GSP1101 5150 Model and Impulse Testing

[aortizjr]

FINALLY got some time to mess with the GSP1101 and the new C55 beta firmware which allows you to load custom impulse responses. Well and ton of other features.

Did a shootout, details are in the blog.

http://ortizaudio.blogspot.com/2010/09/gsp1101-impulse-testing-with-5150-model.html

The issue I noticed is that not all impulses sound worth a shit since the GSP1101 can only load 128 taps of the impulse. It is also why the sound is drastically different between the loaded one and using LeCab.

But my Powerball one sounded best loaded into the GSP1101. But I only tried a few from Recabinet, Catharsis, and GuitarHack.

For reference I also used Nick Crow's 8505 and Catharsis' S-PresHigh impulse to give a frame of reference since they are popular around here. I wish I had a real 5150/6505 to really test with... ugh one on Craigslist... must resist... need to eat...

 

[Kazrog]

You're really doing some great work with your blog. I have to say, though, if by "taps" they mean "samples" - then no wonder the sound is so different. They're cutting off some essential data, and it will audibly affect the sound of any impulse response I can think of, as your clips demonstrate.

 

[aortizjr]

I am not 100% sure what is meant by "taps." That term is from the developer. I assume that means samples as well.

I haven't really dug into the impulses, they just sound good and I am happy with that and I will often shorten them in KeFir to save on some CPU when tracking.

With the 1101, I was really hoping it would be the saving grace for Direct, but it still needs some work and tweaking. I am gassing hardcore for an AxeFX and I just sold a bunch of gear to clear some room for a move. But $2k, that is a killer tube amp! And I am more the 1 or 2 raw good tone kind of guys, wouldn't use the other stuff the Axe offers, but I also want the multi-band and intelligent pitch shifting... grr.

Thanks on the blog. Try to keep things more central to share information. Also it is a great reference for me as well since I am so busy. I used to keep a digital "notebook" as just a collection of txt files. I noticed people actually referring to them as google found them, prompting me to start cleaning it up a bit.

 

[ahjteam]

It also might be bytes = 8 bits. 8 * 128 = 1024 bits ?

 

[Lazzari]

Dude it sounds awesome ! For the first 3 samples you are only using the 5150 model from the GSP1101 ?? Or are you using other amp there ??

 

[onqel]

Tap - A FIR "tap" is simply a coefficient/delay pair. The number of FIR taps, (often designated as "N") is an indication of 1) the amount of memory required to implement the filter, 2) the number of calculations required, and 3) the amount of "filtering" the filter can do; in effect, more taps means more stopband attenuation, less ripple, narrower filters, etc.)

source: http://www.dspguru.com/dsp/faqs/fir/basics


For convolution, a 128 tap filter means that the current sample is convolved with the last 128 samples.. which isn't very good

 

[aortizjr]

Dude it sounds awesome ! For the first 3 samples you are only using the 5150 model from the GSP1101 ?? Or are you using other amp there ??

It is just the 5150 II model from the beta firmware (stock doesn't have it) boosted with a TS808 model. All within the GSP1101.

The same model is in the RP500 and RP1000. Sonically the GSP1101 has a higher output and a 1M input impedance vs. the RP series 500K.

 

[aortizjr]

Tap - A FIR "tap" is simply a coefficient/delay pair. The number of FIR taps, (often designated as "N") is an indication of 1) the amount of memory required to implement the filter, 2) the number of calculations required, and 3) the amount of "filtering" the filter can do; in effect, more taps means more stopband attenuation, less ripple, narrower filters, etc.)

source: http://www.dspguru.com/dsp/faqs/fir/basics


For convolution, a 128 tap filter means that the current sample is convolved with the last 128 samples.. which isn't very good

Thanks! How many taps do you think is necessary for a good sound then? Or I guess how many are typically used?

Stan, the developer, said that he could potentially make it use 256 taps if you were to only run mono. Still may not be enough though.

 

[onqel]

Well.. I tried to make a plugin to load impulses, and I found out that I needed it use a buffer of at least 384 samples to make it usable..
He most likely needs to do FFT convolution if he wants to go much higher because it's superior to convolution at high buffersize..

 

[Scottxx]

Is the GSP1101 good? How does it compare to the AxeFX? Can it load any Impulse?
It seems to be a good unit to use it directly for IR loading. Which one is better? I know there are units that can load vst plugins too, but wich one is working really well for a guitarist?

 

[aortizjr]

Is the GSP1101 good? How does it compare to the AxeFX? Can it load any Impulse?
It seems to be a good unit to use it directly for IR loading. Which one is better? I know there are units that can load vst plugins too, but wich one is working really well for a guitarist?

I personally am still a fan of the GSP1101, especially at 1/4 the price of an AxeFX Ultra. Or even 1/3 at the price of a Standard. I don't think the AxeFX is overpriced for what it is, but at its price, you can get a lot of amazing sounding tube amps.

And really most guitarists only use a hanful of tones, I don't know of anyone that really needs and uses all 70 amps and 50 cabs or whatever in day to day use.

Now keep in mind, I haven't tried the AxeFX or the Eleven Rack, but I have tried ALL the others. Zoom, Line 6, Rocktron, Carvin, Boss, Vox, Johnson, Behringer, etc. and it was the first hardware unit to actually respond and feel like a real amp. It also had the balance of mud and fizz for a mix without the harshness, making it into my live rig.

When A/Bing with my tube amps (Powerball, Mesa Mark III, e530, Sovtek MIG, Fender HRD, Silverface Champ, etc.) It was also the first to not scream modeler, but more of a different tube amp or different setting.

In one band I play in, the guitarist will switch between the Legacy Sim and his X100B (in 4CM Mode) and I can barely tell the difference only when pointed out.

Recording however, I find that I am always fighting something. Same with VST's and impulses. I haven't mic'd the GSP1101 through a cab. But in the room, I am more than satisfied and club soundguys love it as well. Also though, it does do better at the low chug 7-string stuff than 6-string stuff. It is also less forgiving with pickups compared to say my ENGL's which sound great with just about anything. Hence why I picked my best sounding guitar for the demo

As an impulse loader, that was this test was all about. Really YMMV. You can load any 16-bit 44.1K impulse response. But some sound good, some sound not so good with the 128 tap/sample limitation. And it is a pretty big difference, as demonstrated in the clip.

The AxeFX however, can load anything and I think full length, getting the most out of the impulse. They also did a lot of other things right in their software, converters, and signal path. Basically no compromises, justifying the cost IMO. But can normal people really tell live in a shitty bar or at rehearsal at 125 dB in a square sheetrock jail cell using a moldy couch as a bass trap and a mini-fridge as a diffuser.... I don't think so.

 

[Kazrog]

Well.. I tried to make a plugin to load impulses, and I found out that I needed it use a buffer of at least 384 samples to make it usable..
He most likely needs to do FFT convolution if he wants to go much higher because it's superior to convolution at high buffersize..

True that... doing brute force convolution (especially on rack hardware) is nowhere near efficient enough for realtime processing with a high resolution IR. I didn't know anyone was doing brute force convolution for audio purposes, yikes. That said, getting efficient FFT going can be challenging.

 

[Lazzari]

It is just the 5150 II model from the beta firmware (stock doesn't have it) boosted with a TS808 model. All within the GSP1101.

The same model is in the RP500 and RP1000. Sonically the GSP1101 has a higher output and a 1M input impedance vs. the RP series 500K.

you have to pay for the new firmware ?? Or just download it for free?

 

[Force666]

http://www.mustbebeta.com


It's free. It is not supported by Digitech if that matters.

 

[Scottxx]

Thanks aortizjr for the info!

 

[RiF]

The AxeFX however, can load anything and I think full length, getting the most out of the impulse.

The AxeFx loads impulses up to 1024 samples of length. So they are truncated and you'll hear a difference when comparing them with the full length impulse applied in your DAW for example. Usually they lack some room information. IMHO, going below 1024 samples would reduce the quality significantly and probably below my threshold of acceptabililty.

But can normal people really tell live in a shitty bar or at rehearsal at 125 dB in a square sheetrock jail cell using a moldy couch as a bass trap and a mini-fridge as a diffuser.... I don't think so.

Yes, for live use, maybe it doesn't matter that much but in a recording situation you'll well hear the difference.