Izotope SRC

[bdiament]

...I'm still unclear regarding which is considered better, max sample rate or multiple of 2. I simply have mine on max. Anybody care to chime in on that?

 

Hi DreamOperator,

 

In the "old" days, sample rate conversion algorithms did less damage (injected fewer artifacts, such as the brightening and hardening many do today) when performing integer conversion (i.e., even multiples - halving or doubling the sample rate). This may still be true of the lesser algorithms today. However, the better algorithms are better at doing the math and to my ears, will deliver more transparent results even when performing non-integer conversion than lesser algorithms will when performing the easier, integer conversion. I find this particularly true of my long-time favorite, Alexey Lukin's 64-bit SRC from iZotope. The algorithm is smart and just doesn't care whether the conversion is integer or not. It can handle the math.

 

As always, I can only report on how *I* hear it and suggest no one simply take my word for what *they* will hear. My best recommendation is for folks to try it both ways using a few different, high quality sources. Then compare the results and draw their own conclusions.

 

Hope this helps.

 

Best regards,

Barry

Soundkeeper Recordings

The Soundkeeper | Audio, Music, Recording, Playback

Barry Diament Audio

[DreamOperator]

Hi Barry,

 

Thanks for the reply.

 

From past discussion I recall that you prefer offline SRC. Have you tried Audirvana+ with realtime iZotope SRC?

 

Regards,

 

Tim

[bdiament]

Hi Barry,

 

Thanks for the reply.

 

From past discussion I recall that you prefer offline SRC. Have you tried Audirvana+ with realtime iZotope SRC?

 

Regards,

 

Tim

 

Hi Tim,

 

I haven't tried the latest Audirvana+ but I have the SRC as a separate entity, which I can apply on or offline to any of a number of applications.

I still prefer SRC offline. What about you? What do you hear between the two methods?

 

Best regards,

Barry

Soundkeeper Recordings

The Soundkeeper | Audio, Music, Recording, Playback

Barry Diament Audio

[DreamOperator]

Hi Barry,

 

I don't currently have the ability to do it offline, but I'm very pleased with the results in Audirvana+.

 

Are you still using Wave Editor for offline conversion?

 

Regards,

Tim

[bdiament]

Hi Barry,

 

I don't currently have the ability to do it offline, but I'm very pleased with the results in Audirvana+.

 

Are you still using Wave Editor for offline conversion?

 

Regards,

Tim

 

Hi Tim,

 

I have the algorithm in several applications now. In addition to Wave Editor, it is in Triumph, Sample Manager and iZotope's own RX 3.

I have tended toward the batch processor in Sample Manager, which allows me to tell the app to convert multiple files on its own. (Tell it where the files are, tell it what process - such as what target sample rate - and let it loose. Very easy.)

Just upgraded from RX 2 to RX 3, which adds the SRC algorithm to its batch processor too.

 

Best regards,

Barry

Soundkeeper Recordings

The Soundkeeper | Audio, Music, Recording, Playback

Barry Diament Audio

[Superdad]

Hi Tim,

 

I have the algorithm in several applications now. In addition to Wave Editor, it is in Triumph, Sample Manager and iZotope's own RX 3.

 

Hi Barry:

 

I am a big fan of the iZotope SRC as implemented in Audirvana. While offline SRC is of some interest to me (mostly to be able to provide a few reconverted files to friends so they can bypass their DAC's own filters), it does have the drawbacks both of large file size and of baking-in a particular filter.

 

It is useful for me on some occasions, but as I recall from my trial of Triumph many months ago, it does not actually give the same complete control over the filter as their own OEM engine does as implemented in Audirvana. What I mean by that is that A+ provides for control over all the following parameters (with numerical and slider entry):

Steepness

Filter Max. Length

Cut-off Frequency

Anti-alising (that's just the final depth of attenuation)

Pre-ringing (balance of minimum phase to linear phase)

 

I find that adjustment of all the above can be critical for optimizing a really great sounding filter (I've tuned a sows-ear DAC into a silk purse on more than one occasion). The anti-aliasing (max. attenuation) is the only one I leave at its maximum setting.

 

I seem to recall Triumph leaving out adjustability of one or two of the above. Can you confirm this and also let us know about what parameters SampleManager offers? While I have the basic SampleManager app, iZotope does not trial the Action Pack which is necessary for access to the advanced SRC settings, so I can't cheek that of myself.

 

Thanks and regards,

Alex Crespi

[bdiament]

Hi Barry:

 

I am a big fan of the iZotope SRC as implemented in Audirvana. While offline SRC is of some interest to me (mostly to be able to provide a few reconverted files to friends so they can bypass their DAC's own filters), it does have the drawbacks both of large file size and of baking-in a particular filter.

 

It is useful for me on some occasions, but as I recall from my trial of Triumph many months ago, it does not actually give the same complete control over the filter as their own OEM engine does as implemented in Audirvana. What I mean by that is that A+ provides for control over all the following parameters (with numerical and slider entry):

Steepness

Filter Max. Length

Cut-off Frequency

Anti-alising (that's just the final depth of attenuation)

Pre-ringing (balance of minimum phase to linear phase)

 

I find that adjustment of all the above can be critical for optimizing a really great sounding filter (I've tuned a sows-ear DAC into a silk purse on more than one occasion). The anti-aliasing (max. attenuation) is the only one I leave at its maximum setting.

 

I seem to recall Triumph leaving out adjustability of one or two of the above. Can you confirm this and also let us know about what parameters SampleManager offers? While I have the basic SampleManager app, iZotope does not trial the Action Pack which is necessary for access to the advanced SRC settings, so I can't cheek that of myself.

 

Thanks and regards,

Alex Crespi

 

Hi Alex,

 

I have my favorite settings and very specific uses for the SRC so I have no problem with file size or wanting to change the filtering.

To answer your question, I've looked at Triumph, Sample Manager with the Action Pack and at iZotope's own RX 3 to see what controls are offered.

Here are screen captures from all three.

 

First, Sample Manager with the Action Pack

 

Next, Triumph

 

Now, RX 3

 

Hope these are of use to you.

 

Best regards,

Barry

Soundkeeper Recordings

The Soundkeeper | Audio, Music, Recording, Playback

Barry Diament Audio

[Superdad]

Thanks Barry. That's about what I remember from trials of RX3 and Triumph. Is RX3 really the only one that shows a numeric value for the sliders? Or do SampleManager and Triumph show a number when you grab the slider?

 

I do wish they allowed for control of filter length (number of taps), as I recall Alexey saying it was fixed at some maximum amount (A+ max filter length is 2,000,000, but I don't think it is really a 2 million-tap filter!). Repeated experiments always land me down around 1.3 million.

 

And slope settings finer than whole number steps would be very much welcome--A+'s implementation of iZotope SRC does not offer this either.

 

Do you use the iZotope SRC more for down conversion than upsampling? That would make sense to me since you are on the recording end of the business.

 

Regards,

ALEX

[bdiament]

Thanks Barry. That's about what I remember from trials of RX3 and Triumph. Is RX3 really the only one that shows a numeric value for the sliders? Or do SampleManager and Triumph show a number when you grab the slider?

 

I do wish they allowed for control of filter length (number of taps), as I recall Alexey saying it was fixed at some maximum amount (A+ max filter length is 2,000,000, but I don't think it is really a 2 million-tap filter!). Repeated experiments always land me down around 1.3 million.

 

And slope settings finer than whole number steps would be very much welcome--A+'s implementation of iZotope SRC does not offer this either.

 

Do you use the iZotope SRC more for down conversion than upsampling? That would make sense to me since you are on the recording end of the business.

 

Regards,

ALEX

 

Hi Alex,

 

All three show numeric values. Sample Manager has them to the right of the sliders (not shown in the screen grab I attached) and Triumph shows the number when you grab one of the sliders.

 

Since I record everything at 24/192, my prime use for SRC as far as my own recordings is to create the 24/96 and 16/44 versions.

However, when I master, I use upconversion as most sources come in at 24/44 and I find benefit in doing the mastering at a higher res, even if the target is a CD master.

When I'm just listening to the server for my own pleasure, I tend to listen to everything at its native rate, with no additional processing.

 

Alexey has certainly created some wonderful algorithms.

 

Best regards,

Barry

Soundkeeper Recordings

The Soundkeeper | Audio, Music, Recording, Playback

Barry Diament Audio

[Superdad]

Thanks again Barry. With that info I think I will go ahead and purchase the Action Pack for Sample Manager. Will come in handy.

 

Do you know if iZotope has ever licensed its SRC engine for embedded use? That is, if a designer was building a DAC with, say an ARM processor, and wanted to include a range of custom filters, would they license the iZotope SRC for that? I know it is a question for the company, but they recently changed their web site and now I can't even find out what platforms all their packages run on. Just OS X and Windows? Anyway, I was just wondering if perhaps you have seen some pro hardware using iZotope technologies.

One alternative that we have already designed filters with is SoX, and for that my engineer partner wrote utilities to output the filter coefficients to a file (after we carefully tuned the filter by ear on my system) and then he loads them into SRC routines he wrote for a big FPGA.

[bdiament]

Thanks again Barry. With that info I think I will go ahead and purchase the Action Pack for Sample Manager. Will come in handy.

 

Do you know if iZotope has ever licensed its SRC engine for embedded use? That is, if a designer was building a DAC with, say an ARM processor, and wanted to include a range of custom filters, would they license the iZotope SRC for that? I know it is a question for the company, but they recently changed their web site and now I can't even find out what platforms all their packages run on. Just OS X and Windows? Anyway, I was just wondering if perhaps you have seen some pro hardware using iZotope technologies.

One alternative that we have already designed filters with is SoX, and for that my engineer partner wrote utilities to output the filter coefficients to a file (after we carefully tuned the filter by ear on my system) and then he loads them into SRC routines he wrote for a big FPGA.

 

Hi Alex,

 

While I have seen iZotope's SRC and MBIT+ dither algorithms in several software applications, both pro and consumer, I have not yet seen these in any hardware products. That said, I notice their site mentions hardware products, which I suppose are in their future plans, though I have no idea what sort of products those might be.

 

As far as the operating systems they write for, so far, I've only seen mention of OS X and Windows.

 

Best regards,

Barry

Soundkeeper Recordings

WordPress.com

Barry Diament Audio

[tiff_needle]

As anyone played with iZotope SRC on A+ with the new Geek Out?

 

Curious to know which settings some of you might be using.

 

It's just a shame that it doesn't do PCM->DSD128 on the fly...

[Kees de Visser]

Hi,

I'm trying to understand how (pre-)ringing of a >20kHz filter can be audible. Are there any threads in this forum, or online articles (preferably scientific) worth reading ? Thanks.

[Jud]

Hi,

I'm trying to understand how (pre-)ringing of a >20kHz filter can be audible. Are there any threads in this forum, or online articles (preferably scientific) worth reading ? Thanks.

 

Hi, Kees. Miska has posted previously that although the ringing itself occurs in the ultrasonic region and cannot be heard, another effect is that the filtered signal is smeared in the time domain by as much as the time taken by a 1kHz sine.

[Kees de Visser]

That's interesting since I can't confirm <20kHz artifacts after SRC with iZotope. Remember the image I posted in another thread recently ? Thanks for mentioning Miska, I'll ask him for details.

[Jud]

Folks - After having a chance to listen to music with my completed semi-DIY DAC and new speakers, I determined that I wanted to modify the iZotope filter parameters just slightly from those I'd been using. (What specifically got my attention was that music through XXHE with PeterSt's Arc Prediction filter seemed to come from a blacker background than when I used my filter parameters with iZotope. And Peter had mentioned to me that with the very mild steepness setting I'd been using, I could well be hearing harmonic distortion that was euphonic.) So I installed the RX3 trial, and made the steepness slightly greater (changed just from 3 to 5), which according to the RX3 scope simulation should put any distortion at frequencies I can hear (i.e., up to about 16000 Hz) pretty low in level. What I'd like to see is what this has done to impulse response/ringing, but I can't figure out how to get RX3 to show this. Any help? Thanks.

[Superdad]

...and made the steepness slightly greater (changed just from 3 to 5), which according to the RX3 scope simulation should put any distortion at frequencies I can hear (i.e., up to about 16000 Hz) pretty low in level. What I'd like to see is what this has done to impulse response/ringing, but I can't figure out how to get RX3 to show this. Any help? Thanks.

 

But how does it sound to you Jud? As you well know, the slope affects multiple aspects at once, so just looking at the impulse response may be interesting, but won't tell you what is best. If you settle on an A+/iZotope steepness setting of 5, you are now just two steps away from the 7 I have been living with for over a year (and I'd probably prefer a 6.5 if that was possible). We are converging buddy!

[Jud]

But how does it sound to you Jud? As you well know, the slope affects multiple aspects at once, so just looking at the impulse response may be interesting, but won't tell you what is best. If you settle on an A+/iZotope steepness setting of 5, you are now just two steps away from the 7 I have been living with for over a year (and I'd probably prefer a 6.5 if that was possible). We are converging buddy!

 

It sounds wonderful, but then it did before. I would not have been motivated to change anything if I hadn't listened to filtering done by someone who clearly knows what he's doing (PeterSt) and identified that blacker background.

 

Yep, I know steepness affects multiple aspects of the sound, which is why I've asked how I can get RX3 to show me what the impulse response looks like. Do you know?

 

We do seem to be converging, which may mean we're onto something, or just on something.

[Superdad]

Yep, I know steepness affects multiple aspects of the sound, which is why I've asked how I can get RX3 to show me what the impulse response looks like. Do you know?

 

I did it a couple of times in the past with some impulse files that were given to me by someone here in this very thread. I have the memory of a goldfish, so I don't know how to repeat the procedure (besides, my RX trial period ran out long ago and their protection embeds some tiny file so deep in some hidden OS folder that I have never found it to be able to delete and do another trial.) It might have been Barry who guided me, but looking back quickly I think it was Junker or Levandier. Just go to the first pages of this thread.

 

Ciao,

AJC

[Alexey Lukin]

Yep, I know steepness affects multiple aspects of the sound, which is why I've asked how I can get RX3 to show me what the impulse response looks like. Do you know?

Create a unit pulse by silencing a few seconds of audio, then mouse-dragging a single sample upwards. Then upsample this pulse by a factor of 10 or 20 (by typing in the required rate). That will give you the impulse response.

[Alexey Lukin]

If you settle on an A+/iZotope steepness setting of 5, you are now just two steps away from the 7 I have been living with for over a year (and I'd probably prefer a 6.5 if that was possible).

In the next version of RX, you'll be able to type .5 numbers in the Steepness field. The latency will also be fully compensated to simplify the null-testing (no more fractional latency).

[Superdad]

In the next version of RX, you'll be able to type .5 numbers in the Steepness field. The latency will also be fully compensated to simplify the null-testing (no more fractional latency).

 

Dear Mr. Lukin:

 

Thank you for the good news about finer granularity for Steepness becoming available in next edition of RX. Will you also make that available to your iZotope SRC OEM clients--specifically Damien Plisson of Audirvana? Realtime upsampling via the iZotope engine in Audirvana Plus is my main way of using your fine product at this time, and finer control of the advanced settings--particularly the Steepness--would be a very welcome enhancement.

 

Also, can you tell me if your iZotope SRC (with full controls) could be made available for license to a hardware OEM? In other words, if someone was developing a player/DAC that had processor running a Linux variant, could they license your s/w for "embedded" use in that product? Feel free to send me a private message regarding this.

 

Regards,

 

--Alex Crespi

UpTone Audio LLC

[Jud]

Create a unit pulse by silencing a few seconds of audio, then mouse-dragging a single sample upwards. Then upsample this pulse by a factor of 10 or 20 (by typing in the required rate). That will give you the impulse response.

 

Thank you for your help, much appreciated.

[Bob Stern]

 

 

[Bob Stern]