it in any player you choose. The player itself will, obviously have an influence. Since you are offline up-sampling to the max PCM sample rate, you obviously wouldn't want to try any further PCM up-sampling with the resulting files (except perhaps up to 705.6 in XXHighEnd).

 

I'm also wondering if you had the "Dither" option turned on or off.

 

What you could do though is take those files and convert them inline to DSD using HQPlayer, if you have a DSD capable DAC.

 

One other point to bear in mind is that when using inline/realtime up-sampling with or without conversion between PCM and DSD, it might push your CPU too hard and I found this does detrimentally effect the sound.

 

On the Mac you could check Activity Monitor when doing inline up-sampling and see what CPU reading you get. On the Windows side I use Process Explorer.

 

I have considered creating a separate library on its own drive with offline up-sampled files of select music. It is a lot of work and additional expense though. I'd have to be convinced the results were really worth it.

 

With good immunity as Miska mentioned and a powerful enough CPU to handle the up-sampling and conversions inline, I question if it really is worthwhile.

 

geoff

Owner of: Sound Galleries, High-End Audio Dealer, Monaco

Link to comment
Hey Jud, you mentioned you compared this method of offline up-sampling with HQPlayer and preferred the result of the offline samples.

 

Do you have a DSD capable DAC? If so did you compare the offline up-sampled PCM with the original PCM files converted to DSD realtime in HQPlayer?

 

I ask because I'm preferring my PCM converted to DSD right now.

 

I have tried offline up-sampling with AE's sample manager, which uses iZotope and the offline up-sampler supplied with Amarra. I got some good results.

 

I also did some offline up-sampling using the Japanese "Bug Headed Emperor" Although here I was limited to upping to 88.2. Bug Head produces raw files and I couldn't find a way of converting these to a playable format such as flac at higher sample rates.

 

For me the interesting aspect of this, is that you can take the resulting file and play it in any player you choose. The player itself will, obviously have an influence. Since you are offline up-sampling to the max PCM sample rate, you obviously wouldn't want to try any further PCM up-sampling with the resulting files (except perhaps up to 705.6 in XXHighEnd).

 

I'm also wondering if you had the "Dither" option turned on or off.

 

What you could do though is take those files and convert them inline to DSD using HQPlayer, if you have a DSD capable DAC.

 

One other point to bear in mind is that when using inline/realtime up-sampling with or without conversion between PCM and DSD, it might push your CPU too hard and I found this does detrimentally effect the sound.

 

On the Mac you could check Activity Monitor when doing inline up-sampling and see what CPU reading you get. On the Windows side I use Process Explorer.

 

I have considered creating a separate library on its own drive with offline up-sampled files of select music. It is a lot of work and additional expense though. I'd have to be convinced the results were really worth it.

 

With good immunity as Miska mentioned and a powerful enough CPU to handle the up-sampling and conversions inline, I question if it really is worthwhile.

 

geoff

 

- My mid-2009 MacBook Pro does not have sufficient CPU resources to convert PCM inline to DSD128 via HQPlayer without audible defects. It will convert inline to DSD64 via HQPlayer without identifiable audible defects, but the result doesn't sound as good as HQPlayer upsampling to 352.8/384kHz PCM. Perhaps this is due to the CPU operating near its limits, because most DSD files sound superb. - I always use dither when upsampling 16 bit word length files to 24. When going from 24 bit to 24 bit I have tried both with and without dither, but haven't really paid attention to determine whether I think there's an audible difference. - Please note Miska's response in this thread where he says that for a given filter offline conversion will sound better, but use of an NAA or particular DACs such as the exasound can make up a good part of the difference. - AuI does not currently do offline conversion to DSD, but I think Yuri may be open to adding that capability if he sees a demand. If you think you'd be interested, let him know in this thread or via private message (member name audiventory). - For me AuI is more useful than Sample Manager or Triumph, and perhaps even the more costly RX3, for two reasons: (1) Sample Manager and Triumph are limited, as far as I am able to tell, to 24/192kHz max. That would leave me still having to use inline filtering to get to the "native" rates of my DAC chip (and the "native" rate of most other folks' DAC chips as well), 352.8/384kHz. So I would not get the benefit of eliminating inline filtering. (2) For all three, Sample Manager, Triumph, and RX3, I would be using my own filter settings and thus not taking advantage of the expertise of someone who really knows about filters.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment

I found myself getting into a discussion of this topic on another thread. I've been doing offline upsampling for several years now, generally to 48kHz, since that's the highest rate the iPod Touch supports and I want my iTunes library to stay under 1 TB for at least another year. This is also exactly half the maximum rate supported by the DAC connected to my computer, minimizing the calculations needed to interpolate when I play back from my computer. I also do offline downconversion when I download an album in 96kHz and want to play it on my iPod. My goal is always to make my music files as independent of playback electronics as possible. I use Sox, since it is free and easy to control with scripts. I typically process an album with a batch script, such as the following, which I recommend for early digital recordings made with a PCM1610 or the like (this is the Ruby version of the script):

thr = []
fname = "resampled"
Dir.mkdir fname unless File.directory? fname
Dir.foreach(".") do |x|
 next unless x.match(/aiff?\Z/)
 thr << Thread.new do
   system("sox",'-D',
           x,
           fname + '/' + x,
    'gain','-0.1',
    'rate','-v','-s','48000',
    'sinc','-a','60','-t','2400','-18300')
   puts x + ": done"
 end
end
thr.each { |i| i.join }


 

This processes every track in parallel using threads, uses a "very high quality" (-v) sample rate converter (180dB stopband attenuation) with a "steep" (-s) cutoff (1% transition bandwidth), and uses the "sinc" effect to actually process the audio with a brickwall filter that I make as un-brickwall as possible by reducing its stopband attenuation to 60 dB and increasing its transition bandwidth to 2400 Hz. 18300 Hz is the cutoff at which I find I lose any artifacts of the phase distortion characteristic of the old analogue brickwall filters; the digital filter that I use here is of course phase-linear.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...